Who put people in the seats, measured inside their own promotion and season.
Raw attendance ranks the promotions that hired a wrestler, not the wrestler. This holds the promotion and the year fixed and asks whether the cards someone main-events out-draw the other cards that company ran that season.
Attendance has been sitting in the corpus the whole time and nothing has ever read it. 28,063 of the 59,700 documented cards carry a house count, going back to a 1948 show and forward to this August, and no published artifact on this site has ever carried the figure forward. This board is the first thing built on it.
Raw attendance cannot answer the question, because raw attendance is a fact about the promotion. New Japan's median documented house is 1,850, All Japan's is 1,200, NOAH's is 1,103. Ranking wrestlers by the size of the rooms they worked in ranks the companies that hired them. So the promotion and the year are held fixed and the question becomes narrower: do the cards this wrestler main-events out-draw the other cards that company ran that season? A multiplier of 2.32 means the houses under that name ran about 130 per cent above what the promotion was drawing around them.
The order that produces is not the K-WAR order and the disagreements are the interesting part. Shinsuke Nakamura is second here and 27th in the model, which is close to agreement. Antonio Inoki is fourth here and 643rd in the model. Naoya Ogawa is eighteenth here and 2,103rd. Nothing about drawing power is an input to K-WAR, and the model says so in its own limitations: it rates documented results and nothing else. Two measures of the same career, built from different evidence, are allowed to disagree, and where they do it is usually because one of them is looking at a wrestler the corpus barely documents.
One absence is worth naming before anybody looks for it. The highest score the model produced belongs to the Great Muta, at 3.02, and he is not on this board. The corpus splits him across three rows that were never linked, one of which carries ten matches, so there is no single career here to attach the figure to. A drawing figure on the wrong row is the same failure as a photograph of the wrong face, and the rule is the same: refuse rather than guess.
590 main events with a documented house, which is the second-largest sample on this board and the reason 2.98 barely moved under shrinkage. A thin record gets pulled towards no effect here; this one had nothing to pull it with. The houses themselves are ordinary New Japan houses, median 1,478, against a promotion median of 1,850, and the figure is not about their size. It is about how reliably the number went up when his name was on top of the card, across seventeen years and one 46,913-seat night.
The closest thing to agreement between the two measures on this page: 27th in K-WAR, second in drawing power. That is rare enough to be worth stating, because the rest of this board is mostly the two measures pointing in different directions. 473 main events at a median house of 1,400, with a best of 53,000. His All-Time score of 99.03 is the highest here, so this is one career where documented results and documented attendance are telling the same story.
481 main events split across two companies that are the same story told twice. In All Japan the median house under him is 3,100; in NOAH, the promotion he left to form, it is 1,800 with a best of 62,000. The model measures him inside each of those separately, which is the only way the comparison means anything, and it returns the same figure on both sides of the split. A 2,100 median across the whole record is the third-highest on this board.
The largest disagreement between the two measures that involves a famous name. K-WAR has him 643rd on a score of 71.66, because the corpus documents 498 of his matches and a model of documented results cannot rate what was never written down. The attendance record is far denser: 273 main events at a median house of 3,260, the highest median on this board, topping out at 63,900. Two measures of one career, and the thin one is the ranking.
919 main events with a documented house, the largest sample anywhere in this projection, which makes 2.32 the least uncertain figure on the board. The span it covers is the useful part: his New Japan years run through the stretch in which the promotion's median house fell from 5,308 to 802 and climbed back to 1,494. A multiplier is measured against a moving floor, and his stayed above it the whole way down and the whole way back.
242nd in the model and sixth here, on 308 main events across all four of the promotions this board can see. The median house under him is 1,700 and the best is 55,000. What separates a figure like this from a thin one is that it was earned in four different companies with four different baselines, so it is not a single favourable booking pattern in a single place; the same lift shows up wherever the record follows him.
The highest All Japan figure belonging to somebody who did not leave in 2000. 303 main events, median house 2,100, best 64,000. All Japan's geometric mean reads 3,226 in 1999 and 2,355 in 2000, and NOAH appears that same year at 2,316, which is most of an audience walking out of one company and into another. He is the man the model measures against what was left, and 2.09 is what he did with it.
