pog-mcp 0.9.4 → 0.9.5

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pog-mcp",
-  "version": "0.9.4",
+  "version": "0.9.5",
   "type": "module",
   "description": "MCP server that lets an AI agent play Proof of Goal — wallet, sign-in, squad building, and matches as typed tools.",
   "license": "MIT",

package/skill/reference/measurements.md

@@ -2544,14 +2544,68 @@ that was an artefact of an unattainable +4 buff and is withdrawn.
   the short side drops by the same Δ. A real fatigued squad is uneven.
 
 
-## What load does a player actually accrue? (`engagement-load.mts` + `c2-team-load.ts`)
+## What growth-neutral load would a player accrue? (`engagement-load.mts` + `c2-team-load.ts`)
 
 **This section does NOT close C-2.** An earlier version said it did; that was wrong. `τ` was
 never measured — the probe's 24h default was simply used — and the neutral point was left
 unchosen between its two branches. An F-1 implementer following that decision would have made
 a probe default the production recovery rate. And `τ` cannot be measured from this data: F-1
 does not exist, so production has no fatigue and there is no observed recovery curve to fit.
-What measurement CAN do is price the choice, which is what the τ sweep does.
+What measurement CAN do is price the choices, which is what the τ sweep does. A third product
+choice sits beside τ and the neutral point: F-1 has not selected predicate C versus D. The
+published values are conditional on C; switching to D moves τ=24h median/p95 `L` by
+−5.3%/−10.4%. That measures the price of the choice, not which workload meaning the product wants.
+
+### Growth-leak gate: use a zero-growth shadow, not the real action trace
+
+The old workload source was unsafe. Growth raises the hidden effective `Player.total`; an early
+decision then changes the engine branch and RNG-consumption order, so even an unsigned action
+count becomes a channel for `potentialBp`. On the production window, the effective-total trace
+and a zero-growth replay differed in **668/1,201 matches and 8,195 predicate-C player cells**.
+Before making that attribution, the probe rejects any snapshot slot whose `total − Σattrs` is
+not an integer in the growth model's supported 0..5 range, so it cannot silently erase an
+unsupported total-only modifier. The range does not identify provenance by itself: the current
+producer's code contract—that this validated innate input receives only the growth buff—supplies
+that fact, and the guard fails closed when the encoded input contradicts it.
+A plausible slot-wise 0..5 total-only variation inside the same public squad-hash boundary changed
+**787/1,201 matches and 9,485 cells**. That is a counterexample to option (b), not a claim that
+production used that synthetic vector.
+
+The adopted source is option (a): copy the committed match snapshot, rebuild only each player's
+`total` as `pass + dribble + shoot + defense`, preserve condition, position, takers, fair-play
+tendency, name and slot order, and run a **separate deterministic engine instance** with the same
+stored seed and `gameFlg`. F-1 consumes unsigned engagement counters only from that shadow.
+The real result still uses the growth-buffed input; the shadow cannot change its score, events,
+ratings or growth. It is therefore a **zero-growth counterfactual workload**, never the action
+record of the match that happened. Runtime cost is one extra pure simulation per competitive match;
+F-1 still has to persist its counters atomically with completion. The workload timestamp must be
+growth-invariant too: fix every due match's first-condition-snapshot (claim-start) time before ANY
+real or shadow engine in that scheduler batch runs, then use that match's one stored value for both.
+A sequential loop that timestamps match B only after simulating match A would carry A's
+growth-dependent control-flow time into B's condition and is forbidden; F-1 therefore needs a
+claim-first/run-second phase. The batch epoch itself must not be a wall-clock read taken only after
+an earlier real engine finishes: it must be a persisted epoch independent of that earlier runtime,
+such as an externally scheduled tick, or else a public growth-independent time such as scheduled
+kickoff. Claim-first fixes ordering inside one batch; it does not make the delayed wall clock of a
+later batch invariant by itself. A post-simulation `completed_at` can react to
+the realized control-flow length; it is only this historical audit's proxy because the old rows do
+not carry the first-snapshot timestamp, not the production F-1 clock contract.
+
+The boundary is an induction conditional on the growth-independent inputs committed before the
+match (participants, XI, seed, mode and the first-condition-snapshot time fixed in that pre-run
+batch). Any two inputs
+that differ only in hidden growth project to identical teams, so a deterministic engine and tally
+produce identical counters. A fixed initial load plus a growth-invariant EWMA increment and decay
+timestamp keeps the
+next load — and the condition derived from it — invariant too. Growth can change public results and
+therefore which later matches exist, but those results and appearances are already public; the
+condition observation adds no hidden-growth channel given that ledger. The audit applied the
+projection to the synthetic variations above and got byte-identical candidate-F-1 counters for
+**1,201/1,201 matches and all 2,402 sides**. Unit tests pin erasure, preservation, non-mutation and
+idempotence; `match-growth.test.ts` pins that the real `applyMatchGrowth` producer projects exactly
+back to its innate team, and the existing squad-hash test independently pins that `total` is outside
+the hash. This closes only the §3-5 prerequisite, not F-1/F-3/G-1/F-4's own implementation and
+exposure contracts.
 
 **The gate is not a per-mode ceiling — it is defined on the exposure-weighted mix.**
 `band-worstcase.mts` holds the canonical statement (`01-engine-expansion.md:319-329` puts
@@ -2564,8 +2618,8 @@ Three things are established:
 
 - **Carrying E's ceiling across exceeds NOTHING at τ=24h.** Each team state is turned into a
   fresh and a tired variant and both are played against the SAME REAL OPPONENTS on the same
-  seed, with the modes mixed by exposure: at τ=24h, 66 cells give **0 over, 3 established
-  below, 63 undecided**. The short τ exceed badly — **32 of 52** at 6h, **12 of 63** at 12h.
+  seed, with the modes mixed by exposure: at τ=24h, 57 cells give **0 over, 3 established
+  below, 54 undecided**. The short τ exceed badly — **33 of 57** at 6h, **12 of 55** at 12h.
   Only three established below is the price of the guarantee, not a smaller effect: the cell is
   now (team × curve × STATE RULE) with six opponents inside it, so the family is **1,836**. **"none over" and "all established below" are different claims**: the first says
   exceeding cannot be shown, the second says clearing can. τ=24h is the first; getting the
@@ -2574,10 +2628,10 @@ Three things are established:
   diverge** — knockout point estimates run about **1.5×**
   ladder's, which is why holding each mode to 5pp separately showed 38 of 78 over. That is
   what the earlier "almost every team, in knockout" conclusion was: **the wrong bar**.
-- **So what this section delivers is a safe amplitude**: re-evaluating ALL 66 cells at each
+- **So what this section delivers is a safe amplitude**: re-evaluating all 57 realized cells at each
   candidate under a simultaneous correction and then validating the choice on INDEPENDENT
-  seeds gives **`b` ∈ 0.1988–0.2014** at τ=24h, worst cell 3.86pp. **All four τ values have
-  such an amplitude** (0.16–0.21), so fairness closes on the AMPLITUDE, not on the choice of τ.
+  seeds gives **`b` ∈ 0.1911–0.1936** at τ=24h, worst cell 3.43pp. **All four τ values have
+  such an amplitude** (0.15–0.20), so fairness closes on the AMPLITUDE, not on the choice of τ.
   That is **0.07×** the value derived by putting E's 0.2 on the median player. Testing one cell with a pointwise
   interval leaves winner's-curse bias: on a small sample a search that passed at 4.11pp
   validated at 10.48pp.
@@ -2585,10 +2639,10 @@ Three things are established:
   0.59pp [0.25, 0.93] and the paired difference is −0.43pp [−0.76, −0.10]: the prize does exceed
   the cost there. But **that is not where the decision lives.** A rested player can only recover
   the deficit the shipped amplitude creates, and the **safe amplitude produces a deficit of
-  0.135** (τ=24h, the largest across all validated cells — taking it from the cell with the
+  0.133** (τ=24h, the largest across all validated cells — taking it from the cell with the
   largest BOUND instead understates what can be recovered). At that deficit the prize is
-  **0.16pp [−0.08, 0.41]** against a cheapest cost of **0.16pp [−0.36, 0.67]**, paired difference
-  **−0.00pp [−0.47, 0.46] — undecided**, and the same at all four τ (the eight repricing calls
+  **0.16pp [−0.09, 0.40]** against a cheapest cost of **0.16pp [−0.36, 0.67]**, paired difference
+  **0.00pp [−0.47, 0.47] — undecided**, and the same at all four τ (the eight repricing calls
   publish a prize interval and a paired-difference interval apiece, so all **16 intervals** split
   one 5% budget). The cost of a one-point-weaker slot runs 0.16 to
   2.21pp depending on WHICH point moves, with selection and publication on SEPARATE seed columns
@@ -2607,26 +2661,31 @@ TIMELINE gives the numbers above. The first two are withdrawn.
 
 ### λ_team × multiplier cannot produce L
 
-That product multiplies two upper quantiles, so it puts `L_top` at 63 — while integrating
-the real timeline puts the peak `L` at **p95 21.3** (observed max 50.5). The two quantiles do
+That product multiplies two upper quantiles, so it puts `L_top` at 67.2 — while integrating
+the zero-growth shadow on the committed fixture ledger ordered by the verified claim clock puts
+the at-appearance `L` at **p95 25.31**
+(observed max 57.45). The two quantiles do
 not arrive on the same player-day. And bucketing by day discards the SPACING, which is
 everything to an EWMA: cup fixtures cluster into a few hours after opening, so eight closely
 spaced calls do not equilibrate like eight spread over a day.
 
