The top-player tier lists behind the comparison layer — a check on the
performance ranking, never blended into the composite. How each reads the field is
the kind of single-number call the methodology
library takes apart; the Index keeps them separate.
Win rate is Wilson-adjusted — did it win. Top cut is its share of top-4 finishers over its share of the field, in 16+ player events — does it convert to trophies. Blowout is mean battle points per game (0–20) — winning big, losing small. Consistency is the inverse of game-to-game variance — steady vs boom-bust. The sliders weight them; the composite re-ranks live. See the methodology note below for the exact definitions.
WHAT DO THE Z-SCORES MEAN?
Each facet number is a z-score — a comparison to the whole competitive roster this window. Zero is the roster average; +1.0 is one standard deviation above (roughly the top 16% of factions on that facet), +2.0 is exceptional; the scale mirrors going down. The composite is the weighted blend of the four. A faction under the game-count floor keeps its number but is held in the AVERAGE band with a “THIN” chip — a short, lucky sample shouldn’t front the tier list.
TOP PLAYERS — THE COMPARISON LAYER
The Top players column shows the curated top-player tier-list rating (S+ through F), recency-weighted across in-window expert publications. It is never blended into the composite — it sits beside it so you can see where results and reputation diverge. A RESULTS ▲ REP chip flags a faction winning more than top players rate it; HYPE ▲ RESULTS the reverse. Switch the VIEW toggle to order the table by top-player tier instead, with the composite rank as the comparison.
COMPOSITE
Z-SCORES ACROSS THE FACTION ROSTER · RANGE BARS ARE 95% CI · ALPHABETICAL ORDER
WEIGHTS · ADJUST AND WATCH THE COMPOSITE RE-RANK
WIN RATE 40%
TOP CUT 30%
BLOWOUT 20%
CONSISTENCY 10%
Win rate leads at 40% because winning is the baseline; top-cut conversion, margin, and consistency refine how a faction wins. Adjust to reflect what matters to you — the defaults are an editorial position, not a measurement.
COPIEDShare this reading — a custom-weights link opens at your weights.
FORMAT:
VIEW:
Singles is the honest matchup read; the team view is captain-influenced (pairings are assigned, not drawn). Top players is the comparison layer — expert tier lists, never folded into the composite.
Bands by Composite z-score (the performance-facet weighted blend):
STRONGER · z > +0.5 AVERAGE · −0.5 to +0.5 WEAKER · z < −0.5THIN under 30 games — held in AVERAGE
Ordered by top-player tier (recency-weighted top-player tier-list aggregation); the comparison column shows each faction's Performance composite rank:
S+ (5.5)S (5.0)A+ (4.5)A (4.0)A- (3.5)B (3.0)B- (2.5)C (2.0)C- (1.5)D (1.0)F (0.5)UNRATED · no tier data this window
VIEW:
PER-FACTION BREAKDOWN
COMPOSITE-SORTED · ZONE-PARTITIONED · RESHUFFLES LIVE ON WEIGHT CHANGE
#
FACTION
WIN RATE
TOP CUT
BLOWOUT
CONSISTENCY
COMPOSITE
TOP PLAYERS
Δ WK
Δ WK = change in composite z vs 2026-W29
(now 2026-W30), at the default weights —
▲ stronger, ▼ weaker; movement does not track your slider changes.
STRONGER · z > +0.5 AVERAGE · −0.5 to +0.5 WEAKER · z < −0.5
Click a faction above to see the full breakdown — the four performance facets, their z-scores and confidence, and the top-player comparison.
METHODOLOGY NOTE
Every number here is a facet of the Archive’s own tournament data. The composite is a weighted blend of four measures, each z-scored across the competitive roster so they combine on one scale:
Win rate — Wilson-adjusted (wins + ½·draws) / games. The baseline: did it win.
Top-cut conversion — the faction’s share of top-4 finishers divided by its share of the field, over events of 16+ players. Above 1.0 means it punches into the sharp end more than its numbers alone would predict. A faction with no standings coverage this window sits at a neutral 1.0.
Blowout differential — mean WTC battle points (0–20) per game. Winning big and losing small. Team games report battle points directly; singles map the victory-point margin through the standard band table.
Consistency — the inverse of per-game battle-point variance, so a steady army scores higher than a boom-bust one. The floor complement to blowout’s ceiling.
The default weights — win rate 40, top-cut 30, blowout 20, consistency 10 — are an editorial position, not a measurement; the sliders are there so you can disagree. A faction under 30 games in the active format is shown with its score but held in the AVERAGE band (a “THIN” chip marks it) — a lucky short sample shouldn’t front the tier list. This is the answer to the single-number tier lists the methodology library takes apart (Big Soup, Six Bins): the same results, kept as separate measures with visible uncertainty and adjustable weights, rather than collapsed into a confidence they haven’t earned.
