Analysis · Oct 19, 2026

Players press faster against Wolf

Players work the controller at 6.8 inputs a second in ranks 21–29 and 8.1 at 40–46. But the pace is not only theirs: the same people press +2.0% against Wolf and -2.0% against Eileen.

Inputs a second, ranks 21–296.8

Inputs a second, ranks 40–468.1

Same players, against Wolf+2.0%

A player's own pace, per rank climbed+0.3%

Faster up the ladder, mostly by who gets there

inputs a second and a minute, ranks 21–46

Across the ladder, pace rises +1.0% for every rank. Follow the same player as they climb and it rises +0.3%: about 34% of the ladder's slope. The rest is selection, faster players rising further, rather than anyone speeding up because they ranked up. The rank-by-rank curve is on the input rate report (/reports/input-rate/).

Players	Inputs/sec	Inputs/min	95% interval	Players
Ranks 21–29	6.8	409	404–414	3,977
Ranks 30–39	7.3	437	431–444	2,096
Ranks 40–46	8.1	488	475–502	439

Players	Per rank	95% interval	p
Across the ladder	+1.0%	+0.8 – +1.1	< 0.0001
The same player, climbing	+0.3%	+0.2 – +0.4	< 0.0001

From the 1-in-10 input sample. A cell is one player at one rank with 3 minutes of live rounds; 3,871 players have one, and 2,929 have cells at two or more ranks, which is all the within-player slope is built from. Both slopes are fitted to the log of the rate on the same cells, weighted by live time, with errors clustered by player.

The opponent sets the pace

each player against one character vs their others · 5,104 players

Hold the player fixed and change only who they are facing. Against Wolf the same people's pace runs +2.0% on what it is against everyone else; against Eileen, -2.0%. A gap of a couple of percent is under ten inputs a minute. It is small for any one round and consistent across thousands of players.


Wolf	+2.0%	+1.9 – +2.2	< 0.0001	3,406
Akira	+1.1%	+1.0 – +1.3	< 0.0001	3,947
Lau	+1.0%	+0.8 – +1.2	< 0.0001	3,080
Jacky	+0.8%	+0.7 – +1.0	< 0.0001	3,675
Pai	+0.7%	+0.5 – +0.8	< 0.0001	3,484
Goh	+0.6%	+0.4 – +0.8	< 0.0001	3,305
Lion	+0.3%	+0.1 – +0.4	0.00083	3,200
Jeffry	-0.0%	-0.2 – +0.2	0.91	3,237
Dural	-0.1%	-0.4 – +0.2	0.44	1,478
Aoi	-0.3%	-0.5 – -0.2	< 0.0001	3,390
Brad	-0.4%	-0.5 – -0.2	< 0.0001	2,987
Shun	-0.4%	-0.5 – -0.2	< 0.0001	3,420
Jean	-0.6%	-0.8 – -0.4	< 0.0001	3,147
El Blaze	-0.9%	-1.1 – -0.7	< 0.0001	2,974
Vanessa	-1.0%	-1.1 – -0.8	< 0.0001	2,858
Sarah	-1.4%	-1.5 – -1.2	< 0.0001	3,502
Lei Fei	-1.4%	-1.6 – -1.3	< 0.0001	2,834
Kage	-1.6%	-1.8 – -1.5	< 0.0001	3,160
Taka	-1.9%	-2.2 – -1.7	< 0.0001	2,418
Eileen	-2.0%	-2.2 – -1.7	< 0.0001	2,365

From every regular's stats pull, not a sample. Each player with 3 minutes against a character and 3 against the rest is compared only with themselves; a character needs 30 such players. 18 of 20 survive the Benjamini-Hochberg correction. The character the player picked is not held fixed here, so a switch made after seeing the opponent is in it; the matchup table below holds it fixed. This says the pace changes, not why: an opponent who throws more asks for more breaks, one who keeps you blocking asks for fewer, and a round's inputs count both.

Matchups that change the pace most

98 of 310 matchups tested survive correction

Now hold the character a player picked fixed as well. Each row compares the same person on the same character against one opponent with that character against everyone else, so neither the player nor their pick is in the gap.

