CSE:CCI Canadian Copper Inc.

Lifecycle: a funded, PEA-stage base-metals developer. Price-to-earnings, EBITDA and return on equity are meaningless here: there is no production ^{[1] [9]} and the provider returns zero revenue ^[21]. That one basis is what makes every “not applicable” in this report — no ROIC, no free-cash-flow yield, no PEG, no dividend — a statement about the company's stage rather than a gap in the data. Quality is scored on the pre-production ladder: study strength, distance to production, permitting, funding and control of infrastructure. Every component and its weight is published below, because a pillar that decides a Decision-Matrix row should not arrive as a bare number.

Rounded, that is 70. The weighted build of 69.50 sits 4.5 points clear of the 65 boundary between High and Medium, so the matrix row does not turn on rounding.

Balance sheet — and the number I have improved on my own last report

The 11 September report put net cash near C$18M — a figure this report supersedes, and the reason it can is on the record. That report explicitly labelled the number an estimate rolled forward from the 31 January 2026 filing, and named the April interim statements as not retrieved ^[19]. That figure was on the record: the 16 July financing release discloses available cash of C$17.3M at 30 April 2026 ^[2]. Rebuilt from that anchor — plus the C$12.5M OR receipt of 16 July ^[2], plus the C$3,181,250 August warrant exercise ^[3], less the US$2.0M of debt repaid ^[4], less C$7–11M of five months of outflow — net cash is C$21.20M on a range of C$19.20M to C$23.20M. It is still an estimate, because the July-quarter statements were not available, and it is still labelled one.

Share count — where the data provider would have been wrong by 6%

Derived from a primary filing, not taken from a provider. The early warning report of 13 August 2026 states Ocean Partners held 45,397,223 shares representing 21.34% of the issued and outstanding after exercising 12,725,000 warrants; 45,397,223 ÷ 0.2134 = 212,733,004 ^[3]. The independent cross-check from the same filing agrees to within 0.03%: 32,672,223 ÷ 0.1633 = 200,075,462 before the exercise, plus 12,725,000 = 212,800,462. The provider would have been wrong: Yahoo's market capitalisation of C$140.26M at C$0.70 implies 200.4M shares, which is precisely the pre-11-August count ^[21] — using it would have overstated NAV per share by 6%. Fully diluted adds the 23,685,200 warrants at C$0.25 ^[4] to the 212,733,004 basic derived above ^[3], for 236,418,204.

Competitive moat — 54 / 100

A price-taker with no network effects cannot score well here, and it should not. The moat is real but narrow: it is one permitted mill, not a franchise.

Capital allocation and skin in the game

There is no ROIC to compute. What there is: management funded a C$64M capex with a stream and a debt facility rather than by issuing equity at C$0.25–0.70, which for a junior is the difference between building a mine and building a share count. And the strategic holder is a buyer — Ocean Partners early-exercised 12,725,000 warrants for C$3,181,250 of cash on 11 August 2026 and took its stake to 21.34% ^[3]. Scored 75 for allocation and 70 for alignment.

The anchor is price to net asset value, built bottom-up from the Ausenco PEA held at the study's own base-case deck — not at spot. That is the SKILL's miner rule, and it is the conservative choice: the metal basket currently sits at 1.5722× the deck, so holding the project at deck deliberately leaves that uplift out of the NAV and shows it as scenario upside instead.

The NAV, term by term

So the actual multiple is C$0.70 ÷ C$0.7853 = 0.8914×, against a warranted 0.65×, for a ratio of 1.37× — the FULL band. (Exactly 1.3714× before rounding, and the band test uses that unrounded value — the framework's rule, and it matters because a name within a hundredth of an edge would otherwise be banded by its own rounding. Round 1 of this report's audit found the previous version of this sentence falsified by its own page: it claimed every figure showed two decimals while the scorecard, this section's own header and five calibration fields showed four. Every display of the ratio now renders from one field.)

Threshold fragility — derived, not asserted

Ratio 1.3714, FULL band. NOT ROBUST. It is 3.8 NAV units from the 1.40x edge and 26.5 from 1.20x. Largest single sensitivity: capex revision at the December 2026 feasibility study (pea_capex_revision), worth 0.158 on the ratio and closing 506% of the nearer gap (to 1.40x) — it CROSSES the edge on its own. OR financing basis - credit the surplus, or treat the pair as value-neutral: 0.049, covering 170% of that same gap. corporate overhead and owner's costs to first production: 0.050, covering 98% of that same gap. marginal tax rate applied to the stream give-up: 0.037, covering 78% of that same gap. net cash at the report date: 0.029, covering 53% of that same gap. discount rate applied to the stream give-up: 0.049, covering 0% of that same gap. Taken together the declared inputs close 904% of the gap to 1.40x and 52% of the gap to 1.20x. The fragility is ONE-SIDED and points WORSE. Do not claim both adjacent bands sit inside the envelope — that claim was published once and was false.

In plain terms: the published Full band is one accounting choice and one feasibility study away from a HOLD on every horizon. A 40% construction discount — the midpoint of the framework's own range — puts the ratio at 1.4856×. A value-neutral reading of the OR pair puts it at 1.4207×. Either of those would fire Gate 3, which would then cap all three horizons at HOLD — and because no horizon would carry a BUY, it would auto-Stop this Donatien Pick. Nothing in this report matters more than that, and the 11 September report disclosed the same fragility on different arithmetic.

Sensitivity — one row per declared model input

Read the last column first. 2 of the 6 declared inputs reach the Expensive edge on their own, and the Fair edge needs C$26.51M that no single input supplies. That is what ‘one-sided worse’ means in practice, and it is why the FULL band above should be read as a ceiling on conviction rather than a comfortable middle.

The relative lenses — and the three of four that could not run

Three of the four relative lenses are unavailable, and the weights are re-normalised over what is left rather than the missing lenses being silently scored neutral. The PEG lens does not apply to a company with no earnings. The own-history decile is withheld: a price-to-NAV decile needs at least two readings on one basis. There is ONE loosely comparable prior reading — the now-superseded 0.9103× of 11 September — and ZERO strictly comparable, because this run changed the basis again: the give-up is now after-tax at the PEA's own 0.746 USD/CAD rather than pre-tax at 1.39, and the receivable is discounted at this run's cost of equity with a conditionality haircut. The 11 September report said this arm would clear ‘at the next report, when two clean readings exist’. This is that report, and the condition is not met — one exists, not two. It clears from the report after this one.

And the third — the sector/peer median — was WITHDRAWN during this report's audit, which flattered the score. It had been carried at 40 on a 26.7% re-normalised weight, and no declared source anywhere carries a peer or sector median price-to-NAV; the only feed that could have is declared FAILED and is therefore non-citable. Withdrawing a lens that scored below the anchor raises the blend, from 44.70 to 46.40 — about 1.7 points. The band clamp absorbs it and the published score is unchanged, but the direction belongs here, where the number is published, and not only in the audit log.

The blend comes to 46.40, clamped into the Full band's 40–49 range at 46. The anchor is supreme — the relative lenses order names inside a band and can never lift one out of it.

Analyst cross-check

One analyst. Ventum Capital Markets initiated coverage on 4 August 2026 ^[4]; the C$1.00 target comes from the analyst panel on the data provider's snapshot ^[21], where mean, median, high and low are all C$1.00 on a single analyst. The price of C$0.70 is 30.0% below it. The BUY rating is not relied on and is withdrawn as a sourced claim: round 1 of this report's audit established that the release announcing the initiation explicitly disclaims both the rating and the target — they were carried by third-party wires, and no wire is a declared source here — while the provider returns a null recommendation. The lens is therefore scored on the target alone. On the scoring table that is an 85–100 read, marked down to 60 here because n=1 is not a consensus — there is no dispersion to read, no revision history, and the coverage is 52 days old. The FMP consensus, grades and ratings endpoints all return empty for this listing, so there is nothing to cross-check it against. Free-cash-flow yield is not applicable: there is no free cash flow.

Risk-reward, anchored to the stop

The stop sits at C$0.625, below the C$0.68 low that has now been tested twice. That is C$0.075 of risk, 10.7% of price and 1.98 ATR on an ATR of C$0.0379 — squarely in the 1.5–2.5 ATR ‘moderate’ band, so the position-risk sub-signal scores 55. Price is within 3% of that support, which earns the +15 proximity bonus and takes position-risk to 70. It is 7.9% below the C$0.76 three-month high, so no resistance penalty applies.

Relative strength — the strongest thing on this page

Over one month CCI.CN is +12.90% against SPY at +0.69% and the Materials sector ETF XLB at −7.21%; over three months +18.64% against +5.81% and −3.49% ^[17]. That is outperformance of +12.21 points against the index and +20.11 against its own sector on one month, which is the top scoring band at 90. The shares sit at the 79.7th percentile of a 52-week range of C$0.23 to C$0.82 — momentum, and extended. Blended with position-risk, the risk-reward sub-signal is 80.

Macro overlay — 20% weight, because Materials is a HIGH-sensitivity sector

Sentiment — 15% weight

One rating action in twelve months: Ventum Capital Markets initiated coverage on 4 August 2026, with the C$1.00 target coming from the provider's analyst panel ^{[21] [4]} — outside the 30-day window, so it reads as no recent change. The rating itself is not stated: the round-3 source auditor searched for a wire carrying a Buy and found none, and the release announcing the initiation explicitly disclaims both the rating and the target. An earlier draft here read ‘initiated at Buy’, which was the single NOT-SUPPORTED row in the claim census. There are no estimate revisions to track, because there are no estimates. News tone over the period is positive but thin: one release, the 21 September step-out, and its assays are pending ^[5]. Scored 58 — a neutral base lifted slightly by a positive release, not by a trend.

Catalyst calendar — 15% weight, and distinct from clustering

Scored 55 for how favourable the calendar is: the feasibility study and the pending assays could go either way, the EIA is likely positive, and Core PCE is forecast to rise into a hiking Fed, which is not. That is a different question from how crowded it is — clustering reads 18 and is handled in §8. Changing one does not move the other.

Upcoming events

Recent surprises (last 7 days)

This is a crowded fortnight and the catalyst clustering score is scored accordingly at 18 — chaotic. Four high-impact releases that specifically move a base-metals developer fall inside eleven days: Core PCE on 30 September, ISM Manufacturing on 1 October, Non-Farm Payrolls on 2 October and the FOMC minutes on 7 October ^[18]. The SKILL's table is scored as written — four or more inside fourteen days is the chaotic band — not softened by feel. The consequences are applied: position sizing is halved, timing confidence takes a 10-point penalty, and short-horizon signal confidence is capped at 35%. It does not move the timing score, because the score's catalyst term measures whether the calendar is favourable and clustering measures how crowded it is. Core PCE on 30 September ^[18] sits three trading days out, which is inside the window that would normally make a macro release a scheduling trigger for a high-sensitivity sector. It is deliberately not used as one — with five high-impact releases inside seventeen days, scheduling off the first would peg the next update perpetually three days out, and none of them resolves the feasibility study, the EIA or the pending assays, which are what actually move this name.

Intraday data is unavailable for this CSE listing ^[21], so the weights are reallocated to Monthly 35%, Weekly 30%, Daily 35% exactly as the framework prescribes for that case, and the two empty rows are shown rather than deleted so the gap is visible.

The structure is intact and the near-term tape is not. Price of C$0.70 sits +4.9% over a rising 50-day average of C$0.6676 and +12.5% over a 200-day of C$0.6221, and both averages are rising — that is a genuine uptrend, and the stock is up 18.6% over three months. But it is below its 20-day average of C$0.7160, it has printed three consecutive lower weekly closes (C$0.74 → C$0.72 → C$0.70), and the MACD histogram has been negative for eight consecutive sessions, since 16 September: −0.0006, −0.0008, −0.0010, −0.0013, −0.0003, −0.0011, −0.0024, −0.0039. Round 1 of this report's audit corrected the previous description of that series — it read ‘negative and widening for five sessions’, and both halves were wrong: the run is eight sessions, not five, and it NARROWED on 22 September before widening again. The conclusion the cap rests on is untouched, because the test is whether the histogram is positive for two consecutive days or turning up, and on 25 September it is neither. RSI at 50.6 is neutral. Volume on 25 September was 41,274 against a 20-day average of 177,012 — 0.23×, which is not a session that confirms anything in either direction.

This is the read that caps the short horizon. The cross-timeframe pattern is ‘higher-timeframe bullish with a lower-timeframe pullback’, which is normally the high-probability buy setup — but the setup requires the pullback to be tested, and C$0.68 has now been hit twice inside the pullback (16 and 23 September) without a higher low forming between them — and a third time on 10 September, just before it. A double bottom is not yet a higher low. The level that would change this is a close back above C$0.7160 with the histogram turning up, or a volume-backed close above the C$0.76 three-month high.