 So the layer is gone. `matches.input_snapshot` (#694) freezes the exact two teams handed to
-the engine, so the last 30 days' **1,201 real kickoff matchups** are replayed in time order
-and the decay is integrated over them:
+the engine, so the last 30 days' **1,201 real kickoff matchups** are projected to zero growth,
+replayed in time order on their stored RNG columns, and the decay is integrated over them:
 
-    L ← L·exp(−Δt/τ) + (that match's engagements)
+    L ← L·exp(−Δt/τ) + (that match's zero-growth shadow engagements)
 
 Neither `λ` nor a multiplier appears. Each snapshot carries its own `gameFlg`, so the mode
 mix is right by construction — removing an estimate beats getting it right.
 
 Three further corrections: only **credited** appearances accumulate (a human ladder away side
-and an AI cup entrant carry zero F-1 load, and an earlier version pooled them into the human
-multiplier); each matchup is replayed **five times and averaged**, because a single replay's
-integer counts read within-match noise as between-player heterogeneity; and the synthetic
-grid is kept as a control only.
+carries zero F-1 load; cup sides, including AI entrants, are credited, while bots are excluded
+only from the human calibration cohort); the predicate-rate table replays each matchup's shadow
+**five times and averages only those rates**, because a single replay's integer counts read
+within-match noise as between-player heterogeneity; and the synthetic grid is kept as a control
+only. The same five fixed-fixture draws are kept as five integer observations for the one-match
+increment distribution. The EWMA timeline is different: it uses the one canonical integer shadow
+for each stored seed, never a five-replay average or a chain of derived-seed draws.
 
 **τ is therefore not a free parameter.** An earlier version used the equilibrium form and
 concluded τ cancels. On the timeline τ sets the load itself: every table below is τ=24h and
@@ -2681,22 +2740,23 @@ rather than assumed.
 
 ### F-1's predicate: counting actors only drops half, and concentrates the rest
 
-Production replay, 1,201 matchups over 5 replays, 8,283 credited human player-matches:
+Production-matchup zero-growth shadow, 1,201 matchups over 5 replays,
+8,283 credited human player-matches:
 
 | predicate | team total | p95 | median | p95/median | zero-load share |
 | --- | --- | --- | --- | --- | --- |
-| **A.** hyoka credit sites, unsigned | 6.60 | 1.60 | 0.40 | **4.00** | 13% |
-| **B.** actor only | 8.97 | 2.00 | 0.80 | **2.50** | 14% |
-| **C.** actor + responder | **17.88** | 3.20 | 1.60 | **2.00** | **4%** |
-| **D.** C without the ZP fallback | 16.61 | 2.80 | 1.40 | 2.00 | 4% |
+| **A.** hyoka credit sites, unsigned | 6.62 | 1.60 | 0.40 | **4.00** | 12% |
+| **B.** actor only | 8.91 | 2.00 | 0.80 | **2.50** | 14% |
+| **C.** actor + responder | **17.91** | 3.20 | 1.60 | **2.00** | **4%** |
+| **D.** C without the ZP fallback | 16.65 | 2.80 | 1.40 | 2.00 | 4% |
 
 The pool is credited HUMAN team-sides only — λ is a human-cohort figure, so the pool has to
 be as well. **This table alone uses expected RATES** (the 5-replay mean per player-match).
-Counting realized integers instead would put a player whose rate is 0.4 into "no load" on
+Counting individual shadow draws instead would put a player whose rate is 0.4 into "no load" on
 every draw that came up 0, and that player does accrue load across matches; the question here
 is whether a player is STRUCTURALLY load-free, not how often they were picked in one draw.
 The timeline integration asks the opposite question and uses integers.
-Counting actors only drops 49.8% of the engagement — (17.88−8.97)/17.88 — and what
+Counting actors only drops 50.3% of the engagement — (17.91−8.91)/17.91 — and what
 remains is MORE concentrated: p95/median 2.50 against 2.00 (4.00 at the hyoka sites), with a
 zero-load share three and a half times higher. Establishing §3-10's premise across the whole eleven needs responders counted.
 
@@ -2711,8 +2771,9 @@ Predicate D now subtracts fallbacks on BOTH sides of the contest: zone press pic
 through the fallback too, and an earlier version recorded only the responder's, so "C without
 the ZP fallback" was removing half of what it claimed.
 
-**A and C rose again in this pass** — A from 5.72 to **6.60**, C from 16.54 to **17.88** —
-because two sites were not being counted at all:
+**The instrumentation correction remains valid.** In that effective-total run A moved from
+5.72 to **6.60** and C from 16.54 to **17.88** because two sites were not being counted at all.
+The canonical values from the current single zero-growth-shadow run are **6.62 and 17.91**:
 
 - **The goal-scorer's credit does not go through `applyHyoka`.** `scoreGoal`
   (`engine.ts:1581`) writes `atk.hyoka[i] += 25 + rand(50)` DIRECTLY. Defining A as "the
@@ -2723,8 +2784,9 @@ because two sites were not being counted at all:
   In the production distribution: **2,894 selections, 13.2% of all defensive selections,
   7.1% of C**. If C means "defensive selections", this is one.
 
-**B did not move by a single digit** (8.97 → 8.97), which is the control that says both fixes
-touched only what they were meant to.
+**B did not move by a single digit in that correction** (8.97 → 8.97), which is the control
+that says both fixes touched only what they were meant to. It is **8.91** in the current shadow
+run; values from different workload sources are not combined into one control.
 
 ### The parameters
 
@@ -2734,11 +2796,11 @@ E capped the ATTAINABLE spread at 0.2, so the derivation fixes that and solves f
 143 human players over an **11-day window with 100% snapshot coverage** (2026-08-21 onward,
 1,201 matches), τ=24h.
 
-**"the match that happened" is verified per match, not inferred from a date.** Calling a
-stored-seed replay that only holds if the engine replaying it is the engine that ran it, and
-when that breaks fixture consistency breaks with it — a different knockout winner leaves a
-team that cannot exist in that timeline still accruing load, exactly why the derived-seed
-timelines were dropped.
+**The match-that-happened check is separate from the workload shadow.** First the effective-total
+snapshot and stored seed must reproduce the actual score, path and ratings, proving per match
+that today's instrumented engine matches the one that ran then. Only passing snapshots are
+projected to zero growth. The shadow uses the same RNG column but is never called the match that
+happened; the actual replay is evidence for engine-version and instrumentation compatibility.
 
 That boundary was once taken from `git log`'s COMMIT date. A commit date is not a deployment:
 it can ship after the following midnight, be rolled back, or roll out gradually. So the
@@ -2748,7 +2810,8 @@ it and the match happens the same way under today's engine, otherwise drop it. T
 golden-goal control flow, `b587f5d4` scene tags) did not change these matches' outcomes. The
 209 matches cut by date were fine.
 
-**The comparison runs on three axes, each closer than the last to what this probe counts:**
+**The comparison runs on three axes, each closer than the last to the probe's instrumentation
+sites:**
 
 | axis | what it covers | this window |
 | --- | --- | --- |
@@ -2796,18 +2859,20 @@ AVERAGED the counts, and fed those fractional expectations into the EWMA. That i
 every published quantity — per-player peak, p95, the gate deficits — is a NONLINEAR
 functional, so `E[f(X)] ≠ f(E[X])`, and averaging erases within-match selection streaks and
 narrows the load vector. A narrower vector passes the gate more easily, so the error runs in
-exactly the direction that says "safe". What runs now is **one realized timeline**, replaying
-each match's stored `matches.rng_seed` (**1,201 of 1,201**) — that pass is not a different draw
-of the same matchup but the match that happened.
-
-A middle version ran five timelines, four of them on derived seeds, and that was wrong too:
-the fixture list comes from production and is fixed, so a derived seed that changes a knockout
-winner leaves a team that **could not exist** in that timeline still accruing load in the next
-round. Those four were inconsistent histories, and they were feeding the published medians and
-the worst-state search. The cost of dropping them is that the across-draw spread is no longer
-measurable here — it needs bracket progression regenerated, which this probe does not do.
-
-Fixing the averaging moved the vectors up: max deficits rose 5–14% per cell and the per-match
+exactly the direction that says "safe". What runs now is **one committed fixture ledger ordered
+on the verified claim clock** with
+one zero-growth shadow per match, keyed by the stored `matches.rng_seed` (**1,201 of 1,201**).
+It uses the match's committed RNG column, but it is not the action trace that happened.
+
+A middle version mixed in four derived-seed timelines. Those are not values F-1 would persist:
+the canonical workload is the deterministic shadow of each committed match seed. A shadow result
+does not decide the real result or bracket, so this is explicitly conditional on the public fixture
+ledger rather than a counterfactual tournament history. The cost of using only the canonical draw
+is that across-draw workload sensitivity is not measured here; that would be a separate resampling
+analysis and must not be mixed into the production workload value.
+
+In the then-effective-total source, fixing the averaging moved the vectors up: max deficits rose
+5–14% per cell and the per-match
 increment rose **31–33%**. The safe amplitude moved about 2% (measured then against the
 per-mode gate, so the absolute values differ from today's) — the conclusion was insensitive to
 the error, but that is only knowable after re-measuring.
@@ -2816,9 +2881,9 @@ the error, but that is only knowable after re-measuring.
 
 | instant | median | p95 | observed max |
 | --- | --- | --- | --- |
-| **at appearance** (what cond reads, 7,139 reads) | **8.46** | **26.21** | 55.27 |
-| at the window end (an arbitrary clock time) | 7.03 | 16.24 | — |
-| path maximum (after its own match's engagements) | 19.35 | 38.17 | — |
+| **at appearance** (what cond reads, 7,139 reads) | **8.53** | **25.31** | 57.45 |
+| at the window end (an arbitrary clock time) | 7.25 | 15.39 | — |
+| path maximum (after its own match's engagements) | 19.93 | 36.23 | — |
 
 **The load event is not placed at kickoff.** F-1 pins the decay reference to the time the
 match's condition is FIRST SNAPSHOTTED, which is when the scheduler picks the match up —
@@ -2832,23 +2897,23 @@ So it is measured rather than argued. The published `claim` axis uses **1,144**
 `live` first, then that latter column records the completion transition. A per-row
 `claimSource` proves which hand supplied it; if either the completion time or its source is
 missing, the probe refuses the whole window instead of falling back to kickoff [Codex P2,
-2026-09-04]. Integrating on the explicit `kickoff` alternative instead gives median 8.66 and
-p95 26.35 — **2.3%** and **0.5%** away.
+2026-09-04]. Integrating on the explicit `kickoff` alternative instead gives median 8.71 and
+p95 25.35 — **2.2%** and **0.1%** away.
 
