feat(observers): Front O — the third observer, player-informed p (Phase V)

Three p-generators with disjoint information sets price the identical
favourite bet on 17,356 league matches (2018-2025; link fitted 2013-2018):
results-only Elo, the bookmaker (Shin de-vig), and the new players observer
(log market value of the actual starting lineup, Transfermarkt CC0 mirror;
state never reads a result or a price). Brier differs freely (book 0.230 <
elo 0.234 < players 0.235; calibration 0.012/0.016/0.032) but the SHAPE
residual - realised skew minus the law at realised frequency - is ~0.005
for all three. Calibration moves you along the law, never off it.

skewlib/players.py (+ committed tm_club_map/value_link), block 55, /odds
section 'The third observer', WC ledger players overlay (wc_squad_values.json
frozen, p_players/skew_players in the triangulation), FINDINGS Phase V,
Timeline V.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Author: valvesss

site/src/components/Timeline.astro

@@ -44,7 +44,8 @@ const phases = [
   { tag: 'S', kind: 'External validity', title: 'The law is the sport, not football', body: 'A canonical data layer makes the core sport-agnostic — it needs only (probability, odds, outcome) per bet. Plugging in tennis (tennis-data.co.uk, ATP+WTA 2005–2025, 62,865 matches) — a different sport, a two-outcome market with no draw, an independent odds source — and changing no analysis code, the signature reappears. The de-vig stays calibrated (mean favourite probability 0.688 ≈ realised win rate 0.692); the structural law holds, with the favourite bet most negative where the tournament is most lopsided (corr(skew, competitiveness) by tier = −1.00 ATP / −0.98 WTA, vs football’s −0.90; Grand Slams the steepest on both tours); and the underdog is lottery-like at +2.31, all but identical to football’s +2.35. The structural invariance is not an artefact of the 1X2 market or of football — it is a property of the sport as a competitive system.' },
   { tag: 'T', kind: 'External validity', title: 'A third sport, the same law', body: 'The same canonical layer absorbs a third sport with no change to the core: basketball (sportsbookreviewsonline.com, NBA 2007–08…2022–23, 19,621 games) — a two-outcome moneyline market with no draw, again an independent odds source. The signature reappears once more. The de-vig stays calibrated (mean favourite probability 0.694 ≈ realised win rate 0.685, the residual gap being the favourite–longshot bias itself); the structural law holds across seasons, with the favourite bet most negative in the most lopsided seasons (corr(skew, competitiveness) by season = −0.95, vs football’s −0.90 and tennis’s −1.00/−0.98); and the underdog is lottery-like at +2.61 ≈ football’s +2.35 and tennis’s +2.31. On one cross-sport curve, tennis and basketball overlap in the same lopsided regime that football traces at lower competitiveness — three sports, three markets (1X2, match-odds, moneyline), three odds sources, one structural law.' },
   { tag: 'U', kind: 'Equivalence', title: 'No drift — proven, not just unrefuted', body: 'The "no secular trend" claim rested on a high p-value (β = +0.00015/yr, p = 0.73) — but failing to reject a trend is not evidence of its absence. We close the gap with an equivalence test (two one-sided tests). Pre-registering the same margin used for the similarity analysis — half a between-league standard deviation (Δ = 0.026), read as the largest twenty-year drift worth caring about — the slope’s 90% confidence interval falls entirely inside [−Δ, +Δ], rejecting any larger drift at p = 0.006. It holds under a league-cluster bootstrap (p = 0.005) and a balanced fixed-basket panel (p = 0.043), and tightens as the margin relaxes. The skewness is not merely un-shown to drift; it is statistically equivalent to no drift, to within a fraction of one league-to-league difference over the entire record — turning absence of evidence into evidence of absence.' },
-];
+
+  { tag: 'V', kind: 'Observer invariance', title: 'Three ways of knowing, one geometry', body: 'A third observer joins the triangulation, built from a disjoint information set: team strength = log market value of the actual starting lineup (Transfermarkt), its state never reading a result or a price. Three engines — results-only Elo, the bookmaker, and the player observer — price the identical favourite bet on 17,356 league matches (2018–2025, link fitted on 2013–2018). They are not equally good forecasters: the market is sharpest (Brier 0.230) and the player observer the most miscalibrated (level error 2.7× the book\'s). But the shape residual — realised skew minus the law at the realised frequency — is ~0.005 for all three, indistinguishable from zero. Calibration differs freely; the shape does not. Knowing more moves an observer along the law, never off it: the strong form of observer-invariance.' },];
 ---
 <ol class="timeline">
   {phases.map(p => (