126 settled daily closes, 27 March to 25 September 2026, raw and unadjusted, with the 50-day average overlaid ^[17]. The pullback from C$0.76 and the twice-tested C$0.68 are the whole short-horizon argument.

The base case is the report's target price of C$0.79. Probability-weighted 12-month value C$0.780 (+11.4% against C$0.70), on weights of 22/55/23 summing to 100 with the base case the most probable. The bear is weighted only one point below the bull, which is deliberate: the December feasibility study is the dominant event and it is a two-sided capex read, while the basket at 1.5722× the deck has more room to mean-revert than to extend.

Forecast:

What this adds up to: the cheapest path back to a short BUY is the technical one, and it is close — one or two sessions above C$0.7160 with the histogram turning. The fundamental path is shut, because the shares trade at a premium to the deck-case NAV and a 35% construction discount.

Forecast:

No exit trigger is live. The stop at C$0.625 sits below the twice-tested C$0.68 with ATR room and is unlikely inside 4–6 weeks at the current trajectory — it is 10.7% below spot and below both the rising 50-day and the 200-day, so reaching it needs a thesis event rather than drift. The risk trigger to watch is the feasibility study, which has no date and is the largest single sensitivity in §4.

Not computed as a percentage, because no allocation was specified — the framework does not invent one. What can be said is what the framework's own levers do to it.

Risk per share — the part that can be computed without an allocation

Round 1 of this report's audit graded this section thin against a registered precedent (S13-SIZING-MATH), whose recorded fix was to compute the risk-per-share math or narrow the subtitle. Both are done: the table above is the math, and the subtitle no longer promises a portfolio percentage this report cannot honestly produce.

Volatility context. ATR of C$0.0379 is 5.41% of price — this stock moves more in an average day than the S&P 500 does in an average week. The 52-week range is C$0.23 to C$0.82 and the shares sit at the 79.7th percentile of it. The 20-day average volume of 177,012 shares is about C$124,000 a day, so liquidity, not conviction, is the binding constraint on position size here — and the 25 September session traded a quarter of that. Nothing in this report is advice, and a micro-cap on a CSE listing is not a core holding at any size.