 **That comparison holds the warm-up boundary fixed.** Left to itself each variant recomputes
 `collectFrom` from its own first event — and switching the time base, or dropping the
 long-lag matches, is exactly what moves that first event. The printed difference would then be
 a clock effect PLUS a changed sample rather than the substitution alone. The published run's
-boundary is passed in so every variant collects from the same absolute instant. Here the
-correction moved the median by 0.01 (8.65 → 8.66): small, but measured rather than assumed.
+boundary is passed in so every variant collects from the same absolute instant; only that corrected
+alternative is reported, so values from different samples are not mixed.
 
 **Neither completion timestamp is the first claim.** The daemon PRESERVES a failed attempt's
 snapshot, resets the match to `scheduled`, and picks it up later, while cup `completed_at`
 records the attempt that finally succeeded. Playoff `finalized_at` is coarser still: it closes
 the live window roughly two to three minutes after the result was computed. The store has no
 first-snapshot column, so it cannot be recovered; excluding the **104** matches with more than
-60 minutes of lag gives median 8.66 and p95 27.26, **+2.4%** and **+4.0%**. The uncertainty on
-this axis is 2–4%, and F-1 persisting the first snapshot time
+60 minutes of lag gives median 8.72 and p95 26.15, **+2.2%** and **+3.3%**. The uncertainty on
+this axis is 2–3%, and F-1 persisting the first snapshot time
 removes it.
 
 **That this axis is a real claim clock is now proved by the ledger, not by lag** [Codex P2,
@@ -2879,7 +2944,7 @@ engine boundary) leaves players carrying load they arrived with, while the integ
 them at `L=0` — treating the boundary as a league-wide fatigue reset. An earlier version merely
 listed this as unmeasured; measured, it is real: discarding the first **2τ** as warm-up raised
 the at-appearance median from 7.32 to 7.76 (+6%) — **that run's numbers**; the current run
-prints only the warmed value, and with the predicate corrected it is **8.46**. The cold start was biasing load DOWN,
+prints only the warmed value, currently **8.53**. The cold start was biasing load DOWN,
 which makes the gate easier. The collection window is 8.8 days and the residual bias is
 `exp(−2) ≈ 0.135` — recorded, not claimed gone.
 
@@ -2892,23 +2957,23 @@ window end (right for "now", wrong for "on entry").
 
 | `L₀/L_top` | `b` | `a` (neutral) | team spread | one-player deficit | prize* | x per match | in pp |
 | --- | --- | --- | --- | --- | --- | --- | --- |
-| 1× | 0.820 | 0.200 | 0.200 | 0.410 | 0.602pp | 0.029 | 0.043pp |
-| 2× | 1.439 | 0.200 | 0.200 | 0.480 | 0.705pp | 0.046 | 0.068pp |
-| 4× | 2.679 | 0.200 | 0.200 | 0.536 | 0.788pp | 0.063 | 0.093pp |
+| 1× | 0.794 | 0.200 | 0.200 | 0.397 | 0.583pp | 0.029 | 0.043pp |
+| 2× | 1.387 | 0.200 | 0.200 | 0.462 | 0.680pp | 0.046 | 0.068pp |
+| 4× | 2.574 | 0.200 | 0.200 | 0.515 | 0.757pp | 0.063 | 0.093pp |
 
 \* deficit × slope (1.47 pp/cond); those deficits sit above the largest directly measured one
 (0.40) and the response is concave, so they are biased high. The measured figure is 0.59pp,
-and at the safe amplitude the whole-vector team-level figure is 3.86pp.
+and at the safe amplitude the whole-vector team-level figure is 3.43pp.
 
 Repricing at each safe amplitude uses the largest deficit across all independently validated
 cells, not the deficit of the cell with the largest confidence bound:
 
 | τ | safe `b` | deficit | prize there | cheapest cost | paired cost − prize | verdict |
 | --- | --- | --- | --- | --- | --- | --- |
-| 6h | 0.1633 | 0.112 | 0.13pp [−0.10, 0.37] | 0.16pp | 0.02pp [−0.44, 0.49] | undecided |
-| 12h | 0.1639 | 0.110 | 0.13pp [−0.10, 0.37] | 0.16pp | 0.03pp [−0.44, 0.50] | undecided |
-| **24h** | **0.1988** | **0.135** | **0.16pp [−0.08, 0.41]** | **0.16pp** | **−0.00pp [−0.47, 0.46]** | **undecided** |
-| 48h | 0.2061 | 0.137 | 0.16pp [−0.08, 0.41] | 0.16pp | −0.00pp [−0.47, 0.46] | undecided |
+| 6h | 0.1497 | 0.107 | 0.13pp [−0.10, 0.36] | 0.16pp | 0.03pp [−0.44, 0.50] | undecided |
+| 12h | 0.1610 | 0.111 | 0.13pp [−0.10, 0.37] | 0.16pp | 0.03pp [−0.44, 0.50] | undecided |
+| **24h** | **0.1911** | **0.133** | **0.16pp [−0.09, 0.40]** | **0.16pp** | **0.00pp [−0.47, 0.47]** | **undecided** |
+| 48h | 0.1922 | 0.132 | 0.16pp [−0.09, 0.40] | 0.16pp | 0.00pp [−0.46, 0.47] | undecided |
 
 The eight repricing calls publish a prize interval and a paired-difference interval apiece.
 All 16 intervals split one 5% family budget; each two-sided interval splits its share again
@@ -2916,10 +2981,11 @@ between its two tails [Codex P2, 2026-09-02; 2026-09-04].
 
 `x`, the number §3-1 asked to be stated: one more match costs a top-load player **0.029–0.063
 condition points (0.043–0.093pp)**. No "N matches per day" qualifier is needed — τ and the
-real spacing are already inside `L`. The increment is the p95 of the **realized integer**
-distribution (4 engagements; median 1) — one match is one draw, so the p95 of the expected
-RATE ("a busy player's average match") answers a different question. That distinction alone
-raised these two columns 31–33%.
+real spacing are already inside `L`. The increment is the p95 of the **fixed-fixture shadow integer**
+pool (five draws per fixture, 4 engagements; median 1) — one match is one draw, so the p95 of the
+expected RATE ("a busy player's average match") answers a different question. Those extra draws
+estimate the one-draw distribution; they are never linked into the canonical EWMA. That distinction
+alone raised these two columns 31–33%.
 
 **The pp column is a CONVERSION, not a measurement.** The 1.47 slope is a SECANT taken at
 `g = 0.4` (that one slot's loss divided by 0.4), and win probability is nonlinear in condition
@@ -2928,8 +2994,8 @@ the same slot and the same published seed column:
 
 | τ | per-match increment (cond) | **measured directly** | secant conversion | ratio |
 | --- | --- | --- | --- | --- |
-| 6h | 0.189 | **0.21pp** [−0.05, 0.47] | 0.278pp | 1.3× |
-| 12h | 0.124 | **0.14pp** [−0.10, 0.38] | 0.182pp | 1.3× |
+| 6h | 0.185 | **0.20pp** [−0.06, 0.47] | 0.272pp | 1.4× |
+| 12h | 0.122 | **0.14pp** [−0.10, 0.38] | 0.179pp | 1.3× |
 | **24h** | **0.063** | **0.07pp** [−0.15, 0.28] | **0.093pp** | **1.3×** |
 | 48h | 0.031 | **0.06pp** [−0.14, 0.26] | 0.046pp | 0.8× |
 
@@ -2949,14 +3015,14 @@ amplitude (one run, one snapshot).
 
 | τ(h) | `L` med | `L` p95 | published `b` (4×) | per match (pp) | over gate | **safe `b`** | worst at its low end | `b` proven over |
 | --- | --- | --- | --- | --- | --- | --- | --- | --- |
-| 6 | 2.68 | 11.80 | 3.716 | 0.278 | **32/52** | 0.1633–0.1669 | 3.64pp | 1.8581 |
-| 12 | 4.19 | 16.18 | 3.291 | 0.182 | **12/63** | 0.1639–0.1671 | 3.42pp | 1.6455 |
-| **24** | **8.46** | **26.21** | **2.679** | **0.093** | **0/66** | **0.1988–0.2014** | **3.86pp** | — |
-| 48 | 17.55 | 45.94 | 2.294 | 0.046 | **0/61** | 0.2061–0.2084 | 3.56pp | — |
+| 6 | 2.74 | 11.51 | 3.565 | 0.272 | **33/57** | 0.1497–0.1532 | 3.56pp | 1.7827 |
+| 12 | 4.28 | 15.56 | 3.111 | 0.179 | **12/55** | 0.1610–0.1641 | 3.34pp | 1.5557 |
+| **24** | **8.53** | **25.31** | **2.574** | **0.093** | **0/57** | **0.1911–0.1936** | **3.43pp** | — |
+| 48 | 17.58 | 43.67 | 2.187 | 0.046 | **0/51** | 0.1922–0.1943 | 3.50pp | — |
 