Top players is the comparison layer, never a composite input: the curated top-player tier-list rating, shown beside each faction so you can see where results and reputation diverge. The Δ WK column tracks week-over-week movement of the composite at the default weights — it does not follow your slider changes.
Scope: singles by default, because team pairings are captain-assigned and distort matchup-dependent results; the team toggle swaps in the team-format facets, which are a captain-influenced read. The data is the edition-lifetime window of the Archive’s first-party BCP pull. The forest plot keeps an alphabetical row order so a faction’s range bar can be tracked at a fixed position across weight changes; the breakdown table picks up the composite-sorted ranking and reshuffles live as the sliders move.
RECENT DATA REFRESHES
2026-07-26performance
BCP first-party scrape · weekly snapshot — W30: 3198 games / 32 events. First fresh week with the parsePairings pagination fix in place from scrape onward — no amendment needed. Two events excluded (Fp9eW3zahGrU Buenos Aires GT + CHE3eeFeas9N Torneo Tabletop Julio) after roster render_timeout on both scrape attempts caused 'No List' UI badge misparse across every pairing row (91 corrupt rows dropped). Tactical Chilli Championship (clSAOk8xWykX, 46p) contributed 108 games despite roster/placings pagination undercount at 32/46 — pairings tab pagination worked correctly; task 44 tracks the roster/placings-tab pagination bug for follow-up. 40k All Ireland Team Championship (2mOdHtSqO7w7, 120p, team-other) contributed cleanly.
2026-07-24digest
Tournament Coverage Watch · digest sync — Added 1 new Art of War 40k video (Space Wolves dataslate reaction); no new Goonhammer/Tabletop Battles articles this run.
2026-07-19performance
BCP first-party scrape · manual revision — Revises perf-bcp-first-party-scrape-2026-07-19. RETRO AMENDMENT (W27+W28): re-scraped Birmingham Super Major (W27 nxoGV1PSrzij, 163 to 455 games, +292) and PATH TO GLORY (W28 LzoihQT4WmZc, 160 to 453 games, +293) with the parsePairings pagi
2026-07-19performance
BCP first-party scrape · manual revision — RETRO AMENDMENT (W27+W28): re-scraped Birmingham Super Major (W27 nxoGV1PSrzij, 163 to 455 games, +292) and PATH TO GLORY (W28 LzoihQT4WmZc, 160 to 453 games, +293) with the parsePairings pagination fix. +585 games recovered on the two remaining historically-affected 60p+ singles events since scraper inception. ELO chronological forward replay W27 to W28 to W29 restored ratings to reflect the corrected game data. Retro scope was originally 3 events but Edmonton Open (W25) was auto-excluded as 10e (edition_mismatch), so full historical singles retro is now complete.
2026-07-19performance
BCP first-party scrape · manual revision — Revises perf-bcp-first-party-scrape-2026-07-19. W29 AMENDMENT: 2816 -> 5080 games. Fixed parsePairings pagination (bcp-pairings-pagination + bcp-pairings-pagination-selectors) that had been silently capturing only the first ~30 pairings per
2026-07-19performance
BCP first-party scrape · manual revision — W29 AMENDMENT: 2816 -> 5080 games. Fixed parsePairings pagination (bcp-pairings-pagination + bcp-pairings-pagination-selectors) that had been silently capturing only the first ~30 pairings per round on every event with >60 players since scraper inception. Re-scraped 5 W29 events (Tacoma 241->1898, Edinburgh 164->544, Palm Springs 190->339, Bordeaux 159->198, Hee Yaw 162->202). +2264 games recovered (24pct -> 83pct coverage on the affected events). Historical events with >60p (W20-W28) similarly affected - retroactive re-scrape pending.
2026-07-19performance
BCP first-party scrape · manual revision — W29 AMENDMENT: 2816 -> ~5700 games. Fixed parsePairings pagination (bcp-pairings-pagination + bcp-pairings-pagination-selectors) that had been silently capturing only the first ~30 pairings per round on every event with >60 players since scraper inception. Re-scraped 5 W29 events (Tacoma 241->1898, Edinburgh 164->544, Palm Springs 190->339, Bordeaux 159->198, Hee Yaw 162->202). Composite +2265 games recovered (24pct -> 83pct coverage on the affected events). All other W29 events under 60p were already at full coverage. Historical events with >60p (W20-W28) similarly affected - retroactive re-scrape pending.
The Infinite Archive uses your browser’s local storage to remember your preferences — your theme, composite-weight sliders, organize mode, selected faction. We don’t track you, run analytics, sell data, or share with third parties. The Cloudflare CDN sets one functional cookie for bot protection.
Privacy details ›