{
  "ticker": "CCI.CN",
  "date": "2026-09-27",
  "version": "v6",
  "exchange_ticker": "CSE:CCI",
  "isin": "UNVERIFIED",
  "finder_ticker": "CCI",
  "finder_exchange": "CSE",
  "analysis_status": "donatien-pick",
  "lifecycle_stage": "pea-stage-developer-funded",
  "sector": "Materials \u2014 base metals (polymetallic VMS: zinc, copper, silver, lead)",
  "price_at_rating": 0.7,
  "signal_short": "HOLD",
  "signal_medium": "BUY",
  "signal_long": "BUY",
  "primary_signal": "BUY",
  "base_signal_short": "BUY",
  "base_signal_medium": "BUY",
  "base_signal_long": "BUY",
  "short_entry_confirmed": false,
  "short_hold_reason": "technical_pending",
  "short_cap_reason": "Short technical-confirmation cap FIRES. The Technical entry group requires the MACD histogram positive for two consecutive days or turning up off support. It has printed NEGATIVE for EIGHT consecutive sessions, from 16 September: -0.0006, -0.0008, -0.0010, -0.0013, -0.0003, -0.0011, -0.0024, -0.0039 to 25 September. CORRECTED IN ROUND 2: an earlier version of this field said 'steadily MORE negative for five sessions' and then quoted a series that narrows at step two - self-falsifying on its own numbers. The run is eight sessions, not five, and it NARROWED on 22 September before widening again. The test the cap turns on is unaffected: on 25 September the histogram is neither positive for two consecutive days nor turning up. Price also sits below its 20-day average of C$0.7160. The Catalyst group requires volume above 2x the 20-day average; the 25 Sep session traded 41,274 against a 20-day average of 177,012, i.e. 0.23x. Both groups UNMET, so signal_short is capped at HOLD - 'buy on confirmation'. This REVERSES the 11 Sep report's short_entry_confirmed=true, which rested on the 31 Aug volume-backed close above C$0.68: the group is tested on the CURRENT tape, and that breakout has faded through three consecutive lower weekly closes (C$0.74 -> C$0.72 -> C$0.70). It re-arms on a close back above C$0.7160 with the histogram turning up, or on a volume-backed close above the C$0.76 three-month high. Medium and Long are untouched by this cap. The half-size quality-starter override does NOT apply: it is scoped to a BASE signal of HOLD produced by Neutral timing, and the base here is BUY on Improving timing of 68.",
  "quality_score": 70,
  "valuation_score": 46,
  "timing_score": 68,
  "driver_score": 77,
  "driver_score_short": 68,
  "driver_score_medium": 76,
  "driver_score_long": 85,
  "relative_strength_vs_spy": 12.21,
  "relative_strength_vs_sector": 20.11,
  "usd_cad_rate": 1.4146,
  "usd_cad_rate_asof": "2026-09-25 (settled Friday close; the Sunday 27 Sep bar of 1.41410 is rejected as unsettled)",
  "pea_usd_cad": 1.3405,
  "fx_note": "THREE rates appear in this report and they do different jobs. CORRECTED IN ROUND 1: an earlier draft asserted the NAV makes EXACTLY ONE FX conversion, and that was false - the cash build also converts the US$2.0M of debt repaid on 11 August at 1.39, close to that day's spot, worth about C$0.05M, and declared nowhere. THE PRINCIPAL CONVERSION, the one that matters, takes the PEA's USD metal deck into CAD at the STUDY's own 0.746 USD/CAD = 1.3405, because the project NPV it is netted against was struck on that rate; a live rate there would mix two bases inside one subtraction, which is precisely the error this run corrects in the 11 September give-up. The market rate at the report date is 1.4146, the settled Friday 25 September close. THE GAP IS A TAILWIND THE NAV DOES NOT COUNT: a Canadian producer selling USD-priced metal earns more CAD per USD at 1.4146 than at the study's 1.3405, so holding the deck at the PEA's rate understates CAD revenue by 5.5%. That sits alongside the metal-price gap as un-priced upside, and it is disclosed in section 4 rather than capitalised into the NAV.",
  "quality_confidence": 55,
  "valuation_confidence": 45,
  "timing_confidence": 50,
  "overall_confidence": 45,
  "quality_detail": {
    "weighted_build": [
      {
        "component": "PEA project economics + distance to production",
        "weight": 0.2,
        "score": 75
      },
      {
        "component": "Funding / runway",
        "weight": 0.2,
        "score": 72
      },
      {
        "component": "Balance-sheet health",
        "weight": 0.1,
        "score": 80
      },
      {
        "component": "Competitive moat (5-dimension average)",
        "weight": 0.2,
        "score": 54
      },
      {
        "component": "Capital allocation + skin in the game",
        "weight": 0.2,
        "score": 72.5
      },
      {
        "component": "Permitting / execution progress",
        "weight": 0.1,
        "score": 68
      }
    ],
    "weighted_build_sum": 69.5,
    "moat_score": 54,
    "moat_build": [
      {
        "dimension": "Pricing power",
        "score": 20,
        "basis": "concentrate seller into LME-priced markets - a pure price-taker"
      },
      {
        "dimension": "Network effects",
        "score": 50,
        "basis": "not applicable to a mine; scored neutral per SKILL rather than zero"
      },
      {
        "dimension": "Switching costs",
        "score": 50,
        "basis": "concentrate is fungible; the Ocean Partners offtake creates some stickiness"
      },
      {
        "dimension": "Cost advantage",
        "score": 80,
        "basis": "an existing, permitted, past-producing 3,000 tpd plant and permitted tailings 13 km from the pit; PEA AISC US$3.14/lb CuEq"
      },
      {
        "dimension": "Intangible assets",
        "score": 70,
        "basis": "the plant and tailings permits themselves - the hardest thing to replicate in the Bathurst camp"
      }
    ],
    "industry_benchmark_name": "Mining (pre-production): PEA NPV/IRR + AISC margin + distance to production",
    "industry_benchmark_value": "NPV7 after-tax C$171M / 36% IRR / C$64M initial capex / 13.2yr LOM (Ausenco PEA, 22 May 2025); AISC US$3.14/lb CuEq against a spot basket of US$6.67/lb CuEq = a 52.9% AISC margin",
    "industry_benchmark_score": 75
  },
  "moat_score": 54,
  "industry_benchmark_name": "Mining (pre-production): PEA NPV/IRR + AISC margin + distance to production",
  "industry_benchmark_score": 75,
  "timing_detail": {
    "weighted_build": [
      {
        "component": "Multi-timeframe trend",
        "weight": 0.3,
        "score": 78
      },
      {
        "component": "Risk-reward + relative strength",
        "weight": 0.2,
        "score": 80
      },
      {
        "component": "Macro regime overlay (Materials = HIGH sensitivity)",
        "weight": 0.2,
        "score": 58
      },
      {
        "component": "Sentiment",
        "weight": 0.15,
        "score": 58
      },
      {
        "component": "Catalyst calendar favourability",
        "weight": 0.15,
        "score": 55
      }
    ],
    "mtf_confluence": 78,
    "risk_reward_score": 80,
    "relative_strength_vs_spy": 12.21,
    "relative_strength_vs_sector": 20.11,
    "catalyst_clustering_score": 18,
    "catalyst_clustering_basis": "CHAOTIC (0-19): four high-impact US releases that specifically move a base-metals developer fall inside eleven days - Core PCE 30 Sep (forecast 3.4% against 3.3% prior), ISM Manufacturing PMI 1 Oct, Non-Farm Payrolls 2 Oct and the FOMC minutes 7 Oct, with the Fed having just hiked 25bp on a unanimous vote and 16 of 18 participants expecting more. The SKILL's table is scored as written, not by feel: 4+ inside fourteen days is the chaotic band. Consequences applied: position sizing x0.5, timing confidence -10, and short-horizon signal confidence capped at 35%. This does NOT move timing_score - the score's catalyst term measures whether the calendar is FAVOURABLE (55), clustering measures how CROWDED it is, and the two are independent.",
    "dynamic_macro_weight": 0.2,
    "atr14": 0.0379,
    "atr_pct_of_price": 5.41,
    "rsi14": 50.6,
    "macd_hist": -0.0039,
    "macd_hist_prior": -0.0024,
    "sma20": 0.716,
    "sma50": 0.6676,
    "sma200": 0.6221,
    "vol_last": 41274,
    "vol_20d_avg": 177012,
    "vol_ratio": 0.23
  },
  "driver_detail": {
    "weighted_build": [
      {
        "horizon_input": "Historical (24m basket vs the PEA deck)",
        "weight": 0.25,
        "score": 85
      },
      {
        "horizon_input": "Current (basket level + AISC margin)",
        "weight": 0.5,
        "score": 82
      },
      {
        "horizon_input": "Forward (macro + futures + Fed path)",
        "weight": 0.25,
        "score": 58
      }
    ],
    "per_horizon": {
      "short": 68,
      "medium": 76,
      "long": 85
    },
    "deck": {
      "cu": 4.25,
      "zn": 1.3,
      "ag": 27.0,
      "pb": 1.1
    },
    "spot": {
      "cu": 6.766,
      "zn": 1.8416,
      "ag": 64.801,
      "pb": 0.863
    },
    "annual_production": {
      "cu": 8.0,
      "zn": 47.0,
      "ag": 0.783,
      "pb": 10.0
    },
    "deck_revenue_usd_m": 127.241,
    "spot_revenue_usd_m": 200.051,
    "deck_revenue_share": {
      "cu": 26.7,
      "zn": 48.0,
      "ag": 16.6,
      "pb": 8.6
    },
    "spot_revenue_share": {
      "cu": 27.1,
      "zn": 43.3,
      "ag": 25.4,
      "pb": 4.3
    }
  },
  "driver_commodity_trend": "Re-derived this run from settled Friday 25 September 2026 closes; nothing on this line is carried. The basket is a ZINC project first: at the PEA deck the revenue split is Zn 48.0% / Cu 26.7% / Ag 16.6% / Pb 8.6%. COPPER HG=F US$6.766/lb - +3.5% against a RISING 50-day average of US$6.537, +11.1% against the 200-day, 4-week +2.7%, 8-week +5.0%; 1.59x the PEA's US$4.25 deck. ZINC LME cash settlement US$4,060.00/t = US$1.8416/lb, 1.42x the US$1.30 deck; the monthly cash average has risen July ~US$3,600 -> August US$3,875 -> September US$4,029, and 25 Sep prints above all three. SILVER SI=F US$64.801/oz, 2.40x the US$27 deck - above a RISING 50-day average (+1.8%) but 10.8% BELOW its 200-day, with 4-week momentum NEGATIVE at -6.7% against 8-week +10.2%. LEAD LME cash US$1,902.50/t = US$0.8630/lb is the one metal BELOW deck at 0.78x, though its September cash average rose against August. Basket-weighted at deck revenue shares: 4-week momentum -0.4% (flat, dragged by silver), 8-week +7.9%. NO metal sits below a falling 50-day average, so the SKILL Step-2b short-horizon downtrend cap does NOT fire - a marginally negative 4-week window against a strongly positive 8-week window, with all four metals above rising 50-day averages, is not a live downtrend. The weighted spot basket is 1.5722x the PEA deck. Macro 2026-09-20: Copper / Industrial Metals N / N / O (the short leg upgraded from Underperform since the 9 Sep run, the long leg downgraded from Strong Outperform); Materials XLB O / O / SO.",
  "economic_alignment_stance": "Trend-Following",
  "economic_alignment_conviction": 72,
  "economic_alignment_pressure": "Tailwind",
  "economic_alignment_source": "sector-map",
  "macro_report_date": "2026-09-20",
  "val_band": "full",
  "actual_multiple": 0.8914,
  "warranted_multiple": 0.65,
  "warranted_ratio": 1.37,
  "val_multiple_basis": "forward P/NAV. NAV built bottom-up from the Ausenco PEA of 22 May 2025 held at its OWN base-case deck (Cu US$4.25/lb, Zn US$1.30/lb, Ag US$27/oz, Pb US$1.10/lb, 0.746 USD/CAD) per the SKILL's miner rule, which is forward by construction - the NPV is the present value of cash flows not yet earned, and no trailing P/NAV exists for a pre-revenue developer, so there is no basis to mix. Warranted 0.65x = a 35%% pre-production construction discount, the middle-to-generous part of the framework's 30-50%% band.",
  "warranted_basis": "35% construction discount => warranted P/NAV 0.65x. REWRITTEN IN ROUND 1 on two auditor findings, and the position in the band is unchanged at 35%. WHAT THE EARLIER DRAFT GOT WRONG: it declared funding risk 'RESOLVED' while separately listing the EIA as an open leg. Those are not independent. Per the 14 April 2026 financing release the C$31.5M construction tranche - 73% of the OR money still to come - is conditional on 'obtaining all remaining required permits for the Bathurst Complex development', on Board approval of construction, on material agreements for execution, on full funding to completion being available at each instalment, and on a financial model supporting a minimum debt-service coverage ratio. Counting the funding as resolved while counting permitting as open credited one leg twice. The earlier draft also rested on a market cross-check - that funded and permitted developers trade around 0.6-0.9x NAV against 0.3-0.6x for greenfield names - which carried NO citation. It was the only external validation of the deciding input in the whole report, and it is WITHDRAWN rather than softened. THE CRITERION AS IT NOW STANDS, on three legs rather than four: (1) FULLY RESOLVED, and unconditionally - the processing plant, its permits, the tailings facility and the title. The Caribou mill is a permitted, past-producing 3,000 tpd plant 13 km from the pit and title transferred on 28 July 2026. For a brownfield restart this is the single largest de-risking factor available, and it is why initial capital is C$64M rather than the several hundred million a greenfield mine on this deposit would need. It is the ONLY leg doing work at the generous end of the band. (2) PARTIALLY RESOLVED - funding. Committed at roughly 150% of capex, C$12.5M of it received unconditionally on 16 July 2026, but the C$31.5M tranche is conditional as set out above and the C$48M Ocean Partners facility is undrawn. Committed is not drawn, and conditional is not committed. (3) OPEN - there is no bankable study (the PEA includes inferred material; the DRA Americas feasibility study is due December 2026 and restates the capex) and the Murray Brook EIA registered 13 July 2026 is still in Technical Review Committee review, the first mine EIA in New Brunswick since 2013, with no determination. One dominant leg unconditionally resolved, one partially, two open, puts this a quarter of the way into a 30-50% band - which is 35%, not the 30% most-generous end the 11 September report used. THE CONSEQUENCE, STATED PLAINLY BECAUSE IT IS NOW NARROW: at the band MIDPOINT of 40% the ratio is 1.4856x, which fires Gate 3, caps all three horizons at HOLD, and would auto-Stop this Donatien Pick. On the base case the name sits only C$3.80M of NAV inside the 1.40x edge.",
  "val_lens_weights_renormalised": [
    {
      "lens": "anchor (actual / warranted P/NAV)",
      "nominal_weight": 0.4,
      "renormalised_weight": 0.7273,
      "score": 41.3
    },
    {
      "lens": "analyst consensus",
      "nominal_weight": 0.15,
      "renormalised_weight": 0.2727,
      "score": 60
    }
  ],
  "val_lens_blend": 46.4,
  "val_decile_na_reason": "The own-history decile lens is WITHHELD, and this is recorded as a check that could not run rather than a check that passed. A P/NAV decile needs at least two prior readings on the SAME basis. There is ONE loosely comparable prior reading (the now-superseded 0.9103x of 11 Sep 2026) and ZERO strictly comparable ones, because this run again changed the basis: the OR Royalties give-up is now valued at the PEA's own 0.746 USD/CAD and AFTER tax rather than pre-tax at 1.39, and the consideration receivable is now discounted at this run's r of 11.684% over a two-year average draw with a 10% conditionality haircut. The three readings before 11 Sep sit on a basis that report corrected. The 11 Sep report recorded this arm as un-clearable and said it would clear 'at the next report, when two clean readings exist'; that condition is NOT met - one exists, not two. It becomes clearable from the report after this one.",
  "val_guardrail_na_reason": "Not N/A - the guardrail RAN and cleared. The Materials line is P/NAV >= 1.5x and the actual multiple is 0.8914x, so the guardrail arm does not fire with room to spare. ON THE TIGHTENING RULE, CORRECTED IN ROUND 1: the 10-year IS above 5% - it settled at 5.184% on Friday 25 September, having run 4.963 -> 4.968 -> 5.114 -> 5.162 -> 5.184 across the week, and an earlier draft of this report used the 20 September macro report's 4.998% stamp and stated flatly that the 10-year was NOT above 5%. That was false on this report's own stamp date. The rule tightens the line one notch per ~1.5pp above 5%, and 5.184% is 0.184pp above it, so the correct number of notches is ZERO and the unadjusted 1.5x line is still the one used. The outcome is unchanged; the stated reason was wrong and is now right.",
  "eps_trailing": null,
  "trailing_pe": null,
  "eps_forward": null,
  "clean_pe": null,
  "clean_peg": null,
  "nonop_pct_of_net_income": null,
  "eps_pe_na_reason": "Pre-revenue. There is no revenue, no operating income and no meaningful EPS, so every earnings-anchored field is null BY CONSTRUCTION, not for want of retrieval: trailing EPS prints -C$0.02 on the data provider against zero sales, and a P/E on it is not a number a reader should be shown. The currency-coherence check (CCY-PE) therefore cannot run on an earnings multiple; it is replaced here by an explicit currency reconciliation of the P/NAV anchor, whose numerator (price C$0.70) and denominator (NAV C$0.7853/share) are BOTH in CAD, with the USD metal deck converted once at the PEA's own 0.746 USD/CAD. These three fields are RE-DERIVED to null every run and never carried - a carried numeric is invisible to stale_text, which is how a wrong clean_pe survived Layer 0 on WCP.TO.",
  "net_debt_ebitda": null,
  "net_debt": -21.2,