-Three readings. **`L` is near-linear in τ** (2.68 → 17.55), so §3-6's `L* = λτ` holds on the
+Three readings. **`L` is near-linear in τ** (2.74 → 17.58), so §3-6's `L* = λτ` holds on the
 timeline too and changing τ means rebuilding the whole table. **Longer τ makes a match cheaper
-and the field fairer** — marginal cost 0.278 → 0.046pp, cells over **32 → 0** — because a
+and the field fairer** — marginal cost 0.272 → 0.046pp, cells over **33 → 0** — because a
 longer memory dilutes both the single match and the heterogeneity between players; a short τ
 buys "one match hurts" at the price of unfairness, and here that price is steep. **All four
 have a safe amplitude**, so fairness is closed by the AMPLITUDE, not by τ; τ is the "how much
@@ -2965,11 +3031,11 @@ does one match hurt" dial.
 **And the bisection errs in one direction under noise.** One unlucky evaluation above 5 near
 the threshold means everything above it is never revisited, while an unlucky PASS is caught by
 the independent-seed validation — the asymmetry is structural. This run's τ=24h trace shows the
-scale (4.97pp at b=0.1988, 5.14pp at b=0.2014, 1.3% higher). The published amplitude is
+scale (4.92pp at b=0.1911, 5.04pp at b=0.1936, 1.3% higher). The published amplitude is
 therefore a LOWER bound on the threshold, and the error runs toward the safe side. And the
 right-hand endpoint is NOT a confidence limit: "not established below" is not "over" — that
 would need a LOWER bound above 5. Across this sweep two candidates cleared that bar and survived an independent-seed confirmation
-(τ=6h at b=1.8581 and τ=12h at b=1.6455), and only those thresholds acquire an upper bound.
+(τ=6h at b=1.7827 and τ=12h at b=1.5557), and only those thresholds acquire an upper bound.
 
 **Whether the weight decides the verdict is measured too.** Re-counting the same cells at
 knockout weights of 5%, 8.1% (canonical), 10% and 15%: at τ=24h cells over run **0·0·0·0** —
@@ -2979,28 +3045,29 @@ dependence grows is recorded rather than assumed away.
 The safe `b` is a **bracket**, not a point: the largest candidate that passed and the smallest
 that failed. And that bracket is SEARCH GRANULARITY, not a confidence interval — each
 candidate's bound is itself estimated from 6,000 seeds PER OPPONENT. This run's lower endpoints
-increase (0.163 · 0.164 · 0.199 · 0.206), but the procedure does not test ordering across τ, so
+increase (0.150 · 0.161 · 0.191 · 0.192), but the procedure does not test ordering across τ, so
 that shape must not be read as structural. The
 bisection errs in ONE DIRECTION under noise: a single unlucky evaluation above 5 near the
 threshold means everything above it is never revisited, while an unlucky pass is caught by the
-independent-seed validation. τ=24h's trace shows the scale — 4.97pp at b=0.1988 against 5.14pp
-at b=0.2014, 1.3% higher. So the published value is a LOWER BOUND on the threshold, and the
+independent-seed validation. τ=24h's trace shows the scale — 4.92pp at b=0.1911 against 5.04pp
+at b=0.1936, 1.3% higher. So the published value is a LOWER BOUND on the threshold, and the
 direction it errs in is the safe one. **The right endpoint is not an upper bound either** —
 "not established below" is not "over". What is established is that every τ in this range has an
-amplitude in the reported safe range that puts all cells under the gate, and that cells over fall 32 → 0
+amplitude in the reported safe range that puts all cells under the gate, and that cells over fall 33 → 0
 with τ.
 
 **The sweep caught two defects in the gate** that a single τ could never have shown. First,
 capping a degenerate sample's bound at infinity made cells where the amplitude is small enough
 that the two variants coincide fail the search outright — τ=6, 12 and 48 all reported "no safe
 amplitude" for that reason, and τ=24 only appeared to succeed because it happened to land
-ABOVE that point. The earlier τ=24 result was luck, not structure. A rule-of-three bound
-(`c + (−ln α / n)·(max − c)`) fixes it and all four then resolve. Second, five bisection steps
+ABOVE that point. The earlier τ=24 result was luck, not structure. The empirical-Bernstein bound
+used now retains a bounded-support range term even when SD is zero, and the probe prints that
+finite **degenerate upper-bound floor** before searching; all four then resolve. Second, five bisection steps
 were too coarse: τ=6 stalled at 5.35pp on b=0.122 and that was printed as "none exists", which
 is search depth, not a finding. Ten steps now, with an early exit on the first failing cell to
 pay for them.
 
-### Testing the vector against the gate instead of borrowing a ceiling
+### Testing the shadow-load vector against the gate instead of borrowing a ceiling
 
 E applies the same Δ to all eleven; fatigue makes a vector. An earlier version averaged
 engagement by ARRAY INDEX across every formation and applied that to a synthetic 4-3-3 — but
@@ -3009,7 +3076,8 @@ vector. Worse, it described the result as "played against its fresh copy" while
 each variant against a rotating synthetic field and subtracted field-relative win rates.
 Match outcomes are nonlinear and opponent-dependent; that difference is not a head-to-head.
 
-Corrected: each team state is turned into a fresh and a tired variant, and BOTH are played
+Corrected: each committed XI plus its synchronized zero-growth shadow-load state is turned into
+a fresh and a tired variant, and BOTH are played
 against EACH OF THE SIX FIXED REAL-OPPONENT BASKET MEMBERS as its own cell — never averaged or
 screened away — on the same seed,
 then differenced. **Each seed runs BOTH home and away; those two
@@ -3023,6 +3091,10 @@ in the window, folded on a hash of the fielded eleven rather than on `teamId`
 neutral offset `a` is applied to both sides, since the equation is `5 + a − bL/(L+L₀)` and
 omitting `a` overstates the gap by exactly that.
 
+Form was not deployed in this window, so the load vector did not historically exist. “Real” here
+means the committed XI, kickoff/claim chronology and opponent are historical; `L` is the
+zero-growth counterfactual integrated on that public chronology.
+
 **The opponent is a CELL AXIS — neither averaged nor screened.** The canonical gate
 (`band-worstcase.mts:161`) loops `for (A) for (B) for (growth) for (mode)`, so both the count of
 cells over and the choice of worst live on that grid. This got it wrong twice:
@@ -3032,8 +3104,8 @@ cells over and the choice of worst live on that grid. This got it wrong twice:
    the worst, so the error ran toward "safe".
 2. **Then screened.** Testing all 62 and confirming only the winner leaves the discarded 61 with
    NO guarantee and a family that never counts the opponent axis. Worse, that screen was NOISE.
-   This run measures one term's SD at a median of **24.0pp** (a win/draw/loss difference), so at
-   120 screen seeds an opponent mean has SE 2.19pp and the max−min of 62 draws is **≈9.4pp from
+   This run measures one term's SD at a median of **24.4pp** (a win/draw/loss difference), so at
+   120 screen seeds an opponent mean has SE 2.22pp and the max−min of 62 draws is **≈9.6pp from
    noise alone**. The 10.85pp range the previous round read as "the opponent axis matters" was
    SMALLER than what noise produces.
 
@@ -3046,7 +3118,7 @@ widens that floor by the same factor; dividing it produced "no safe amplitude" a
 which is arithmetic rather than a finding.
 
 **Bounding all 62 under 5pp is not affordable** — at that SD a 1.5pp half-width per opponent
-needs ≈4,092 seeds each, tens of times the cell budget across 62. So the basket is a SAMPLE of
+needs ≈4,222 seeds each, tens of times the cell budget across 62. So the basket is a SAMPLE of
 the 62, the same kind of limit the canonical gate accepts by using a designed grid. The question
 it leaves open — does an opponent outside the basket exceed? — is asked separately below.
 
@@ -3063,85 +3135,83 @@ things were wrong at once. Teams were ranked by *peak single load* while states
 With the currency fixed, the cut turned out not to exist:
 
 ```
-70363723 5.93 · 30178fc1 4.62 · 6097e735 4.45 · 704b8e45 4.34 · 6aec8bd8 4.31
-118f1f13 4.20 · 6a37ed0a 4.20 · 2aa911db 4.12 · 2b865779 4.01 · 23baeba2 3.74
-930b4313 3.58 · 796efbfe 3.34 · e8dc63e1 0.15        (summed deficit, max over curves)
+70363723 5.79 · 6097e735 4.74 · 704b8e45 4.58 · 30178fc1 4.54 · 118f1f13 4.45
+2aa911db 4.42 · 6a37ed0a 4.38 · 2b865779 4.34 · 6aec8bd8 4.30 · 930b4313 3.58
+23baeba2 3.58 · 796efbfe 3.43 · e8dc63e1 0.19        (summed deficit, max over curves)
 ```
 
-Second (4.62) through twelfth (3.34) sit close together, so "the worst team" would
+Second (4.74) through twelfth (3.43) sit close together, so "the worst team" would
 have been decided by which three I picked. Nothing is cut now, and the simultaneous correction
 widened accordingly. (τ=24h below; other τ in the sweep above.)
 