The only figure in the top ten measured against a small building. His median house is 900 and his best is 9,600, where every other name above him has been in front of 40,000 people at least once. That is the point of holding the promotion fixed: 2.04 in Dragon Gate is the same claim as 2.04 in New Japan, made about a room a twentieth of the size. 529 main events is a large enough record to make the claim stick.
485 main events across four promotions, median house 2,000, best 64,000. The model ranks him 137th on documented results and ninth on documented attendance, which is a smaller gap than most of this board and comes from the same place his K-WAR score does: a very long career spent near the top of several cards rather than one.
The thinnest record on the board and the most shrunk figure on it. 50 main events is barely above the floor of 15, so the ridge pulled hard, and 1.92 is what survived the pulling. What it is pulling on is unusual: a median house of 1,295 and a best of 50,000, which is a career of ordinary cards punctuated by very large ones. K-WAR has him 1,234th on 546 documented matches.
716 main events spanning 1979 to 2024, which is the longest documented drawing record here by two decades. Median house 1,500, best 61,500. A figure earned across forty-five years is a different kind of evidence from one earned across five: it survives four promotions, several booking regimes, and the entire collapse of the audience it was measured against.
152 main events, median house 1,356, best 40,008. The interesting number beside it is 116th in K-WAR on 769 documented matches, which is a high placement built from a comparatively short record. Both measures are reading the same thing here, which is a career with a low match count and a high concentration of the cards that mattered.
64 main events inside eligible cells, one of the smallest samples on the board, against a median house of 1,706 and a best of 38,162. The eligibility floor is doing visible work in a case like this: most of his AEW career sits in promotion-years that report a house on 29 per cent of their cards and are therefore not used at all. What 1.84 measures is the Japanese part of the record, because that is the part the record kept.
243 main events at a median house of 2,300, the fourth-highest median here, with a best of 63,500. The span runs 1981 to 2015 across four promotions including one he founded. Tied to two decimal places with Kenta Kobashi at 1.82, which the model treats as a tie and so should anybody reading it.
374 main events, median house 2,050, best 58,000, and the same 1.82 as Tenryu. His record follows the 2000 split the same way Misawa's and Akiyama's do, out of All Japan and into NOAH, and the figure holds across the move. 233rd in K-WAR on 939 documented matches, which is one of the larger gaps on this board between what the results say and what the houses do.
91st in the model and sixteenth here, on 375 main events with a median house of 1,100. That median is the second-lowest in the top sixteen, which makes 1.79 a claim about consistency rather than scale: the cards he tops are not the biggest cards, they are reliably bigger than the ones around them. 2,217 documented matches, running 1988 to 2026.
2,999 documented matches, more than anybody else on this board, and a K-WAR rank of 740. The attendance record is similarly dense and similarly small-scale: 480 main events at a median house of 900, best 9,400. Dragon Gate is the promotion this measure is most confident about, reporting a house on 91 per cent of its cards, so 1.77 here rests on close to a census rather than a sample.
The widest gap on the board. K-WAR ranks him 2,103rd on a score of 52.03, built from 141 documented matches, which is a career the model can barely see. The attendance record sees something else: 46 main events at a median house of 2,950, second-highest here, with a best of 60,000, all of it inside 1999 to 2003. Four years of very large rooms against a documented record too thin to rank.
551 main events at a median house of 2,500, which is the second-highest median paired with a large sample rather than a small one. Most of the high medians on this board come with a few dozen main events; this one comes with five hundred. 450th in K-WAR, and the distance between that and nineteenth here is the usual shape of this page.
516 main events, median house 1,244, best 34,995, and 55th in K-WAR on a score of 97.45. He closes the board at 1.65, which against a New Japan median of 1,850 is still a substantial figure; the top of this list is unusually crowded. His record runs 2006 to 2026 and takes in the 2021 trough, when the promotion's typical house fell to 802, its lowest documented year.