  "net_debt_ebitda_na_reason": "Net debt is NEGATIVE - C$21.20M of cash against ZERO drawn debt, the US$2.0M facility having been eliminated on 11 August 2026 - and EBITDA is negative on no revenue, so the ratio has no meaningful value and is emitted null rather than fabricated. Do-Not-Buy Trigger 1 is adjudicated on the underlying facts instead, in dnb_adjudication: its first arm (net debt / EBITDA > 4x) fails outright on negative net debt, so the trigger cannot fire however hawkish the Fed is. The arm to watch is the C$48M Ocean Partners facility at SOFR+7.75%, which becomes live the moment it is drawn for construction in 2027.",
  "nav_per_share": 0.7853,
  "nav_total": 185.6576,
  "nav_basis": "PEA project NPV7% after-tax C$171.0M (held at the study's own base-case deck; the C$64M initial capex is already NETTED inside it) + C$22.73M present value of the OR Royalties consideration still receivable (C$31.5M drawable across construction, discounted at r = 11.684% over a 2.0-year average draw and haircut 10% for milestone conditionality) - C$23.30M present value of the OR stream give-up (20% of payable silver sold at 20% of spot = 16% of silver revenue forgone; silver revenue C$28.34M/yr at deck and the PEA's 0.746 USD/CAD, taxed at the 29% combined federal-plus-New-Brunswick statutory rate, over a 13.2-year mine life deferred 2.27 years to first production, a 7.236x marginal multiple) - C$11.90M present value of corporate costs still to be incurred before first production (C$4.4M/yr of G&A plus C$3.5M of owner's costs) + C$21.20M net cash + C$5.92M warrant exercise proceeds = C$185.66M. Divided by 236,418,204 fully diluted shares = C$0.7853 per share. Actual P/NAV C$0.70 / C$0.7853 = 0.8914x; divided by the warranted 0.65x = 1.3714x.",
  "valuation_model": {
    "basis": "P/NAV",
    "unit": "CAD_m",
    "inputs": [
      {
        "id": "financing_basis",
        "label": "OR financing basis - credit the surplus, or treat the pair as value-neutral",
        "basis": "The consideration (C$29.75M PV) and the give-up (C$23.30M at the study's US$27/oz silver) are priced on different decks, so the pair books C$+6.45M. The range is the accounting choice itself: fully credited at the top, exactly value-neutral at the bottom. It subsumes the narrower question of how the RECEIVABLE is discounted, because that range (conditionality 0.75-1.00 over a 1.5-2.5 year draw) sits inside it. It does NOT cover how the GIVE-UP is discounted - that is a separate declared input, giveup_discount_basis, added in round 1 after the auditor found the earlier draft implying this entry covered both.",
        "nav_low": 179.2044,
        "nav_high": 185.6576
      },
      {
        "id": "giveup_discount_basis",
        "label": "discount rate applied to the stream give-up",
        "basis": "ADDED IN ROUND 1, and it is the finding that moved this report most. The give-up reduces the very cash flows stage1_npv discounts at the study's 7%, so it is now valued at 7% (multiple 7.2361x) rather than at the 11.684% cost of equity (multiple 5.1111x) an earlier draft used - a mismatch worth C$6.84M of NAV in the FLATTERING direction, and the same class of error as the FX basis this run corrects in the 11 September give-up. The range spans the two bases: 7% at the conservative end, the equity rate at the flattering end. It is declared as its own input because the earlier draft's financing_basis entry claimed to subsume only how the RECEIVABLE is discounted, which left this uncovered.",
        "nav_low": 185.6576,
        "nav_high": 192.4985
      },
      {
        "id": "pea_capex_revision",
        "label": "capex revision at the December 2026 feasibility study",
        "basis": "The C$64M is a PEA-level estimate, an accuracy class typically quoted at plus or minus 35-50%%, and the DRA Americas feasibility study due December 2026 restates it. The range runs from no revision to plus 30%%, applied as a direct reduction of the after-tax NPV because the capital sits at the front of the schedule. A downward revision is possible on a brownfield restart and is deliberately not modelled, which makes this input one-sided against the name.",
        "nav_low": 166.4576,
        "nav_high": 185.6576
      },
      {
        "id": "corporate_costs",
        "label": "corporate overhead and owner's costs to first production",
        "basis": "G&A runs about C$4.4M a year on the reported quarterly loss, and owner's costs - completing the feasibility study, permitting, pre-construction engineering - are estimated rather than disclosed. The range spans C$2.0-6.0M of owner's cost against a first-production date half a year either side of the end-2028 target, since a slip lengthens the period being funded.",
        "nav_low": 181.95,
        "nav_high": 188.68
      },
      {
        "id": "net_cash",
        "label": "net cash at the report date",
        "basis": "Built forward from the disclosed C$17.3M at 30 April 2026 plus the C$12.5M OR receipt of 16 July and the C$3.18M August warrant exercise, less the US$2.0M debt repaid. The uncertainty is the five months of drilling, feasibility-study and overhead spend, estimated at C$7-11M, because the interim statements for the quarter ended 31 July 2026 were not available at the report date. This is an estimate and is labelled one.",
        "nav_low": 183.6576,
        "nav_high": 187.6576
      },
      {
        "id": "giveup_tax_rate",
        "label": "marginal tax rate applied to the stream give-up",
        "basis": "The give-up is charged after tax because it is netted against an after-tax NPV. The central 29%% is the combined federal-plus-New-Brunswick statutory corporate rate; the range runs 20-35%% to span a lower effective rate while capital shields are unwound and a higher one once the New Brunswick Metallic Minerals Tax is included. It is the smallest of the five inputs and is listed so the count is complete.",
        "nav_low": 182.7047,
        "nav_high": 187.6263
      }
    ],
    "basis_asymmetry_note": "DECLARED, not silent. The two sides of the OR Royalties transaction are priced on different metal decks and the gap is large: the consideration is C$29.75M in present value (C$7.02M received on 16 July, inside net cash, plus C$22.73M still receivable) against a give-up of only C$23.30M when silver is held at the study's US$27/oz. The pair therefore books C$+6.45M, which is 22% of the larger leg - far from the value-neutral result an arm's-length financing should produce. DIRECTION AND CAUSE: the asymmetry is not an error in the model, it is the difference between the PEA's silver deck and the market's. Solving for the silver price at which the pair is exactly neutral gives US$34.48/oz - 53% of the 25 September spot of US$64.80 [[src:px-2509]], which is precisely how a streamer prices a long-term deck. OR did not overpay; the study's US$27 is simply stale. WHY THE SURPLUS IS NEVERTHELESS KEPT: it is real on a deck-consistent NAV, because the company genuinely sold silver into a market that values it above the deck. And the inconsistency cuts the CONSERVATIVE way overall - if the market's silver view were applied to the PROJECT's silver revenue as well as to the stream, the NPV would rise by far more than the give-up would, because the project keeps 84% of payable silver and has sold 16%. WHAT IT COSTS: on a value-neutral treatment NAV falls to C$179.20M, NAV per share to C$0.7580 and the ratio rises to 1.4207x - PAST the 1.40x Expensive edge, firing Gate 3, capping all three horizons at HOLD and auto-Stopping the Donatien Pick. That is the single most important number in this report and it is one defensible accounting choice away from the published signal.",
    "terms": [
      {
        "id": "stage1_npv",
        "label": "PEA project NPV7% after tax",
        "sign": "+",
        "kind": "asset_pv",
        "value": 171.0,
        "nets": [
          "initial_capex",
          "tax"
        ],
        "means": "the present value of the Murray Brook open pit processed at the Caribou plant, at the study's own base-case metal deck, already subtracting the sixty-four million dollars of initial capital as a future outflow and stated after tax"
      },
      {
        "id": "or_consideration",
        "label": "OR Royalties consideration still receivable",
        "sign": "+",
        "kind": "asset_pv",
        "value": 22.73,
        "consideration_for": "or_stream",
        "means": "the present value of the thirty-one and a half million dollars of stream deposit still drawable by the company across construction, being the cash side of the same transaction whose cost is carried by the give-up term below; discounted at this run's cost of equity over a two-year average draw and haircut ten per cent because each instalment is conditional on construction milestones"
      },
      {
        "id": "or_stream",
        "label": "OR Royalties 20% payable silver stream give-up",
        "sign": "-",
        "kind": "giveup",
        "value": 23.3,
        "consideration_elsewhere_value": 7.02,
        "means": "the value the company permanently forgoes by selling twenty per cent of payable silver at twenty per cent of spot, so eighty per cent of that twenty per cent is given up; valued on a marginal after-tax annuity over the mine life deferred to first production, not on the average capex-inclusive multiple that applies to the project as a whole. The PEA schedules no payable GOLD, so although the instrument covers silver and gold this term is sized on silver alone and the gold leg is valued at zero, which is a disclosed omission rather than an oversight"
      },
      {
        "id": "corp_ga",
        "label": "Corporate costs still to be incurred",
        "sign": "-",
        "kind": "outflow",
        "value": 11.9,
        "nets": [
          "corporate_ga"
        ],
        "means": "head-office costs and owner's costs that the project-level study does not carry, counted only from today FORWARD to first production so that money already spent is not deducted twice - the spend to date is already out of the cash balance below"
      },
      {
        "id": "net_cash",
        "label": "Corporate net cash",
        "sign": "+",
        "kind": "cash",
        "value": 21.2,
        "means": "cash at the corporate level at the report date, built from the disclosed seventeen point three million at thirty April, plus the twelve and a half million received from OR Royalties on sixteen July, plus the three point one eight million August warrant exercise, less the two million US dollars of debt repaid, less roughly nine million of outflow of which the largest single item is the six million dollar balance of the Caribou purchase price and the remainder is five months of drilling, study and overhead spend. There is no drawn debt to net against it"
      },
      {
        "id": "warrant_proceeds",
        "label": "Warrant exercise proceeds",
        "sign": "+",
        "kind": "dilution_proceeds",
        "value": 5.92,
        "funds": [
          "shares"
        ],
        "means": "the cash arriving if the twenty-three million six hundred and eighty-five thousand two hundred warrants at twenty-five cents are exercised before they expire on twelve November; credited ONLY because the per-share figure divides by the fully diluted count, so the shares and their proceeds enter together"
      }
    ],
    "no_double_count": "Checked pair by pair. (1) stage1_npv NETS the initial capex as a future outflow and NO term adds it back; the capex is unspent, so unlike a spent-capex case it is correctly not restored. (2) or_consideration and or_stream are the two sides of ONE transaction and each is carried exactly once - crediting the consideration does not double-count the capex, because a project NPV nets a USE of funds and credits no SOURCE, so the consideration is an inflow the NPV omits whose offsetting cost is the give-up already deducted. This was established by cash-flow trace on the 11 September report and is unchanged. (3) warrant_proceeds is credited only against the fully diluted denominator. (4) THE DOUBLE COUNT THIS RUN ACTUALLY MADE AND FIXED: the first build of net_cash added WARR_PROCEEDS (C$5.92M, the 23,685,200 warrants still outstanding) into the cash line as well as carrying it as its own term, overstating NAV by C$5.92M. The cash line now carries the C$3.18M August exercise, which is a DIFFERENT and already-received amount, and the model asserts the two are unequal at import time. (5) corp_ga counts only FORWARD costs; the drilling, study and overhead spend since 30 April is already deducted inside net_cash, so deducting it again in corp_ga would be the same dollar twice. (6) Ocean Partners' C$48M facility is EXCLUDED entirely: it is undrawn, and its proceeds would be offset by repayment plus SOFR+7.75% interest, so crediting it without the servicing cost would be a one-sided entry."
  },
  "shares_basic": 212733004,
  "shares_fully_diluted": 236418204,
  "shares_basis": "DERIVED from a primary filing, not taken from a data provider. The early warning report of 13 August 2026 states that Ocean Partners held 45,397,223 common shares representing 21.34% of the issued and outstanding after exercising 12,725,000 warrants; 45,397,223 / 0.2134 = 212,733,004. The independent cross-check from the same filing agrees: 32,672,223 / 0.1633 = 200,075,462 before the exercise, plus 12,725,000 = 212,800,462, a 0.03% spread that is percentage rounding. THE DATA PROVIDER IS STALE AND WOULD HAVE BEEN WRONG: Yahoo's market capitalisation of C$140.26M at C$0.70 implies 200.4M shares, which is precisely the PRE-11-August count - it has not picked up the warrant exercise. Using it would have overstated NAV per share by 6%. Fully diluted adds the 23,685,200 warrants at C$0.25 from the 13 November 2025 financing, confirmed outstanding in the 11 August 2026 release.",
  "economic_study": "Ausenco Preliminary Economic Assessment ('Combined Strategy'), 22 May 2025 - after-tax NPV7%% C$171M, IRR 36%%, payback 2 years, initial capex C$64M, NPV/capex 2.7x, 13.2-year mine life, 3,300 tpd throughput through the 3,000 tpd Caribou plant with de-bottlenecking, strip ratio 5.0:1, AISC US$3.14/lb CuEq, deck Cu US$4.25/lb / Zn US$1.30/lb / Ag US$27/oz / Pb US$1.10/lb at 0.746 USD/CAD (Scotiabank long-term forecast of 14 May 2025). RE-CONFIRMED this run against the company's own release, which titles the figure C$171M. Note the company's website summary tile reads C$169M for the same study; the release title and body are the primary source and C$171M is used throughout. A DRA Americas feasibility study is under way and due December 2026 - the study ladder is PEA-stage, not FS-stage, and that is what withholds the top of the construction-discount band.",
  "gates_triggered": [],
  "gates_caution": [
    "Gate 1 Financial Distress - CAUTION, not triggered. Net debt is negative (C$21.2M cash, zero drawn debt) and interest coverage is not applicable with no interest. The 'negative FCF for 4+ quarters' arm is TRUE on its face and always will be for a developer, but the arm's own wording requires 'no clear path to positive', and C$96M committed against a C$64M capex with an FS due December 2026 and production targeted 2028 is a path. THE GOING-CONCERN ARM IS NOT CLEARED, IT IS UNVERIFIED: I did not retrieve the auditor's note, because the interim statements for the quarter ended 31 July 2026 were not available at the report date. Recorded as open rather than passed.",
    "Gate 3 Valuation Ceiling - CAUTION, not triggered, and all four arms are adjudicated rather than three. (a) Expensive band at >= 1.40x warranted: does NOT fire at 1.3714x. On the model as built NAV would have to fall C$3.80M (-2.0%) to trip it, and a combined conservative reading of the 6 declared uncertain inputs swings C$34.31M, so this edge is reachable several times over - see threshold_fragility, which grades the band NOT ROBUST and one-sided worse. CORRECTED IN ROUND 1: this sentence previously read 'four uncertain inputs' swinging 'C$12.17M' - both the count and the magnitude were stale, in the one sentence that adjudicates whether the edge is reachable. (b) Materials sector guardrail at P/NAV >= 1.5x: does NOT fire at 0.8914x. The 10-year IS above 5% - it settled at 5.184% on 25 September - but the rule tightens the line one notch per ~1.5pp above 5% and 0.184pp earns ZERO notches, so the unadjusted 1.5x line is the one used. An earlier draft of this sentence said the 10-year was not above 5% at all, on the 20 September macro stamp of 4.998%, and that was false on this report's own date. (c) Price above the highest analyst target: does NOT fire - C$0.70 against a sole target of C$1.00. (d) Own-history decile: CANNOT BE CLEARED and is recorded as such, with the count - one loosely comparable prior reading, zero strictly comparable. A withheld lens must not silently clear a gate arm.",
    "Gate 4 Dilution / Accounting - CAUTION, not triggered, on a stated carve-out rather than a self-exemption. The share-count arm fires on its letter: 200.1M before 11 August against 212,733,004 now, with 23,685,200 warrants (+11.1% fully diluted) still to come. It is recorded as a caution because the dilution is ALREADY INSIDE THE ANCHOR - the NAV divides by the fully diluted count and credits the C$5.92M of proceeds, so firing a gate on top would charge the same fact twice. That reasoning is a lifecycle carve-out the framework does not yet contain, so it is filed to the improvement-debt register as GATE4-DEV-LIFECYCLE with this report as the dated instance, rather than being invented again next run.",