+The columns are folded deliberately. `ladder` and `knockout` are the mode point estimates for
+the opponent with the largest upper endpoint; `weighted` is the maximum exposure-weighted point
+estimate across opponents; `lo` and `hi` are the maxima of the corresponding simultaneous
+endpoints. Those maxima can come from different opponents: together they bound the max functional,
+not one imaginary opponent's point estimate and interval. `state (UTC)` is the verified claim-clock
+proxy used by this integration, not the scheduled kickoff.
+
 | team | `L₀` | state (UTC) | max deficit | ladder | knockout | **weighted** | lo | hi | 5pp |
 | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
-| `70363723` | 1× | 2026-08-25 12:58 | 0.524 | 5.00 | 7.86 | **5.24** | 2.59 | 7.88 | undecided |
-| `70363723` | 1× | 2026-08-27 09:57 | 0.556 | 5.10 | 7.33 | **5.28** | 2.60 | 7.95 | undecided |
-| `70363723` | 2× | 2026-08-25 11:55 | 0.684 | 6.30 | 9.09 | **6.52** | 3.72 | 9.33 | undecided |
-| `70363723` | 2× | 2026-08-27 09:57 | 0.739 | 6.32 | 8.86 | **6.53** | 3.76 | 9.29 | undecided |
-| `70363723` | 4× | 2026-08-25 11:55 | 0.834 | 6.78 | 10.16 | **7.05** | 4.16 | 9.94 | undecided |
-| `70363723` | 4× | 2026-08-27 09:57 | 0.925 | 6.71 | 10.45 | **7.02** | 4.14 | 9.89 | undecided |
-| `30178fc1` | 1× | 2026-08-31 12:56 | 0.404 | 4.55 | 7.14 | **4.76** | 2.16 | 7.35 | undecided |
-| `30178fc1` | 1× | 2026-08-27 07:26 | 0.434 | 5.47 | 7.74 | **5.65** | 3.00 | 8.31 | undecided |
-| `30178fc1` | 2× | 2026-08-31 12:56 | 0.471 | 5.24 | 8.66 | **5.52** | 2.84 | 8.19 | undecided |
-| `30178fc1` | 2× | 2026-08-27 07:26 | 0.518 | 5.34 | 7.91 | **5.55** | 2.84 | 8.26 | undecided |
-| `30178fc1` | 4× | 2026-08-31 12:56 | 0.524 | 6.25 | 8.27 | **6.41** | 3.67 | 9.16 | undecided |
-| `30178fc1` | 4× | 2026-08-27 07:26 | 0.588 | 5.93 | 8.35 | **6.13** | 3.36 | 8.89 | undecided |
-| `6097e735` | 1× | 2026-08-28 07:24 | 0.415 | 5.31 | 6.83 | **5.44** | 2.75 | 8.12 | undecided |
-| `6097e735` | 1× | 2026-08-31 12:56 | 0.441 | 4.22 | 5.88 | **4.36** | 1.75 | 6.96 | undecided |
-| `6097e735` | 2× | 2026-08-28 07:24 | 0.487 | 5.42 | 7.34 | **5.57** | 2.83 | 8.31 | undecided |
-| `6097e735` | 2× | 2026-08-31 12:56 | 0.530 | 4.61 | 6.84 | **4.79** | 2.11 | 7.47 | undecided |
-| `6097e735` | 4× | 2026-08-28 07:24 | 0.546 | 5.46 | 7.89 | **5.66** | 2.84 | 8.47 | undecided |
-| `6097e735` | 4× | 2026-08-31 12:56 | 0.605 | 5.04 | 7.67 | **5.26** | 2.54 | 7.97 | undecided |
-| `704b8e45` | 1× | 2026-08-28 07:26 | 0.470 | 4.47 | 6.35 | **4.62** | 2.05 | 7.19 | undecided |
-| `704b8e45` | 2× | 2026-08-28 07:26 | 0.578 | 5.14 | 6.71 | **5.27** | 2.61 | 7.93 | undecided |
-| `704b8e45` | 4× | 2026-08-28 07:26 | 0.673 | 5.19 | 7.49 | **5.38** | 2.65 | 8.10 | undecided |
-| `6aec8bd8` | 1× | 2026-08-23 07:26 | 0.438 | 5.52 | 6.53 | **5.60** | 2.90 | 8.29 | undecided |
-| `6aec8bd8` | 1× | 2026-08-26 07:26 | 0.466 | 5.02 | 6.25 | **5.12** | 2.44 | 7.80 | undecided |
-| `6aec8bd8` | 2× | 2026-08-23 07:26 | 0.524 | 5.95 | 7.61 | **6.08** | 3.30 | 8.87 | undecided |
-| `6aec8bd8` | 2× | 2026-08-26 07:26 | 0.571 | 5.70 | 7.40 | **5.84** | 3.10 | 8.58 | undecided |
-| `6aec8bd8` | 4× | 2026-08-23 07:26 | 0.596 | 6.19 | 7.85 | **6.32** | 3.50 | 9.14 | undecided |
-| `6aec8bd8` | 4× | 2026-08-26 07:26 | 0.663 | 5.94 | 7.38 | **6.06** | 3.28 | 8.83 | undecided |
-| `118f1f13` | 1× | 2026-08-30 07:24 | 0.468 | 3.78 | 6.70 | **4.02** | 1.54 | 6.50 | undecided |
-| `118f1f13` | 2× | 2026-08-30 07:24 | 0.576 | 4.42 | 7.42 | **4.66** | 2.09 | 7.24 | undecided |
-| `118f1f13` | 4× | 2026-08-30 07:24 | 0.669 | 5.18 | 7.98 | **5.41** | 2.78 | 8.04 | undecided |
-| `6a37ed0a` | 1× | 2026-08-25 07:26 | 0.441 | 4.17 | 6.13 | **4.33** | 1.76 | 6.89 | undecided |
-| `6a37ed0a` | 1× | 2026-08-28 07:26 | 0.459 | 4.04 | 5.30 | **4.14** | 1.59 | 6.70 | undecided |
-| `6a37ed0a` | 2× | 2026-08-25 07:26 | 0.530 | 4.94 | 7.10 | **5.12** | 2.46 | 7.77 | undecided |
-| `6a37ed0a` | 2× | 2026-08-28 07:26 | 0.560 | 4.42 | 6.39 | **4.58** | 1.94 | 7.22 | undecided |
-| `6a37ed0a` | 4× | 2026-08-25 07:26 | 0.605 | 5.50 | 7.58 | **5.67** | 2.99 | 8.36 | undecided |
-| `6a37ed0a` | 4× | 2026-08-28 07:26 | 0.647 | 4.75 | 7.67 | **4.99** | 2.35 | 7.62 | undecided |
-| `2aa911db` | 1× | 2026-08-29 07:26 | 0.404 | 4.04 | 5.57 | **4.17** | 1.61 | 6.72 | undecided |
-| `2aa911db` | 1× | 2026-08-30 07:26 | 0.452 | 3.92 | 5.66 | **4.07** | 1.52 | 6.61 | undecided |
-| `2aa911db` | 2× | 2026-08-29 07:26 | 0.471 | 4.24 | 6.51 | **4.43** | 1.86 | 6.99 | undecided |
-| `2aa911db` | 2× | 2026-08-30 07:26 | 0.548 | 4.63 | 6.02 | **4.74** | 2.15 | 7.32 | undecided |
-| `2aa911db` | 4× | 2026-08-29 07:26 | 0.524 | 4.93 | 6.97 | **5.10** | 2.45 | 7.74 | undecided |
-| `2aa911db` | 4× | 2026-08-30 07:26 | 0.630 | 5.06 | 6.86 | **5.20** | 2.55 | 7.86 | undecided |
-| `2b865779` | 1× | 2026-08-26 07:26 | 0.453 | 4.33 | 6.60 | **4.51** | 1.92 | 7.11 | undecided |
-| `2b865779` | 2× | 2026-08-26 07:26 | 0.549 | 5.02 | 7.92 | **5.26** | 2.58 | 7.94 | undecided |
-| `2b865779` | 4× | 2026-08-26 07:26 | 0.632 | 4.90 | 8.32 | **5.18** | 2.49 | 7.88 | undecided |
-| `23baeba2` | 1× | 2026-08-31 12:56 | 0.405 | 4.06 | 5.37 | **4.17** | 1.16 | 7.18 | undecided |
-| `23baeba2` | 1× | 2026-08-29 07:06 | 0.409 | 3.61 | 5.93 | **3.80** | 1.25 | 6.34 | undecided |
-| `23baeba2` | 2× | 2026-08-31 12:56 | 0.472 | 4.85 | 6.32 | **4.97** | 1.85 | 8.08 | undecided |
-| `23baeba2` | 2× | 2026-08-29 07:06 | 0.478 | 4.30 | 5.85 | **4.43** | 1.81 | 7.05 | undecided |
-| `23baeba2` | 4× | 2026-08-31 12:56 | 0.525 | 4.75 | 6.79 | **4.91** | 1.73 | 8.10 | undecided |
-| `23baeba2` | 4× | 2026-08-29 07:06 | 0.534 | 5.04 | 5.46 | **5.08** | 2.42 | 7.73 | undecided |
-| `930b4313` | 1× | 2026-08-23 07:01 | 0.383 | 3.77 | 5.64 | **3.92** | 1.38 | 6.46 | undecided |
-| `930b4313` | 1× | 2026-08-24 07:14 | 0.429 | 3.62 | 6.07 | **3.82** | 1.29 | 6.35 | undecided |
-| `930b4313` | 2× | 2026-08-23 07:01 | 0.439 | 4.14 | 6.92 | **4.37** | 1.78 | 6.95 | undecided |
-| `930b4313` | 2× | 2026-08-24 07:14 | 0.510 | 4.19 | 5.71 | **4.32** | 1.76 | 6.87 | undecided |
-| `930b4313` | 4× | 2026-08-23 07:01 | 0.482 | 4.85 | 6.92 | **5.02** | 2.41 | 7.62 | undecided |
-| `930b4313` | 4× | 2026-08-24 07:14 | 0.577 | 4.35 | 6.38 | **4.51** | 1.94 | 7.09 | undecided |
-| `796efbfe` | 1× | 2026-08-25 07:06 | 0.351 | 3.59 | 6.89 | **3.86** | 1.44 | 6.28 | undecided |
-| `796efbfe` | 1× | 2026-08-28 06:58 | 0.367 | 2.98 | 4.89 | **3.13** | 0.80 | 5.46 | undecided |
-| `796efbfe` | 2× | 2026-08-25 07:06 | 0.392 | 3.65 | 8.03 | **4.01** | 1.53 | 6.48 | undecided |
-| `796efbfe` | 2× | 2026-08-28 06:58 | 0.415 | 3.06 | 4.94 | **3.21** | 0.83 | 5.60 | undecided |
-| `796efbfe` | 4× | 2026-08-25 07:06 | 0.422 | 4.03 | 7.65 | **4.32** | 1.80 | 6.84 | undecided |
-| `796efbfe` | 4× | 2026-08-28 06:58 | 0.451 | 3.21 | 5.15 | **3.37** | 0.99 | 5.75 | undecided |
-| `e8dc63e1` | 1× | 2026-08-24 02:05 | 0.017 | 0.20 | 0.25 | **0.20** | -1.39 | 1.79 | **below** |
-| `e8dc63e1` | 2× | 2026-08-24 02:05 | 0.015 | 0.15 | 0.25 | **0.17** | -1.35 | 1.72 | **below** |
-| `e8dc63e1` | 4× | 2026-08-24 02:05 | 0.015 | 0.15 | 0.03 | **0.17** | -1.34 | 1.69 | **below** |
-
-**None of the 66 cells is over**; 3 are established below and 63 undecided. So many undecided is
+| `70363723` | 1× | 2026-08-26 13:57 | 0.531 | 5.09 | 7.49 | **5.28** | 2.63 | 7.94 | undecided |
+| `70363723` | 1× | 2026-08-27 09:57 | 0.551 | 4.51 | 6.53 | **4.67** | 2.04 | 7.30 | undecided |
+| `70363723` | 2× | 2026-08-26 13:57 | 0.698 | 6.21 | 8.38 | **6.39** | 3.63 | 9.15 | undecided |
+| `70363723` | 2× | 2026-08-27 09:57 | 0.737 | 5.82 | 8.48 | **6.04** | 3.30 | 8.78 | undecided |
+| `70363723` | 4× | 2026-08-26 13:57 | 0.865 | 7.04 | 8.99 | **7.20** | 4.33 | 10.06 | undecided |
+| `70363723` | 4× | 2026-08-27 09:57 | 0.932 | 6.75 | 10.07 | **7.02** | 4.16 | 9.88 | undecided |
+| `6097e735` | 1× | 2026-08-28 07:24 | 0.447 | 5.47 | 6.63 | **5.57** | 2.88 | 8.26 | undecided |
+| `6097e735` | 2× | 2026-08-28 07:24 | 0.543 | 4.90 | 7.98 | **5.15** | 2.44 | 7.87 | undecided |
+| `6097e735` | 4× | 2026-08-28 07:24 | 0.627 | 5.80 | 9.08 | **6.06** | 3.24 | 8.89 | undecided |
+| `704b8e45` | 1× | 2026-08-28 07:26 | 0.480 | 4.53 | 6.38 | **4.68** | 2.07 | 7.28 | undecided |
+| `704b8e45` | 2× | 2026-08-28 07:26 | 0.602 | 5.17 | 7.59 | **5.37** | 2.66 | 8.07 | undecided |
+| `704b8e45` | 4× | 2026-08-28 07:26 | 0.713 | 5.15 | 6.96 | **5.29** | 2.52 | 8.06 | undecided |
+| `30178fc1` | 1× | 2026-08-31 12:56 | 0.390 | 4.90 | 6.75 | **5.05** | 2.44 | 7.67 | undecided |
+| `30178fc1` | 1× | 2026-08-27 07:26 | 0.413 | 4.78 | 6.76 | **4.94** | 2.31 | 7.57 | undecided |
+| `30178fc1` | 2× | 2026-08-31 12:56 | 0.452 | 5.52 | 8.13 | **5.73** | 3.05 | 8.42 | undecided |
+| `30178fc1` | 2× | 2026-08-27 07:26 | 0.487 | 5.71 | 8.00 | **5.90** | 3.23 | 8.57 | undecided |
+| `30178fc1` | 4× | 2026-08-31 12:56 | 0.500 | 5.77 | 8.46 | **5.99** | 3.26 | 8.72 | undecided |
+| `30178fc1` | 4× | 2026-08-27 07:26 | 0.548 | 5.59 | 8.86 | **5.85** | 3.10 | 8.60 | undecided |
+| `118f1f13` | 1× | 2026-08-30 07:24 | 0.452 | 4.21 | 6.38 | **4.39** | 1.86 | 6.91 | undecided |
+| `118f1f13` | 2× | 2026-08-30 07:24 | 0.552 | 4.30 | 7.71 | **4.58** | 2.00 | 7.16 | undecided |
+| `118f1f13` | 4× | 2026-08-30 07:24 | 0.639 | 5.08 | 8.50 | **5.35** | 2.72 | 7.99 | undecided |
+| `2aa911db` | 1× | 2026-08-29 07:26 | 0.392 | 4.00 | 6.04 | **4.17** | 1.62 | 6.72 | undecided |
+| `2aa911db` | 1× | 2026-08-30 07:26 | 0.451 | 3.91 | 6.33 | **4.11** | 1.57 | 6.65 | undecided |
+| `2aa911db` | 2× | 2026-08-29 07:26 | 0.455 | 4.99 | 7.36 | **5.18** | 2.56 | 7.81 | undecided |
+| `2aa911db` | 2× | 2026-08-30 07:26 | 0.551 | 4.46 | 6.76 | **4.65** | 2.03 | 7.27 | undecided |
+| `2aa911db` | 4× | 2026-08-29 07:26 | 0.505 | 4.67 | 6.22 | **4.80** | 2.16 | 7.44 | undecided |
+| `2aa911db` | 4× | 2026-08-30 07:26 | 0.637 | 5.19 | 6.53 | **5.30** | 2.63 | 7.97 | undecided |
+| `6a37ed0a` | 1× | 2026-08-25 07:26 | 0.464 | 4.50 | 6.74 | **4.69** | 2.11 | 7.27 | undecided |
+| `6a37ed0a` | 2× | 2026-08-25 07:26 | 0.573 | 4.54 | 7.83 | **4.81** | 2.16 | 7.45 | undecided |
+| `6a37ed0a` | 4× | 2026-08-25 07:26 | 0.671 | 5.08 | 7.70 | **5.29** | 2.60 | 7.98 | undecided |
+| `2b865779` | 1× | 2026-08-24 07:24 | 0.407 | 4.33 | 6.71 | **4.52** | 1.91 | 7.14 | undecided |
+| `2b865779` | 1× | 2026-08-26 07:26 | 0.476 | 4.15 | 6.59 | **4.35** | 1.76 | 6.94 | undecided |
+| `2b865779` | 2× | 2026-08-24 07:24 | 0.479 | 4.98 | 8.32 | **5.25** | 2.55 | 7.96 | undecided |
+| `2b865779` | 2× | 2026-08-26 07:26 | 0.594 | 5.05 | 7.41 | **5.25** | 2.57 | 7.92 | undecided |
+| `2b865779` | 4× | 2026-08-24 07:24 | 0.537 | 5.90 | 7.50 | **6.03** | 3.27 | 8.79 | undecided |
+| `2b865779` | 4× | 2026-08-26 07:26 | 0.702 | 5.03 | 7.99 | **5.27** | 2.58 | 7.96 | undecided |
+| `6aec8bd8` | 1× | 2026-08-23 07:26 | 0.464 | 5.22 | 6.33 | **5.31** | 2.62 | 8.00 | undecided |
+| `6aec8bd8` | 2× | 2026-08-23 07:26 | 0.574 | 5.94 | 7.42 | **6.06** | 3.29 | 8.83 | undecided |
+| `6aec8bd8` | 4× | 2026-08-23 07:26 | 0.671 | 6.56 | 7.88 | **6.67** | 3.83 | 9.51 | undecided |
+| `930b4313` | 1× | 2026-08-23 07:01 | 0.388 | 3.91 | 5.80 | **4.06** | 1.56 | 6.56 | undecided |
+| `930b4313` | 1× | 2026-08-24 07:14 | 0.406 | 3.96 | 5.01 | **4.04** | 1.52 | 6.57 | undecided |
+| `930b4313` | 2× | 2026-08-23 07:01 | 0.448 | 4.33 | 5.73 | **4.44** | 1.88 | 7.00 | undecided |
+| `930b4313` | 2× | 2026-08-24 07:14 | 0.477 | 3.81 | 5.81 | **3.98** | 1.45 | 6.50 | undecided |
+| `930b4313` | 4× | 2026-08-26 07:16 | 0.461 | 4.58 | 6.69 | **4.76** | 2.19 | 7.32 | undecided |
+| `930b4313` | 4× | 2026-08-24 07:14 | 0.535 | 3.99 | 5.76 | **4.14** | 1.56 | 6.71 | undecided |
+| `23baeba2` | 1× | 2026-08-31 12:56 | 0.371 | 4.15 | 4.83 | **4.21** | 1.27 | 7.14 | undecided |
+| `23baeba2` | 1× | 2026-08-28 07:14 | 0.411 | 3.85 | 5.47 | **3.98** | 1.42 | 6.54 | undecided |
+| `23baeba2` | 2× | 2026-08-31 12:56 | 0.422 | 4.55 | 6.22 | **4.69** | 1.63 | 7.74 | undecided |
+| `23baeba2` | 2× | 2026-08-28 07:14 | 0.485 | 4.33 | 5.77 | **4.44** | 1.84 | 7.04 | undecided |
+| `23baeba2` | 4× | 2026-08-31 12:56 | 0.462 | 4.82 | 5.66 | **4.89** | 1.74 | 8.03 | undecided |
+| `23baeba2` | 4× | 2026-08-28 07:14 | 0.545 | 4.67 | 6.30 | **4.81** | 2.19 | 7.43 | undecided |
+| `796efbfe` | 1× | 2026-08-25 07:06 | 0.351 | 4.03 | 7.20 | **4.28** | 1.81 | 6.75 | undecided |
+| `796efbfe` | 2× | 2026-08-25 07:06 | 0.394 | 3.84 | 7.49 | **4.14** | 1.64 | 6.63 | undecided |
+| `796efbfe` | 4× | 2026-08-25 07:06 | 0.426 | 3.92 | 8.01 | **4.26** | 1.76 | 6.76 | undecided |
+| `e8dc63e1` | 1× | 2026-08-24 02:05 | 0.026 | 0.27 | 0.35 | **0.27** | -1.35 | 1.91 | **below** |
+| `e8dc63e1` | 2× | 2026-08-24 02:05 | 0.024 | 0.25 | 0.33 | **0.26** | -1.35 | 1.88 | **below** |
+| `e8dc63e1` | 4× | 2026-08-24 02:05 | 0.022 | 0.12 | 0.28 | **0.16** | -1.37 | 1.72 | **below** |
+
+**None of the 57 cells is over**; 3 are established below and 54 undecided. So many undecided is
 what the guarantee cost — the family is **1,836** = (78+78+78+72) × 6 opponents.
 