    "Gate 5 Regulatory / Binary - CAUTION, not triggered. The Murray Brook EIA registered 13 July 2026 is in review and a refusal would be existential for the timeline, which clears the >20%% move test. It is a caution rather than a trigger because the gate requires an outcome that is 'genuinely binary - could go either way': this is a brownfield restart of a past-producing plant whose own permits are already in hand, in a province that hosted a formal government event for the title transfer. Approval is the base case, so the outcome is not 50/50. No FDA, antitrust or takeover arm applies.",
    "Debt service, prospective - the C$48M Ocean Partners tranche is a 36-month facility at SOFR+7.75%% with a 12-month grace period, a short tenor against a mine build, and it is FLOATING into a Fed that has just hiked 25bp to 3.75-4.00%% on a unanimous vote with 16 of 18 participants expecting at least one more. It is UNDRAWN today, so there is no debt service now; this is the gate that arms when construction draws it in 2027.",
    "Single asset, pre-revenue - one complex, no operating cash flow, CSE-listed. 55%% of the C$96M package is optional at the Company's discretion and OR's construction instalments are conditional on five separate milestones. Committed is not the same as drawn."
  ],
  "hard_gate_state": "caution",
  "do_not_buy_triggers": [],
  "dnb_adjudication": "All five triggers considered; none fires. Trigger 1 LEVERAGE + RISING RATES: rates ARE rising (the Fed hiked 25bp to 3.75-4.00% on a unanimous 12-0 vote, 16 of 18 expecting more) and the Ocean facility IS floating at SOFR+7.75%, but the first arm requires net debt / EBITDA > 4x and net debt is NEGATIVE at C$21.2M of cash against zero drawn debt, so the conjunction fails outright. Re-arms when the facility is drawn in 2027. Trigger 2 VALUATION EXTREME, absolute arm (a) deep-expensive: does NOT fire - 1.3714x against the 2.0x threshold, and actual P/NAV 0.8914x against 1.5x the Materials guardrail line, which is 2.25x. Arm (b) expensive plus a live de-rating catalyst: does NOT fire, because it requires the Expensive band (>= 1.40x) and the ratio is 1.3714x; the question of a catalyst therefore does not arise. On cohort membership for the armed systemic tail - the 20 September macro report carries 'S&P 500 concentration / AI earnings-quality unwind' as armed - this name is NOT in that cohort and the tail is not inherited: it is a CSE-listed base-metals developer with no AI capex or monetisation leverage, no index weight, and no reported earnings to be inflated by non-operating gains. 'It is a small-cap' is not cohort membership. Trigger 2 RELATIVE arm: CANNOT BE ADJUDICATED on the decile, which is withheld (see val_decile_na_reason); its second condition, forward growth <= trailing growth, is not applicable pre-revenue. Recorded as could-not-run rather than cleared. Trigger 3 PERSISTENT NEGATIVE EARNINGS REVISIONS: does not fire - there are no earnings and no consensus EPS series to revise, and the only rating action in the past twelve months is Ventum Capital Markets INITIATING COVERAGE on 4 August 2026, with a C$1.00 target that comes from the provider's analyst panel rather than from the release. The RATING is deliberately not stated: the round-3 source auditor searched for a wire carrying a Buy and found none, and the release announcing the initiation disclaims both the rating and the target. An earlier version of this field said 'at Buy', which was the census's only NOT-SUPPORTED row. Trigger 4 INSIDER SELLING SPIKE: does not fire, on affirmative evidence in the opposite direction - Ocean Partners, a 21.34% strategic holder, EARLY-exercised 12,725,000 warrants for C$3,181,250 of cash on 11 August 2026, which is a dated primary-sourced purchase. A named search for director or officer dispositions on SEDI and the insider trackers over August-September 2026 returned nothing; it is recorded in the section 15 Data Sources audit as src:sedi-none, which is where an OPEN-WORLD absence belongs, and it is a search that found nothing rather than proof that nothing exists. CORRECTED IN ROUND 1: an earlier draft of this field said the search was recorded in absence_searches[], and it was not - that array is empty by design, because it is for FETCHABLE absences which verify_absence re-runs against a named artifact. Asserting a record that did not exist is the worst class of defect an audit field can carry, and it also left the deterministic guard reporting zero declared searches. Trigger 5 STRUCTURAL BUSINESS MODEL THREAT: does not fire - the copper and zinc demand case is intact, the project is a brownfield restart in a supportive jurisdiction, and no regulatory ban, patent or technological displacement applies to a concentrate producer.",
  "competitive_share_trajectory": "gaining",
  "competitive_threat_level": "moderate",
  "fcf_yield": null,
  "analyst_consensus_target": 1.0,
  "analyst_target_high": 1.0,
  "analyst_target_low": 1.0,
  "analyst_target_upside_pct": 42.9,
  "analyst_coverage_count": 1,
  "analyst_grades_consensus": "Buy (n=1)",
  "analyst_bullish_pct": 100.0,
  "recent_upgrades_30d": 0,
  "recent_downgrades_30d": 0,
  "fmp_rating": null,
  "fmp_overall_score": null,
  "fair_value_est": 0.5104,
  "stop_loss": 0.625,
  "target_price": 0.79,
  "scenario_base_target": 0.79,
  "scenario_bull_target": 1.1,
  "scenario_bear_target": 0.45,
  "scenario_probs": {
    "bull": 22,
    "base": 55,
    "bear": 23
  },
  "probability_weighted_target": 0.78,
  "entry_groups_met": 0,
  "entry_conviction": "Wait",
  "exit_groups_live": 0,
  "exit_action": "Hold",
  "risk_reward_to_base": 1.2,
  "risk_reward_to_bull": 5.33,
  "discount_rate_r": 11.684,
  "risk_free_10y": 5.184,
  "equity_risk_premium": 4.5,
  "risk_addon": 2.0,
  "g_near": null,
  "g_term": null,
  "g_na_reason": "A two-stage earnings DCF is not the anchor here. The miner rule instantiates the anchor as P/NAV off a published study NPV, so there is no g_near or g_term to haircut - the growth path is the PEA's own production schedule. r still enters, and it enters twice: discounting the OR consideration receivable and the corporate-cost outflow.",
  "implied_deck_scale": 1.5722,
  "amplification_note": "Stage-2 amplification would otherwise lift the base BUY to STRONG BUY on medium and long: driver_score 77 sits in the 65-79 Tailwind band and economic pressure is Tailwind, so BOTH arms are met. It is blocked by the VALUATION arm alone - a Full- or Expensive-band name cannot be amplified, and the warranted ratio is 1.3714x against the 1.20x limit. Recorded explicitly because a consumer reading the JSON would otherwise see driver 77 plus Tailwind plus BUY and read a contradiction. The block is the band, not a per-horizon driver difference.",
  "next_update_date": "2026-09-30",
  "next_update_basis": "Q3 interim financial statements for the quarter ended 31 July 2026, due Tuesday 29 September 2026, plus one trading day. CORRECTED IN ROUND 1, and the correction pulls the date forward thirteen days. Canadian Copper's financial year ends 31 October - its Q1 filing covers the period ended 31 January - so the July quarter is a Q3, and under the 60-day venture-issuer deadline in NI 51-102 it is due 29 September. That is a discrete, dated, impactful event: it supersedes the LARGEST estimate in this report, net cash at C$21.20M on a declared range of C$19.20-23.20M, and it is the filing that would finally let Gate 1's going-concern arm be VERIFIED rather than left recorded as open. An earlier draft scheduled +14d to 13 October on the stated ground that no dated company binary fell inside the window; that was wrong, because this one does. It also settles the round-1 auditor's either/or on the same filing: it is not a primary source this report ignored, because at the 27 September stamp it was not yet due. Two further dated catalysts now sit in the window and are carried in the section 8 table rather than driving the schedule, because both fall after 30 September: the assay results for holes MB26-06 to MB26-10, which the company guides to OCTOBER 2026, and the C$0.25 warrant expiry on 12 November. The recurring US macro releases are still NOT used as a trigger - five fall inside seventeen days, which is exactly the cadence the general prohibition governs.",
  "audit_passes": 3,
  "open_findings": [
    {
      "severity": "MAJOR",
      "location": "Section 11, bull case, leg one",
      "finding": "Section 11's bull case justifies 71% of its C$1.10 target on the construction discount having 'nothing left to compensate' because 'a name is no longer pre-production', but on this report's own timeline - construction 2027, production targeted end-2028 - the company would still be mid-construction and roughly fifteen months from production at the twelve-month horizon, which is precisely when construction risk is live.",
      "effect": "The bull target's largest leg rests on a state the report does not expect to exist within the scenario's own horizon. Read C$1.10 as a market re-rating assumption, not as a discount that has expired."
    },
    {
      "severity": "MAJOR",
      "location": "Section 11, base case, and the scenario ladder's multiples",
      "finding": "Section 11's base case says 'the market pays roughly the deck-case NAV ... so the construction discount unwinds', which conflates a market multiple with this report's warranted multiple - and on the published 0.65x warranted multiple the base target of C$0.79 is a warranted ratio of 1.5477x (Expensive) and the bull target of C$1.10 is 2.1550x (above the 2.0x deep-expensive Do-Not-Buy arm the report itself cites), so 77% of the probability weight sits on prices at which the report's own Gate 3 would fire, and section 11 nowhere says so.",
      "effect": "Readers cannot see from section 11 that the base and bull paths both imply a valuation this report's own gate would cap at HOLD. The ladder's multiple assumptions - 1.0x NAV for base and bull against the 0.65x warranted multiple used for the signal - are asymmetric in the bullish direction and undeclared. Nothing in the published signal chain depends on them: the signal is set by the spot anchor at 1.37x, not by the scenarios."
    },
    {
      "severity": "MAJOR",
      "location": "Section 11, bull case, the C$74M uplift",
      "finding": "Section 11 asserts that marking the deck 16% toward spot 'adds C$74M of after-tax present value', which is 28% of the bull target, without publishing the derivation, and one of its three inputs - the PEA's pre-tax NPV7 of C$264M, which supplies the 35.2% effective tax - appears nowhere in the report, while section 15 lists LOM revenue only as the rounded 'C$2.3B' rather than the C$2,258M actually used.",
      "effect": "A reader cannot reconstruct or falsify the figure from the declared sources, in a report that elsewhere publishes every weighted build. The auditor verified both inputs independently against the PEA's Table 1 and reproduced C$74.40M and the 1.0928x deck factor exactly, so the number is right and only its provenance is missing."
    }
  ],
  "prior_report": "calibration-CCI.CN-20260911-1600.json",
  "prior_primary": "BUY",
  "absence_searches": [],
  "superseded_values": [
    "0.7910",
    "0.9103",
    "1.3003",
    "24.38",
    "28.46",
    "1.50",
    "0.58",
    "C$18M",
    "13.3"
  ],
  "superseded_exempt": {
    "13.3": "The prior NAV headroom to the Gate-3 ceiling, superseded at C$3.80M. Its only surviving occurrence is inside delta_vs_prior, which close_audit_round.py GENERATES by quoting the prior calibration's gates_caution verbatim - so the figure is describing the PRIOR state truthfully, which is the changed-role case this field exists for. The withdrawal itself is stated in full in superseded_detail and in the changes box."
  },
  "superseded_not_swept": {
    "0.70": "The prior warranted P/NAV multiple. NOT declared for the sweep because the digit string collides with THIS run's share price of C$0.70, so every honest mention of the price would register as a surviving superseded value and the check would become noise on its own output. The withdrawal is stated in full in the changes box and in section 4.",
    "13.3": "The prior NAV headroom to the Gate-3 ceiling, superseded at C$3.80M. REMOVED from the sweep list in round 1, not from the withdrawal. Its only surviving occurrence is inside delta_vs_prior, which close_audit_round.py GENERATES by quoting the prior calibration's gates_caution text verbatim - so the figure is describing the prior state truthfully, which is precisely the changed-role case superseded_exempt exists for. The builder's post-render sweep does not read superseded_exempt (the command-line path does), so the exemption is declared here instead. The withdrawal itself is stated in full in superseded_detail and in the changes box.",
    "0.55": "The prior fair-value estimate of C$0.55. NOT declared for the sweep because it collides with the PEA's mined gold grade of 0.55 g/t Au, which is a live figure quoted from the study. The withdrawal is stated in section 4 and in superseded_detail."
  },
  "superseded_detail": [
    {
      "value": "superseded 0.7910",
      "what": "NAV per fully diluted share",
      "was": "superseded from C$0.7910 (11 Sep 2026)",
      "now": "C$0.7853",
      "why": "the give-up is now after-tax at the PEA's own 0.746 USD/CAD rather than pre-tax at 1.39, the consideration receivable is discounted at this run's r over a 2.0-year draw with a conditionality haircut, and net cash is rebuilt from the disclosed 30 April figure"
    },
    {
      "value": "superseded 0.9103",
      "what": "actual P/NAV",
      "was": "superseded from 0.9103x (11 Sep 2026)",
      "now": "0.8914x",
      "why": "price fell to C$0.70 and NAV per share rose"
    },
    {
      "value": "superseded 1.3003",
      "what": "warranted ratio",
      "was": "superseded from 1.3003x (11 Sep 2026)",
      "now": "1.37x",
      "why": "the construction discount tightened from 30% to 35%, so the warranted multiple fell from 0.70x to 0.65x"
    },
    {
      "value": "superseded 0.70 (not swept)",
      "what": "warranted P/NAV multiple",
      "was": "superseded from 0.70x (11 Sep 2026)",
      "now": "0.65x",
      "why": "two of the three construction-risk legs remain open - no bankable study, EIA in review - which withholds the top of the 30-50% band"
    },
    {
      "value": "superseded 24.38",
      "what": "OR stream give-up, C$m",
      "was": "superseded from C$24.38M (11 Sep 2026)",
      "now": "C$23.30M",
      "why": "that figure was struck PRE-TAX at 1.39 USD/CAD against an AFTER-TAX NPV computed at the PEA's 0.746; both bases are now matched to the term it is netted against"
    },
    {
      "value": "superseded 28.46",
      "what": "OR consideration receivable, C$m",
      "was": "superseded from C$28.46M (11 Sep 2026)",
      "now": "C$22.73M",
      "why": "discounted at r = 11.684% over a 2.0-year average construction draw with a 10% milestone-conditionality haircut, where the prior figure implied roughly a one-year deferral with no allowance for conditionality"
    },
    {
      "value": "superseded 1.50",
      "what": "bull-case 12-month target",
      "was": "superseded from C$1.50 (11 Sep 2026)",
      "now": "C$1.10",
      "why": "the superseded C$1.50 is 1.84x this run's NAV per share; the bull case for a developer with no bankable study is capped here at 1.35x NAV, which also sits nearer the sole analyst target of C$1.00"
    },
    {
      "value": "superseded 0.55 (not swept)",
      "what": "fair value estimate",
      "was": "superseded from C$0.55 (11 Sep 2026)",
      "now": "C$0.5104",
      "why": "fair value is warranted multiple times NAV per share, and both moved"
    },
    {
      "value": "superseded 0.58",
      "what": "stop loss",
      "was": "superseded from C$0.58 (11 Sep 2026)",
      "now": "C$0.625",
      "why": "the relevant support is now the C$0.68 low twice tested on 16 and 23 September, so the stop sits below it with ATR room"
    },
    {
      "value": "superseded 13.3",
      "what": "NAV headroom to the Gate-3 ceiling, C$m",
      "was": "superseded from C$13.3M (11 Sep 2026)",
      "now": "C$3.80M",
      "why": "both the NAV and the warranted multiple moved, so the distance to the 1.40x edge is re-measured"
    },
    {
      "value": "superseded C$18M",
      "what": "net cash estimate",
      "was": "superseded from near C$18M on a range of C$15-22M, rolled forward from the 31 January 2026 filing (11 Sep 2026)",
      "now": "C$21.20M on a range of C$19.20-23.20M",
      "why": "rebuilt from the C$17.3M of available cash disclosed at 30 April 2026, which was on the record when the prior report used a January anchor and said so"
    },
    {
      "value": "superseded 212,800,462",
      "what": "an intermediate share-count derivation",
      "was": "superseded from the pre-exercise cross-check figure",
      "now": "212,733,004",
      "why": "the post-exercise derivation (45,397,223 / 0.2134) is the one used; the pre-exercise route agrees to 0.03% and is shown as a cross-check, not as the count"
    }
  ],
  "changes_note": "Short-term signal DOWNGRADED BUY -> HOLD; medium and long HOLD at BUY. CONVICTION LADDER MOVED: entry_conviction Half-Size -> Wait, and entry_groups_met 1 -> 0, because the one path open on 11 September was the Technical group and it has closed. The downgrade is the short technical-confirmation cap, not a change of view: the 31 August breakout that confirmed the prior report's short BUY has faded through three consecutive lower weekly closes (C$0.74 -> C$0.72 -> C$0.70), the MACD histogram has been negative for EIGHT consecutive sessions since 16 September (narrowing on 22 September, then widening to -0.0039), and volume on 25 September ran at 0.23x the 20-day average. Price C$0.72 -> C$0.70 (-2.8%). Pillars: Quality 74 -> 70, Valuation 45 -> 46, Timing 70 -> 68, Driver 70 -> 77. The name is a Donatien Pick and it STAYS one - the mandatory auto-Stop needs no BUY on ANY horizon and medium and long are both BUY. FOUR THINGS THIS RUN CORRECTS IN MY OWN 11 SEPTEMBER REPORT, two of which move the NAV UP and are stated as such so the direction is not hidden: (1) the OR Royalties give-up was charged PRE-TAX at 1.39 USD/CAD against an AFTER-TAX project NPV struck at the study's own 0.746 - two mixed bases in one subtraction, worth C$1.08M; (2) the consideration receivable was discounted too lightly, implying roughly a one-year deferral with no allowance for the five construction milestones each instalment is conditional on, worth -C$5.73M; (3) net cash was rolled forward from the JANUARY filing, and the report said so - the 30 April disclosure of C$17.3M available cash was on the record and is now the anchor; (4) the bull case of C$1.50, now superseded, was 1.84x NAV per share for a name with no bankable study. AND ONE THING THIS RUN TIGHTENS AGAINST THE NAME: the construction discount moves from 30% - the most generous end of the framework's band - to 35%, because two of the three construction-risk legs are still open. The band is FULL either way, but it is NOT robust: a 40% discount, the midpoint of the framework's own range, puts the ratio at 1.4341x, which fires Gate 3, caps all three horizons at HOLD and would auto-Stop the Pick. Carried without change: exit_action Hold, exit_groups_live 0, hard_gate_state caution, gates_triggered empty, analysis_status donatien-pick, donatien_pick true, competitive_share_trajectory gaining, competitive_threat_level moderate, economic_alignment_stance Trend-Following, economic_alignment_pressure Tailwind, economic_alignment_source sector-map.",
  "company": "Canadian Copper Inc.",
  "currency": "CAD",
  "brand": "",
  "gics_sector": "Materials",
  "exchange": "CSE",
  "api_ticker": "CCI.CN",
  "donatien_pick": true,
  "donatien_pick_evidence": "this name's own calibration history carried analysis_status 'donatien-pick'",
  "carried_verbatim": [
    "api_ticker",
    "brand",
    "company",
    "competitive_share_trajectory",
    "competitive_threat_level",
    "currency",
    "donatien_pick",
    "donatien_pick_evidence",
    "exchange",
    "gics_sector",
    "ticker"
  ],
  "carried_verbatim_note": "Generated by scripts/carry_forward.py from calibration-CCI.CN-20260911-1600.json. Every field listed here crossed the run boundary through the manifest, with its disposition and its reason recorded in carry_provenance \u2014 not by being re-typed into a builder.",
  "carry_provenance": {
    "ticker": {
      "disposition": "CARRY",
      "reason": "identity"
    },
    "brand": {
      "disposition": "CARRY",
      "reason": "identity"
    },
    "company": {
      "disposition": "CARRY",
      "reason": "identity -- the legal name"
    },
    "currency": {
      "disposition": "CARRY",
      "reason": "the listing currency; a change is a re-listing, not a refresh"
    },
    "exchange": {
      "disposition": "CARRY",
      "reason": "identity"
    },
    "api_ticker": {
      "disposition": "CARRY",
      "reason": "identity -- the data-source symbol"
    },
    "donatien_pick": {
      "disposition": "CARRY",
      "reason": "an OPERATOR decision, not an analytical one \u2014 it persists until the operator changes it, and it is a separate field from analysis_status precisely so a lifecycle change cannot erase it. The auto-Stop scrubbed it from ten names before this field existed."
    },
    "donatien_pick_evidence": {
      "disposition": "CARRY",
      "reason": "where the pick fact came from; carried with the fact"
    },
    "gics_sector": {
      "disposition": "CARRY",
      "reason": "the GICS sector, not the descriptive `sector` line"
    },
    "competitive_share_trajectory": {
      "disposition": "CARRY",
      "reason": "a snapshot of the last competitive disclosure",
      "condition": "no_new_competitive_data",
      "condition_asserted": "Re-checked 2026-09-27: the only release since 11 Sep is the 21 Sep Murray Brook drill update (assays pending). No new competitor disclosure, no change in the Bathurst camp field. Competitor set re-walked: Foran Mining (McIlvenna Bay), Osisko Metals (Pine Point), Callinex Mines (Pine Bay)."
    },
    "competitive_threat_level": {
      "disposition": "CARRY",
      "reason": "as competitive_share_trajectory",
      "condition": "no_new_competitive_data",
      "condition_asserted": "Re-checked 2026-09-27: the only release since 11 Sep is the 21 Sep Murray Brook drill update (assays pending). No new competitor disclosure, no change in the Bathurst camp field. Competitor set re-walked: Foran Mining (McIlvenna Bay), Osisko Metals (Pine Point), Callinex Mines (Pine Bay)."
    }
  },
  "carry_source": "calibration-CCI.CN-20260911-1600.json",
  "threshold_fragility": "Ratio 1.3714, FULL band. NOT ROBUST. It is 3.8 NAV units from the 1.40x edge and 26.5 from 1.20x. Largest single sensitivity: capex revision at the December 2026 feasibility study (pea_capex_revision), worth 0.158 on the ratio and closing 506% of the nearer gap (to 1.40x) \u2014 it CROSSES the edge on its own. OR financing basis - credit the surplus, or treat the pair as value-neutral: 0.049, covering 170% of that same gap. corporate overhead and owner's costs to first production: 0.050, covering 98% of that same gap. marginal tax rate applied to the stream give-up: 0.037, covering 78% of that same gap. net cash at the report date: 0.029, covering 53% of that same gap. discount rate applied to the stream give-up: 0.049, covering 0% of that same gap. Taken together the declared inputs close 904% of the gap to 1.40x and 52% of the gap to 1.20x. The fragility is ONE-SIDED and points WORSE. Do not claim both adjacent bands sit inside the envelope \u2014 that claim was published once and was false.",
  "lint": {
    "exit": 0,
    "assertions": 69,
    "errors": 0,
    "warnings": 5,
    "could_not_run": [],
    "at": "2026-09-27T17:21",
    "report": "CCI.CN_Signal_v6_20260927_1200.html"
  },
  "cite_audit": {
    "round": 3,
    "rows": 60,
    "supported": 22,
    "not_supported": 1,
    "partial": 35,
    "census_sha256": "ce6c70f5150f7260fbef988aa27d49159068dbfc21942383fa7598d2eb77e141",
    "at": "2026-09-27T13:56",
    "disposition": "Round 3 Layer-0.5 result, and what has happened to it since. A narrow source auditor - a fresh general-purpose agent given the census and the declared sources ONLY, never the report - adjudicated 60 rows against a stated rubric: 22 SUPPORTED, 1 NOT-SUPPORTED, 35 PARTIAL, 2 MIS-CITED. It is stricter than the round-1 pass (34/1/22/1 on a 58-row census) on one declared axis: a figure derived from a SERIES a source note declares but whose individual values the note does not contain is graded PARTIAL, not SUPPORTED. The rubric is recorded so the difference between the two passes is auditable rather than mysterious. GATE: the single NOT-SUPPORTED is CLOSED. It was the word 'Buy' in the Ventum rating claim - the release announcing the initiation explicitly disclaims both the rating and the target, the provider returns a null recommendation, and the auditor searched for a wire carrying a Buy and found none. The rating is now withdrawn as a sourced claim in both the section 7 sentence and dnb_adjudication, and only the initiation date and the provider-panel target remain. Both MIS-CITED rows are closed: the C$1.00 target re-pointed to yq-stale, and the December feasibility study to cci-may27 at the second site where the round-2 re-point had been missed. PARTIALS: the cheap and material ones are closed - the three sources that were declared and never once cited (news-0927 for the walked release index, macro-0909 for both 'since 9 September' deltas, ew-aug13 for the derived basic share count), the deck column of the metal table, the pre-revenue basis behind every 'not applicable' in the report, the economic calendar, and an unsourced '10-year touched 5.01% on 16 September' now stated as the settled 5.006% and cited. The remainder are of one shape - a computed figure whose inputs are all declared and cited but whose output carries no marker of its own - and they are left, because demanding a citation on every derived numeral is how a citation convention gets abandoned under load. NOTE ON THE HASH: this receipt carries the census as it stands AFTER those fixes (ce6c70f5150f7260, 61 rows), not the 60-row census the auditor read. That is deliberate - the field exists so the NEXT run can tell whether the census has moved and therefore whether a re-audit is owed, and it has moved. The counts are the auditor's, against the 60 rows it actually adjudicated. A round-2 receipt in this file previously carried a hash that matched no census of this artifact and round-1's counts summing to 57 against its own rows:60; the round-3 auditor caught it, and that is what this field replaces."
  },
  "post_audit_fixes_round1": "TWO BLOCKERS CLOSED. BL-1: dnb_adjudication asserted the insider search was 'recorded in absence_searches' when that array is empty by design; corrected to name src:sedi-none, and \u00a715's newsfile claim restated as a POSITIVE enumeration of the three releases the index contains rather than an assertion that nothing exists. BL-2: the assays ARE dated - the 21 Sep release guides holes MB26-06 to MB26-10 to OCTOBER 2026 - and the report asserted three times that they were undated while citing that release; \u00a78 now carries the row, \u00a712's forecast cell is corrected, and next_update_basis is rewritten. EIGHTEEN MAJORS: MJ-1 the give-up was discounted at the equity rate against a 7% project NPV, a basis mismatch worth C$6.63M in the FLATTERING direction - fixed at 7%, and the give-up's discount basis is now its own declared sensitivity input because financing_basis claimed to cover only the receivable. MJ-2 gates_caution[1] carried a stale count AND magnitude ('four inputs', 'C$12.17M') in the one sentence adjudicating reachability - both now derived. MJ-3 NAV/sh printed 0.8136 and 0.8135; every display now divides the unrounded total. MJ-4 the ratio printed at three values and the sentence explaining the precision was falsified by its own page; the scorecard and pillar header now render from one field. MJ-5/MJ-6 net_debt_ebitda 0.0 and eps_trailing/trailing_pe were fabricated to make two checks RUN, each contradicting its own na_reason and each suppressing the skip code - all reverted to null, the could-not-run declared in \u00a715, and the framework gap filed as LINT-PREREVENUE-NA. MJ-7 the peer-median lens was scored 40 at a 26.7% weight with no declared peer P/NAV anywhere - WITHDRAWN, weights re-normalised, and the flattering side-effect (+1.7 on the blend) stated. MJ-8 the uncited 0.6-0.9x market cross-check, the only external validation of the deciding input - WITHDRAWN. MJ-9 the 10-year was a week stale at 4.998%; it settled at 5.184% on 25 Sep, so r is 11.684% and the guardrail-tightening adjudication is restated (0.184pp above 5% = zero notches). MJ-10 the MACD run is eight sessions not five and it NARROWED on 22 Sep. MJ-11 the Ventum rating and target were cited to a release that disclaims both; the target now cites the provider panel and the BUY rating is withdrawn as a sourced claim. MJ-12 fx_note claimed exactly one FX conversion; the US$2M debt repayment at 1.39 is a second and is now declared. MJ-13 five new sources declared: the 14 Apr C$96M release, the 27 May FS Team Award, the news enumeration, the 9 Sep prior macro state and the failed estimates pull. MJ-14 the Layer-0.5 cite receipt is stamped. MJ-15 the published lock hash was the amendment's prior_sha256 and verified against nothing; it now reads from the lock file. MJ-16 the warranted-multiple criterion declared funding RESOLVED while listing the EIA open, when the C$31.5M tranche is conditional on that very permit - rewritten on three legs, and 35% is retained on the unconditional brownfield-plant leg alone. MJ-17 \u00a713 now carries real risk-per-share math AND a narrowed subtitle, both halves of the registered S13-SIZING-MATH fix. MJ-18 the 52-week percentile is 79.7th, not 81st. MINORS: the C$7.02M/C$5.48M split of the C$12.5M receipt is now shown; 'restates' softened to 'may revise'; the 10 Sep C$0.68 low acknowledged; geopolitical status added to the \u00a715 live-verify list with a dated re-search. ONE FINDING DISPUTED AND DISPROVED with the primary series rather than argued down: the source auditor reported HG=F at 6.6955 and SI=F at 64.245 against the published 6.7660 and 64.8010. The 25 Sep bars are HG=F 6.7805/6.7970/6.7230/6.7660 on 25,665 lots and SI=F 64.2650/65.5050/63.7850/64.8010 on 36,179 - the five preceding sessions traded 88 to 1,239 lots, so those are thin pre-roll front-month bars and 25 Sep is the active contract's settle. The auditor's figures appear at no 25 Sep close. Recorded in \u00a715.",
  "audit_route_note": "Round 1: a FRESH general-purpose agent (never a fork) \u2014 Round 1 used a FRESH general-purpose agent (never a fork) because this was the first audit round on this artifact. Round 2 must ALSO be fresh rather than a continuation: the artifact was SUBSTANTIALLY REBUILT in this fix round, not patched. The stream give-up moved from the 11.684% cost of equity to the study's 7% project rate, which moved the NAV from C$192.34M to C$185.66M, the anchor from 1.32x to 1.37x and the headroom to the Gate-3 edge from C$10.48M to C$3.80M; the 10-year moved 4.998% to 5.184% so r moved with it; a sixth sensitivity input was added; the peer-median valuation lens was withdrawn entirely; the warranted-multiple criterion was rewritten on three legs instead of four; and the next-update date moved 13 October to 30 September on a catalyst the earlier draft denied existed. report-audit.md L51 routes a rebuilt artifact to a fresh auditor, because a returning auditor would be re-reading a different document while carrying commitments formed against the old one.. Verdict FAIL (2 blockers, 18 majors, 6 minors). Round 2: a FRESH general-purpose agent (never a fork) \u2014 Round 2 used a FRESH general-purpose agent, not a continuation, because round 1's fix pass rebuilt the artifact rather than patching it. Round 3 should CONTINUE with the round-2 auditor via SendMessage: this fix pass PATCHED - four blockers and eight majors closed with no change to the model's structure, no change to any pillar, band or signal, and no change to the anchor beyond a precision fix. A returning auditor re-reads the same document and can check its own findings cheaply, which is what report-audit.md L51 routes to for a patch.. Verdict FAIL (4 blockers, 8 majors, 5 minors). Round 3: SendMessage to the SAME auditor \u2014 Round 3 CONTINUED with the round-2 auditor via SendMessage rather than spawning fresh, because the round-2 fix pass PATCHED the artifact: no pillar, band or signal moved, the model's structure was unchanged, and the anchor moved only by a precision fix from 1.3713 to 1.3714. report-audit.md L51 routes a patch back to the same auditor, which retains context and re-checks only what changed. A separate NARROW Layer-0.5 source auditor was spawned fresh for the same round, given the census and the declared sources only and never the report, per the Layer-0.5 design.. Verdict PASS (0 blockers, 3 majors, 4 minors).",