 **The state axis had the same defect.** The previous round chose each team's worst state by a
@@ -3149,13 +3219,13 @@ low-seed screen — exactly what had just been thrown out on the opponent axis:
 carry no guarantee, the family never counts that axis, and the screen itself is noise. The state
 is now chosen by two DETERMINISTIC rules — largest summed deficit (how tired the team is overall)
 and largest single-slot deficit (a vector concentrated on one place) — and BOTH are measured. At
-τ=24h the two rules pointed at different states in **27 of 39** (team, L₀) pairs.
+τ=24h the two rules pointed at different states in **18 of 39** (team, L₀) pairs.
 
 **The previous round's "1 over" is withdrawn.** It came from screening 62 opponents, then
 correcting the winner against a family of 153 — the search over 62 was never counted. Corrected
 for it (α = 0.05/18972), nothing at τ=24h is established over. **"none over" and "all established
 below" remain different claims** — the first says exceedance cannot be established, the second
-says non-exceedance can — and the 63 undecided are neither.
+says non-exceedance can — and the 54 undecided are neither.
 
 **At the short τ, however, an exceeding matchup does exist.** Since the basket is a sample of the
 62, "does an opponent outside it exceed?" is a separate question, asked separately: screen all 62
@@ -3164,45 +3234,30 @@ count (12,000), and correct for having searched 62.
 
 | τ | (cell, opponent) established over | worst |
 | --- | --- | --- |
-| 6h | **23** | `30178fc1` 4× — weighted 13.32pp, lower bound 9.59pp |
-| 12h | **5** | `6aec8bd8` 4× — weighted 9.81pp, lower bound 6.36pp |
+| 6h | **29** | `30178fc1` 4× — weighted 13.82pp, lower bound 10.10pp |
+| 12h | **10** | `30178fc1` 4× — weighted 9.92pp, lower bound 6.48pp |
 | 24h · 48h | 0 | — |
 
-Inside the basket, the point estimates for one state spread by **1.98pp** median (4.14pp max) —
-the real measure of how much the opponent moves the verdict, and next to the ≈9.4pp noise range
+Inside the basket, the point estimates for one state spread by **2.05pp** median (4.03pp max) —
+the real measure of how much the opponent moves the verdict, and next to the ≈9.6pp noise range
 above it shows why a screen cannot rank them.
 
 This does NOT replace the verdict above: screening is drawn to noise, so it does not establish
 "we found the worst". What it establishes is an EXISTENCE claim — at τ=6h and 12h the published
-amplitude fails against a real matchup from this window. At τ=24h and 48h not even that stands. A cell is (team 13 × curve 3), and the simultaneous family
+amplitude fails against a real matchup from this window. At τ=24h and 48h not even that stands. A displayed cell is (team × curve × deterministic state rule), and the simultaneous family
 is the SUM of each τ's cell bounds times the opponent basket, **(78+78+78+72) × 6 = 1,836**
 (τ=48h loses one team to the longer warm-up) — the document's claim spans every τ, and the counts differ per τ. A
-cell is (team × curve) and the two modes are mixed **inside** it by exposure, because that is
+cell keeps six opponent-specific bounds, the family counts all six and the verdict takes their worst;
+the two modes are mixed **inside each opponent bound** by exposure, because that is
 how the gate is defined. The per-mode figures sit beside it because they show WHERE the effect
 lives: knockout point estimates run about 1.5× ladder's. Holding each mode to 5pp separately
 made 38 of 78 look over, and that was the earlier "almost every team, in knockout" conclusion.
 **It is withdrawn — it was measured against the wrong bar.**
 
-**Choosing the state by summed deficit was also wrong.** The sum ignores WHICH slot is tired,
-and influence differs by slot (§3-16's ZP fallback slot is the example) and by mode. Every one
-of a team's states is now measured on 200 seeds with COMMON RANDOM NUMBERS across states, and
-the worst weighted edge wins. That choice differed from the summed-deficit pick in **29 of 39**
-cells.
-
-**But "we found the worst" is not claimable — that was measured too.** The gap between first
-and second place is **0.16pp** median (min 0.00), while the gap between first and last is
-**4.41pp** median (min 0.00). The range is some 27 times the margin, so the screen carries
-real signal and **only the top is tied** — the difference between what was chosen and the true
-worst is about the margin, ≈0.2pp, which cannot move a 5pp gate. It also reframes the
-disagreement with the summed-deficit pick: not "the proxy was badly wrong" but "the top states
-are tied, so any rule picks one of several equals".
-
-The selection comes from an outcome, so winner's-curse bias remains. That is why the states are
-**re-screened at the chosen amplitude** before validation, then measured on independent seeds:
-freezing the selection at the published amplitude would leave validation unable to see a state
-that is worse at a much smaller `b`, because the response is nonlinear. The bisection itself
-runs on the frozen selection as a SEARCH HEURISTIC; only the published verdict is exact at its
-own amplitude, and if validation fails the amplitude halves and the whole step repeats.
+Those two rules never inspect an outcome, so the state's winner-selection bias is gone. The
+guarantee still attaches only to the one or two states they choose: the remaining post-warm-up
+states and everything outside this window have no bound. This is a sampled state basket, not a
+structural upper bound, just as the six opponents are a sample of the 62.
 
 **Those repeats spend a divided alpha budget.** Up to five independently seeded confirmations
 that accept the FIRST to pass, each spending the full `0.05/1836`, would inflate the chance of
@@ -3230,11 +3285,12 @@ unchanged).
 
 **The "last lineup" machinery disappeared as a side effect.** The lineup now comes from the
 same state as the loads, so loads can never be applied to a player who has since left — the
-mismatch that machinery existed to prevent is now unrepresentable. Each state is the worst
-OBSERVED one per team and `L₀` (**649** post-warm-up candidates = that team's appearances),
-chosen on summed deficit, before a single match is played. `b` is a common factor of the
-deficits so it cannot reorder them — the bisection tests the same state throughout. `L₀` can
-and does reorder them: `70363723`'s worst kickoff differs at 1×, 2× and 4×.
+mismatch that machinery existed to prevent is now unrepresentable. For each team and `L₀`, one
+or two states are selected from **649** post-warm-up public-fixture/shadow-load candidates by
+the two deterministic rules (largest summed deficit and largest single-slot deficit), before a
+single match is played. `b` is a common factor of the deficits so it cannot reorder them — the
+bisection tests the same state throughout. `L₀` can reorder candidates, so the state basket is
+computed separately for every curve; repeated timestamps in this τ=24h table are not a general rule.
 
  Two things had been hiding that. The gate rebuilt the
 curve per team, which pins each team's MEDIAN deficit to 0.20 — the median column read
@@ -3248,23 +3304,28 @@ Exceeding the gate does not mean the mechanism fails; it means the amplitude is
 Bisect for the largest `b` whose worst cell stays under, measuring EVERY cell at each step
 rather than assuming linearity — a single proportional guess landed at 5.15, still grazing it.
 