  "delta_vs_prior": "vs calibration-CCI.CN-20260911-1600.json. signal_short BUY -> HOLD. quality 74 -> 70. valuation 45 -> 46. timing 70 -> 68. driver 70 -> 77. economic alignment 70 -> 72. price 0.72 -> 0.7. warranted_ratio 1.3 -> 1.37. entry_conviction Half-Size -> Wait. entry_groups_met 1 -> 0. short_entry_confirmed true -> false. next_update 2026-09-25 -> 2026-09-30. bull target 1.5 -> 1.1. relative strength vs SPY 16.2 -> 12.21. relative strength vs sector 16.4 -> 20.11. catalyst clustering set to 18 (absent in the prior report). moat 58 -> 54. industry benchmark score 72 -> 75. overall confidence 58 -> 45. probability-weighted target 0.866 -> 0.78. gates_caution 5 -> 6 (added Gate 1 Financial Distress - CAUTION, not triggered. Net debt is negative (C$21.2M cash, zero drawn debt) and interest coverage is not applicable with no interest. The 'negative FCF for 4+ quarters' arm is TRUE on its face and always will be for a developer, but the arm's own wording requires 'no clear path to positive', and C$96M committed against a C$64M capex with an FS due December 2026 and production targeted 2028 is a path. THE GOING-CONCERN ARM IS NOT CLEARED, IT IS UNVERIFIED: I did not retrieve the auditor's note, because the interim statements for the quarter ended 31 July 2026 were not available at the report date. Recorded as open rather than passed.; Gate 3 Valuation Ceiling - CAUTION, not triggered, and all four arms are adjudicated rather than three. (a) Expensive band at >= 1.40x warranted: does NOT fire at 1.3714x. On the model as built NAV would have to fall C$3.80M (-2.0%) to trip it, and a combined conservative reading of the 6 declared uncertain inputs swings C$34.31M, so this edge is reachable several times over - see threshold_fragility, which grades the band NOT ROBUST and one-sided worse. CORRECTED IN ROUND 1: this sentence previously read 'four uncertain inputs' swinging 'C$12.17M' - both the count and the magnitude were stale, in the one sentence that adjudicates whether the edge is reachable. (b) Materials sector guardrail at P/NAV >= 1.5x: does NOT fire at 0.8914x. The 10-year IS above 5% - it settled at 5.184% on 25 September - but the rule tightens the line one notch per ~1.5pp above 5% and 0.184pp earns ZERO notches, so the unadjusted 1.5x line is the one used. An earlier draft of this sentence said the 10-year was not above 5% at all, on the 20 September macro stamp of 4.998%, and that was false on this report's own date. (c) Price above the highest analyst target: does NOT fire - C$0.70 against a sole target of C$1.00. (d) Own-history decile: CANNOT BE CLEARED and is recorded as such, with the count - one loosely comparable prior reading, zero strictly comparable. A withheld lens must not silently clear a gate arm.; Gate 4 Dilution / Accounting - CAUTION, not triggered, on a stated carve-out rather than a self-exemption. The share-count arm fires on its letter: 200.1M before 11 August against 212,733,004 now, with 23,685,200 warrants (+11.1% fully diluted) still to come. It is recorded as a caution because the dilution is ALREADY INSIDE THE ANCHOR - the NAV divides by the fully diluted count and credits the C$5.92M of proceeds, so firing a gate on top would charge the same fact twice. That reasoning is a lifecycle carve-out the framework does not yet contain, so it is filed to the improvement-debt register as GATE4-DEV-LIFECYCLE with this report as the dated instance, rather than being invented again next run.; Gate 5 Regulatory / Binary - CAUTION, not triggered. The Murray Brook EIA registered 13 July 2026 is in review and a refusal would be existential for the timeline, which clears the >20%% move test. It is a caution rather than a trigger because the gate requires an outcome that is 'genuinely binary - could go either way': this is a brownfield restart of a past-producing plant whose own permits are already in hand, in a province that hosted a formal government event for the title transfer. Approval is the base case, so the outcome is not 50/50. No FDA, antitrust or takeover arm applies.; Debt service, prospective - the C$48M Ocean Partners tranche is a 36-month facility at SOFR+7.75%% with a 12-month grace period, a short tenor against a mine build, and it is FLOATING into a Fed that has just hiked 25bp to 3.75-4.00%% on a unanimous vote with 16 of 18 participants expecting at least one more. It is UNDRAWN today, so there is no debt service now; this is the gate that arms when construction draws it in 2027.; Single asset, pre-revenue - one complex, no operating cash flow, CSE-listed. 55%% of the C$96M package is optional at the Company's discretion and OR's construction instalments are conditional on five separate milestones. Committed is not the same as drawn.; removed Valuation Ceiling (Gate 3) \u2014 does NOT fire at a 1.30x warranted ratio (Full band), but the margin is thin and one-sided. On the model as built it sits C$13.3M of NAV below the 1.40x Expensive threshold that caps every horizon at HOLD; treat the OR Royalties financing as exactly value-neutral instead of crediting its consideration and the ratio is 1.38x \u2014 only C$2.2M below it. It also blocks amplification to STRONG BUY.; Dilution/Financing (23,685,200 warrants at C$0.25 expiring 12 Nov 2026 = +11.1% fully diluted; plus up to C$5M optional Ocean equity); Debt service (Ocean Partners C$48M at SOFR+7.75%, 36-month term with a 12-month grace period \u2014 a short tenor against a mine build); Regulatory/Binary (Murray Brook EIA registered 13 Jul 2026, in review; Feasibility Study with DRA Americas due Dec 2026 may revise the C$64M PEA capex); Single-asset, pre-revenue (no operating cash flow; 55% of the C$96M package is optional at the Company's discretion and OR's construction instalments are milestone-conditional)). Carried without change: signal_medium, signal_long, val_band, exit_action, hard_gate_state, analysis_status, base target, bear target, economic alignment SOURCE, competitive share trajectory, exit_groups_live.",
  "post_audit_fixes_round2": "FOUR BLOCKERS CLOSED. B1 the geopolitical block asserted a re-search and published the inherited figure: it claimed 'Brent near US$104' which is the 18 September close to the cent. Brent SETTLED at US$97.44 on 25 September, down 8.6% in one session from US$106.60 on the week's heaviest volume, and WTI at US$92.41 against the macro's US$100.30; the cause is the 25 September US-Iran side talks at the UN on a phased Hormuz reopening. The strait itself remains closed, so the status claim survives and the price does not. Two sources declared for it, oil-2509 and hormuz-2509. B2 next_update_date 2026-09-30 had reached the calibration and neither visible site: the header still read 13 October and the section 8 CPI row referred to it. Both propagated. B3 the MACD correction was asserted in section 9 and six other sites still carried 'widened negatively for five sessions', including the scorecard cell that explains the HOLD, the section 12 entry sub-condition that produces entry_groups_met 0, short_cap_reason - which quoted a series that narrows at step two while calling it 'steadily MORE negative' - and changes_note. All seven now read eight consecutive negative sessions since 16 September, narrowing on 22 September. B4 the ratio rendered at four values against a sentence claiming it renders at one: root cause was formatting the ROUNDED warranted_ratio at four decimals in section 6, and the sensitivity table computing from a 2dp nav_total. nav_total and every fragility input range are stored at 4dp now; the page carries 1.3714 fifteen times and no other value, and the value-neutral counterfactual is 1.4207 six times and no other. EIGHT MAJORS. M1 net cash C$21.20M was right and its label was wrong: the largest component of the C$9.0M deduction is the C$6.0M balance of the Caribou purchase price, not operating spend, and a reader adding it back would have reached a ratio of 1.4187x and concluded Gate 3 fires. The Caribou payment is now named and the issuer's own 16 July C$32.6M cash statement is published as the reconciliation. M2 the uncited 0.6-0.9x cross-check was withdrawn from the prose in round 1 and was still standing in nav_cci_cn_20260927.py, the file the brief calls the source of every figure - removed, because a fix is code AND agent - and section 11's funded and unfunded developer ranges are withdrawn with it. M3 the bull's stated mechanism did not produce its number: C$1.10 is 1.40x NAV per share while the withdrawn range topped out at 0.90x = C$0.71. Section 11 is rebuilt on two derived legs - the construction discount unwinding to 1.0x NAV, then a deck marked 1.093x, 16% of the way to spot, adding C$74M of after-tax present value - and the bear on one derived mechanism, capex +30% at an unchanged warranted multiple. Neither invokes a peer multiple. A self-inflicted error inside that new text was caught by the pre-close artifact grep: leg two had been solved at the warranted 0.65x rather than at 1.0x NAV, overstating the required deck mark-up as 1.267x. M4 three of four relative lenses are unavailable, not two, and the peer-median withdrawal that raised the blend from 44.70 to 46.40 is now disclosed where the score is published rather than only in the fix log. M5 the sensitivity count is derived, six inputs not five. M6 the C$38.35M less C$7.02M gap against the stated C$31.5M is the issuer's own inconsistency and is flagged as such rather than published as arithmetic. M7 the census is re-run against the final artifact and cite_audit re-stamped at round 2. M8 delta_vs_prior regenerates on this close, which is the only thing that could clear it. MINORS: the stale share-count comment, the blended volume bounds in the price dispute, the September lead average, and 'five points clear' against a 69.50 build. THE ROUND-1 DISPUTE IS RESOLVED IN THE REPORT'S FAVOUR: the round-2 auditor independently pulled both series and confirmed HG=F 6.7660 on 25,665 lots and SI=F 64.8010 on 36,179, against 739-1,458 and 51-216 in the five preceding sessions, and stated the published figures belong in the report. PROCESS: round 1 was closed before the mandatory built-artifact grep and that cost a whole class of sibling misses. This round ran the grep FIRST and it found three - the bull's wrong solve, the 2dp precision source, and the reconciliation that had landed in a code comment instead of the page.",
  "post_audit_fixes_round3": "PASS. Round 3 returned 3 BLOCKER, 5 MAJOR, 5 MINOR on the audit MACHINERY, with the report body verified clean - all four round-2 blockers genuinely closed and eight of eight majors plus five minors. ALL THREE BLOCKERS NOW CLOSED. BL-A: the decision lock had been re-frozen TWICE under the same stamp, which silently reset the amendment count and destroyed the record of a round-2 amendment whose stated reason was itself a material finding, while the page claimed the lock was 'frozen before any prose was written' of a file that post-dated two renders. freeze_decision's own refusal text says to re-freeze under a NEW stamp for exactly this reason. The full chain is restored by hand as refreeze_history on the lock - three baselines with the amendment each carried - the page now states the truth plainly, and the missing overwrite guard is filed as FREEZE-SAME-STAMP-OVERWRITE at P1. BL-B: the round-2 cite_audit receipt carried a census hash matching no census of this artifact and round-1's counts summing to 57 against its own rows:60, certifying a Layer-0.5 pass that had not been performed. A fresh narrow source auditor re-adjudicated all 60 rows - 22 SUPPORTED, 1 NOT-SUPPORTED, 35 PARTIAL, 2 MIS-CITED against a stated rubric - and the receipt now carries those counts, the current census hash and a disposition explaining both. BL-C: the live watchlist row published the OPPOSITE short signal (BUY against this report's HOLD), the prior run's pillars, C$0.72, Half-Size and a link to the superseded report. build_watchlist.py --write and build_portfolio_grid.py --write both run; the row reads 70/46/68/77/72, HOLD/BUY/BUY, Wait, 27 Sep, 30 Sep, and CCI.CN correctly LOSES its Materials-CA grid tile on the short HOLD, 3 tiles to 2, with PortfolioWatchlist-state-20260927.json rewritten to 2 winners. preflight_report.py now returns 0 errors. THE LAYER-0.5 GATE IS CLEARED: the single NOT-SUPPORTED row was the word 'Buy' in the Ventum rating claim - the release disclaims both rating and target, the provider returns a null recommendation, and the auditor searched for a wire carrying a Buy and found none - and it is withdrawn from both the section 7 sentence and dnb_adjudication. Both MIS-CITED rows closed: the C$1.00 target re-pointed to yq-stale and the December feasibility study to cci-may27 at the site where the round-2 re-point had been missed. Material PARTIALs closed: three sources were declared and never once cited (news-0927, macro-0909, ew-aug13) and now are; the deck column of the metal table, the economic calendar and the pre-revenue basis behind every 'not applicable' are cited; and an unsourced '10-year touched 5.01% on 16 September' is now the settled 5.006%, sourced. TWO OF THE FIVE MAJORS WERE FIXED because they were cheap and consequential: a dangling '- five sessions' left inside the section 12 condition that PRODUCES the short HOLD, created by my own round-2 fix; and the model's anchor comment still carrying the two-leg funding framing round 1 replaced, with the three-leg sentence added beneath it falsified by its own neighbour. THREE MAJORS SHIP DISCLOSED under the operator's 2026-09-27 rule, verbatim in the auditor's words with audit_passes:3 - the bull's leg one resting on a state the report does not expect within its own horizon; the base and bull targets implying warranted ratios of 1.5477x and 2.1550x so that 77 per cent of the probability weight sits on prices at which this report's own Gate 3 would fire, undeclared in section 11; and the C$74M bull uplift carrying no published derivation. They are NOT half-fixed: partially closing a finding to make the disclosure read better is the downgrade the rule forbids. PROCESS, and it held for the third consecutive round: the mandatory grep of the BUILT artifact found something every time - in round 1 a correction that had landed in one of four sibling sites, in round 2 three more including an error inside my own new bull text, and in round 3 the last surviving 'Buy'. Round 1 was closed BEFORE that grep and it cost a whole class of misses; rounds 2 and 3 ran it first.",
  "report_detail": {
    "navps_disp": "C$0.7853",
    "nav_total_disp": "C$185.66M",
    "net_cash_disp": "C$21.20M",
    "cash_lo_disp": "C$19.20M",
    "cash_hi_disp": "C$23.20M",
    "warr_proceeds_disp": "C$5.92M",
    "capex_disp": "C$64M",
    "ratio_exact_disp": "1.3714",
    "ratio40_disp": "1.4856&times;",
    "ratiovn_disp": "1.4207&times;",
    "deck_scale_disp": "1.5722&times;",
    "shares_basic_disp": "212,733,004",
    "shares_fd_disp": "236,418,204",
    "blend_disp": "46.40",
    "deck_rev_disp": "127.241",
    "spot_rev_disp": "200.051",
    "cueq_disp": "6.67",
    "aisc_spot_disp": "52.9%",
    "aisc_deck_disp": "26.0%",
    "analyst_up_disp": "30.0%",
    "basis_asym": "DECLARED, not silent. The two sides of the OR Royalties transaction are priced on different metal decks and the gap is large: the consideration is C$29.75M in present value (C$7.02M received on 16 July, inside net cash, plus C$22.73M still receivable) against a give-up of only C$23.30M when silver is held at the study's US$27/oz. The pair therefore books C$+6.45M, which is 22% of the larger leg - far from the value-neutral result an arm's-length financing should produce. DIRECTION AND CAUSE: the asymmetry is not an error in the model, it is the difference between the PEA's silver deck and the market's. Solving for the silver price at which the pair is exactly neutral gives US$34.48/oz - 53% of the 25 September spot of US$64.80 [[src:px-2509]], which is precisely how a streamer prices a long-term deck. OR did not overpay; the study's US$27 is simply stale. WHY THE SURPLUS IS NEVERTHELESS KEPT: it is real on a deck-consistent NAV, because the company genuinely sold silver into a market that values it above the deck. And the inconsistency cuts the CONSERVATIVE way overall - if the market's silver view were applied to the PROJECT's silver revenue as well as to the stream, the NPV would rise by far more than the give-up would, because the project keeps 84% of payable silver and has sold 16%. WHAT IT COSTS: on a value-neutral treatment NAV falls to C$179.20M, NAV per share to C$0.7580 and the ratio rises to 1.4207x - PAST the 1.40x Expensive edge, firing Gate 3, capping all three horizons at HOLD and auto-Stopping the Donatien Pick. That is the single most important number in this report and it is one defensible accounting choice away from the published signal.",
    "threshold_fragility_txt": "Ratio 1.3714, FULL band. NOT ROBUST. It is 3.8 NAV units from the 1.40x edge and 26.5 from 1.20x. Largest single sensitivity: capex revision at the December 2026 feasibility study (pea_capex_revision), worth 0.158 on the ratio and closing 506% of the nearer gap (to 1.40x) \u2014 it CROSSES the edge on its own. OR financing basis - credit the surplus, or treat the pair as value-neutral: 0.049, covering 170% of that same gap. corporate overhead and owner's costs to first production: 0.050, covering 98% of that same gap. marginal tax rate applied to the stream give-up: 0.037, covering 78% of that same gap. net cash at the report date: 0.029, covering 53% of that same gap. discount rate applied to the stream give-up: 0.049, covering 0% of that same gap. Taken together the declared inputs close 904% of the gap to 1.40x and 52% of the gap to 1.20x. The fragility is ONE-SIDED and points WORSE. Do not claim both adjacent bands sit inside the envelope \u2014 that claim was published once and was false.",
    "benchmark_value": "NPV7 after-tax C$171M / 36% IRR / C$64M initial capex / 13.2yr LOM (Ausenco PEA, 22 May 2025); AISC US$3.14/lb CuEq against a spot basket of US$6.67/lb CuEq = a 52.9% AISC margin",
    "decile_reason_short": "a price-to-NAV decile needs at least two readings on one basis. There is ONE loosely comparable prior reading &mdash; the now-superseded 0.9103&times; of 11 September &mdash; and ZERO strictly comparable, because this run changed the basis again: the give-up is now after-tax at the PEA's own 0.746 USD/CAD rather than pre-tax at 1.39, and the receivable is discounted at this run's cost of equity with a conditionality haircut. The 11 September report said this arm would clear &lsquo;at the next report, when two clean readings exist&rsquo;. This is that report, and the condition is <strong>not</strong> met &mdash; one exists, not two. It clears from the report after this one.",
    "shares_basis_short": "Derived from a primary filing, not taken from a provider. The early warning report of 13 August 2026 states Ocean Partners held 45,397,223 shares representing 21.34% of the issued and outstanding after exercising 12,725,000 warrants; 45,397,223 &divide; 0.2134 = <strong>212,733,004</strong> [[src:ew-aug13]]. The independent cross-check from the same filing agrees to within 0.03%: 32,672,223 &divide; 0.1633 = 200,075,462 before the exercise, plus 12,725,000 = 212,800,462. <strong>The provider would have been wrong:</strong> Yahoo's market capitalisation of C$140.26M at C$0.70 implies 200.4M shares, which is precisely the pre-11-August count [[src:yq-stale]] &mdash; using it would have overstated NAV per share by 6%. Fully diluted adds the 23,685,200 warrants at C$0.25 [[src:cci-aug11]] to the 212,733,004 basic derived above [[src:ew-aug13]], for 236,418,204.",
    "cu_x_disp": "1.59&times;",
    "cu_share_disp": "26.7%",
    "zn_x_disp": "1.42&times;",
    "zn_share_disp": "48.0%",
    "ag_x_disp": "2.40&times;",
    "ag_share_disp": "16.6%",
    "pb_x_disp": "0.78&times;",
    "pb_share_disp": "8.6%"
  },
  "composite_short": 63,
  "composite_medium": 61,
  "composite_long": 62
}

Published HOLD / BUY / BUY at C$0.70 on Q70 V46 T68 D77 E72, overall confidence 45%. The anchor is 0.8914× actual over 0.65× warranted = 1.37×, the FULL band. Decision lock sha256 6aeff96bb07ef6ec. That lock was re-frozen twice during this report's audit, and an earlier version of this sentence wrongly described it as ‘frozen before any prose was written’ — true of the original freeze, false of the file that shipped. The original (0976aeed11a0dfae) was struck before a line of prose existed; it was re-frozen once when the round-1 auditor re-based the stream give-up from the cost of equity to the study's 7% project rate, and once more when a round-2 precision fix moved a derived figure inside a frozen gate string by one cent. Each re-freeze left every decision-bearing field byte-identical — all three pillars, all six signals, the band, both multiples, the gate arrays and the conviction — and the full chain, with the amendment each baseline carried, is recorded as refreeze_history on the lock. Both re-freezes were done under the SAME stamp, which the tool's own refusal text says not to do because it resets the amendment count silently; the round-3 auditor caught it, the destroyed record was restored by hand, and the missing guard is filed as FREEZE-SAME-STAMP-OVERWRITE. the base signal in the lock is BUY on all three horizons, derived by the linter rather than typed, and the short HOLD is the technical-confirmation cap applied after it. The field to read next to the signal is threshold_fragility, which is derived rather than written and reports the band as NOT ROBUST and one-sided worse.