-| `b` | cell hit / worst | upper | lower | max deficit | verdict |
+The named cell is the **witness for the maximum upper bound**. The lower column is maximized
+separately over the same scan and can have a different witness; the two endpoints are not one
+cell's paired interval. Passing rows and independent validation scan every cell. A failing row
+stops at the first upper-bound failure, so its extrema cover only that scanned prefix—which is
+already sufficient to reject the candidate.
+
+| `b` | max-upper witness | upper | max lower | max deficit | verdict |
 | --- | --- | --- | --- | --- | --- |
-| 1.3394 | `70363723` 1× | 13.38 | 3.51 | 0.856 | not established |
-| 0.6697 | `70363723` 1× | 8.66 | -0.12 | 0.428 | not established |
-| 0.3349 | `70363723` 1× | 5.70 | -1.67 | 0.214 | not established |
-| 0.1674 | `23baeba2` 1× | 4.84 | -2.28 | 0.083 | **below** |
-| 0.2511 | `70363723` 1× | 5.28 | -2.09 | 0.160 | not established |
-| 0.2093 | `30178fc1` 1× | 5.03 | -2.12 | 0.111 | not established |
-| 0.1884 | `6aec8bd8` 1× | 4.91 | -2.19 | 0.101 | **below** |
-| 0.1988 | `30178fc1` 1× | 4.97 | -2.12 | 0.098 | **below** |
-| 0.2041 | `30178fc1` 1× | 5.06 | -2.06 | 0.101 | not established |
-| 0.2014 | `30178fc1` 1× | 5.14 | -2.02 | 0.107 | not established |
-| **0.1988** (validated on independent seeds) | `6aec8bd8` 1× | **3.86** | — | 0.106 | **all cells below** |
-
-**`b` ∈ 0.1988–0.2014**, with all 66 cells established below the gate on independent seeds at
-the lower end — after re-screening the states at that amplitude — and the upper end being the
-smallest candidate that failed.
+| 1.2872 | `70363723` 1× | 13.26 | 3.39 | 0.862 | not established |
+| 0.6436 | `70363723` 1× | 7.58 | -0.81 | 0.431 | not established |
+| 0.3218 | `70363723` 1× | 6.21 | -1.54 | 0.215 | not established |
+| 0.1609 | `704b8e45` 1× | 4.71 | -2.22 | 0.097 | **below** |
+| 0.2414 | `70363723` 1× | 5.29 | -2.10 | 0.162 | not established |
+| 0.2011 | `6097e735` 1× | 5.01 | -2.22 | 0.113 | not established |
+| 0.1810 | `23baeba2` 1× | 4.98 | -2.17 | 0.085 | **below** |
+| 0.1911 | `6aec8bd8` 1× | 4.92 | -2.18 | 0.112 | **below** |
+| 0.1961 | `6097e735` 1× | 5.15 | -2.13 | 0.110 | not established |
+| 0.1936 | `70363723` 1× | 5.04 | -2.13 | 0.134 | not established |
+| **0.1911** (validated on independent seeds) | `30178fc1` 1× | **3.43** | — | 0.094 | **all cells below** |
+
+**`b` ∈ 0.1911–0.1936**, with all 57 deterministic shadow cells established below the gate on
+independent seeds at the lower end and the upper end being the smallest candidate that failed.
 
 **Do not read that bracket as precision.** The fail rows are not monotone even though the
 deficit falls with `b`, and the cell that trips changes between them: each bound is estimated
@@ -3275,13 +3336,14 @@ separate seed column. Note
 that the worst cell moves: at the published amplitude it is `L₀`=4×, but at low `b` it is
 **1×**, because a smaller `L₀` saturates sooner and so produces a larger deficit for the same
 `b`. Testing only the cell that was worst at the starting amplitude would have missed that. E's 0.2 was measured with a uniform Δ across all eleven, and
-putting it on the MEDIAN player is what let the top of a real vector (max deficit 0.02–0.88)
+putting it on the MEDIAN player is what let the top of a shadow vector (max deficit 0.02–0.93)
 run past the gate. That value is the worst over all 13 teams that played, but NOT over every appearance: the
 state comes from two deterministic rules (largest summed deficit, largest single-slot deficit)
 applied to the 649 post-warm-up candidates, and only those one or two states per (team, curve)
 carry a bound. The remaining appearances are unmeasured — the same kind of limit the opponent
 basket carries. What is gone is selection on an OUTCOME; what is not gone is that this is a
-sample of the observed states, and everything outside this window is unseen entirely.
+sample of shadow states on the observed public fixture chronology, and everything outside this
+window is unseen entirely.
 
 **The bound's support constant was wrong, too.** One `ser` term averages the home and away
 versions of `(win(fresh) − win(tired))`, then multiplies by `100 · 2`. Each difference lies in
@@ -3309,7 +3371,7 @@ almost every observation is 0:
 
 The price is width. At this family, a half-width of 2.9pp needs **12,000 seeds per opponent**;
 at 6,000 it is 4.9pp, which would leave the 5pp gate no room at all. So the seed budget per
-opponent went up 4×, the safe amplitude fell from 0.15 to 0.13, and only 3 of 66 cells are
+opponent went up 4×, the safe amplitude fell from 0.15 to 0.13, and only 3 of 57 cells are
 established below at the published amplitude. Every one of those is the cost of not borrowing
 normality.
 
@@ -3332,7 +3394,7 @@ calls `applyResult` for both; `selectOpponentByStrength` draws from the five str
 in-division; the cooldown is initiator-only. Observed at-appearance `L` reaches 55 for an
 honest player while a team taking the same match count as the away side accumulates zero. That
 is exemption, not a ratio — bounded only by what bounds the feature, since the prize is 0.59pp at g=0.4 (0.16pp at the τ=24h safe amplitude)
-and the team-level gap is 3.86pp at the safe amplitude, under the gate.
+and the team-level gap is 3.43pp at the safe amplitude, under the gate.
 
 ### What this does not settle
 
@@ -3367,8 +3429,9 @@ gate either, so such teams are excluded outright and the output says how many. N
 this window.
 
 The prize-versus-cost ordering AT THE AMPLITUDE THAT WOULD SHIP — it splits at `g=0.4`
-(−0.43pp [−0.76, −0.10]) but not at the safe amplitude's own deficit (prize 0.16pp, cheapest
-cost 0.16pp, paired difference −0.00pp [−0.47, 0.46]), and that second one is the decision. The neutral-point fork. Maximum LEGAL role concentration (the
+(−0.43pp [−0.76, −0.10]) but not at the safe amplitude's own deficit (0.133; prize 0.16pp
+[−0.09, 0.40], cheapest cost 0.16pp, paired difference 0.00pp [−0.47, 0.47]), and that second
+one is the decision. The neutral-point fork. Maximum LEGAL role concentration (the
 replay is what happened, not what the rules permit). Player identity is the snapshot's `actorIds` playerId, keyed GLOBALLY rather than per team —
 a minted player sold mid-window would otherwise get two chains starting from zero, understating
 the buyer's vector. A missing actor map or even one player that fails to join is rejected by the
@@ -3379,7 +3442,7 @@ condition is read on entry whether or not load is charged, and ownership moves o
 appearance. Both are right by construction, which is not the same as having been seen to
 run. All four τ values `{6, 12, 24, 48}h` are integrated separately, with the gate and safe-amplitude
 search rerun on that τ's own cells rather than converted from 24h. At τ=24h the published run has
-66 exposure-mixed cells: 0 established above the gate, 3 established below it, and 63 undecided.
+57 exposure-mixed cells: 0 established above the gate, 3 established below it, and 54 undecided.
 Snapshot coverage turned out to be a step function — 0% before 2026-08-21 and 100% after — so
 the window is 11 fully covered days rather than a 37% sample, and the script finds that step
 itself. The cost is that nothing with a period longer than 11 days is visible.
@@ -3401,9 +3464,12 @@ MATCHES=3000 SNAPSHOTS=/tmp/prod-tl.json REPLAYS=5 TAU_SWEEP=6,12,24,48 \
 still "the last completed UTC day", so an unqualified re-export extends it by a day — but
 `WINDOW_END` makes that date the window's (exclusive) end. The numbers in this section come
 from **08-21 through 08-31, 1,201 matches**, and re-exporting with `WINDOW_END=2026-09-01`
-reproduces that `matches` array **byte for byte** (verified) — which came for free from
-collapsing a window boundary that had been spelled out in eighteen separate places into one
-definition. The window's length remains a limitation.
+reproduces that `matches` array **byte for byte** (verified three times). The complete 0600 JSON
+used by this run has SHA-256
+`1c7335ba9fa11388363de57628207ecc4820dc9bb02285066bda849adf571b6f`; that is the anchor that
+prevents scalars from being mixed across exports. This reproducibility came from collapsing a
+window boundary that had been spelled out in eighteen separate places into one definition. The
+window's length remains a limitation.
 
 **"Fully covered" cannot be counted from completed rows alone.** A day with rows still
 unfinished, or whose cup bracket has not completed, is not fully covered — knockout rows are
@@ -3477,8 +3543,9 @@ has to carry which predicate it is conditional on.
 Without `TAU_SWEEP` only `TAU_H` runs. **Calling a single-τ run a "decision" makes a probe
 default the production recovery rate** — that is why the sweep exists.
 
-`REPLAYS` now applies only to the predicate table, where each fixture is replayed
-independently. The timeline is integrated once, from each match's stored `rng_seed`.
+`REPLAYS` now applies only to fixed-fixture summaries: predicate rates average the draws,
+while the one-match increment distribution keeps each integer draw as an observation. The
+timeline is integrated once, from each match's stored `rng_seed`, and is independent of it.
 
 **"`tsc` does not see these probes" is wrong, and it was checked.** #719 added
 `skill/reference/probes/**/*.mts` to `packages/mcp/tsconfig.typecheck.json`, and