Pressing more

Playing → against	Vs their others	95% interval	q	Players compared
Vanessa → Wolf	+4.1%	+2.7 – +5.6	< 0.0001	70
Lei Fei → Wolf	+4.1%	+2.5 – +5.7	< 0.0001	60
Lau → Wolf	+4.0%	+2.2 – +5.7	0.00029	87
El Blaze → Wolf	+3.5%	+2.2 – +4.8	< 0.0001	74
Pai → Wolf	+3.3%	+2.2 – +4.3	< 0.0001	103
Sarah → Wolf	+3.3%	+2.0 – +4.6	< 0.0001	95
Goh → Wolf	+3.2%	+1.8 – +4.6	0.00022	97
Eileen → Wolf	+3.1%	+0.7 – +5.5	0.038	40
Kage → Goh	+3.1%	+1.7 – +4.5	0.00043	88
Kage → Wolf	+3.1%	+1.8 – +4.3	< 0.0001	107

Pressing less

Playing → against	Vs their others	95% interval	q	Players compared
Aoi → Taka	-5.5%	-7.9 – -3.2	0.00044	31
Pai → Eileen	-4.1%	-5.9 – -2.3	0.00073	31
Shun → Taka	-3.5%	-5.7 – -1.3	0.015	30
Vanessa → Sarah	-3.4%	-4.7 – -2.1	< 0.0001	60
Lei Fei → Kage	-3.3%	-4.5 – -2.1	< 0.0001	50
Wolf → Kage	-3.2%	-4.3 – -2.1	< 0.0001	100
Akira → Taka	-3.2%	-4.9 – -1.4	0.0037	69
Eileen → Sarah	-3.2%	-6.1 – -0.2	0.092	32
Lau → Lei Fei	-3.1%	-4.7 – -1.4	0.0036	45
Aoi → Kage	-3.0%	-4.1 – -2.0	< 0.0001	81

From the 1-in-10 input sample. A player needs 3 minutes on a character against the opponent and 3 against the rest, and a matchup needs 30 such players. The correction runs over every matchup tested, not just the rows shown.

Up a rank, down a rank

the same players by the opponent's rank against their own

Does facing someone above you speed you up? Barely. Against an equal the pace is a shade higher than against anyone else, and against someone three or more ranks below it is a shade lower. Neither moves it by even one percent.

From the 1-in-10 input sample, each player against their own other steps, corrected over the five.

Pace and rating

3,533 rated regulars with 30 minutes counted, in five groups

Cut the regulars into five equal groups by Elo and the pace climbs group by group, but loosely: the rank correlation between a player's rating and their pace is 0.22. Fast hands are common at the top and do not put you there.

Elo group	Elo	Inputs/sec	Inputs/min	95% interval	Players
Group 1	794–1343	6.8	411	400–421	706
Group 2	1343–1468	7.1	426	415–437	707
Group 3	1470–1581	7.1	428	418–437	706
Group 4	1582–1714	7.4	442	433–452	707
Group 5	1714–2217	8.0	479	469–488	707

Spearman's rho 0.22 (p < 0.0001). Each group's rate is its inputs over its live minutes, with a player-clustered interval.

The fastest hands

regulars with 30 minutes counted · of 4,069

Rk	Player	Region	Played most	Inputs/sec	Inputs/min	Elo	Rounds counted
1	Paul68kdmd	North America	Akira	14.4	865	—	191
2	LEEGOSEGA	Asia	Akira	14.2	852	1762	4,628
3	もっこり	Asia	Akira	13.0	779	1630	5,405
4	ctraum	Asia	Akira	12.9	775	1707	7,006
5	yumi_itigo	North America	Shun	12.7	760	1320	639
6	zxyzguy	North America	Jacky	12.6	757	1364	8,347
7	Manueljairo	—	Jacky	12.6	754	—	102
8	ｎ	Asia	Akira	12.6	753	—	94
9	kinchan0421	—	Jacky	12.5	753	1518	517
10	Zeke_fund	—	Shun	12.5	752	—	155

Over the whole season, every character they played. A player's own page shows their pace per character and where it sits among everyone else's. 8 profiles averaging over 20 inputs a second are set aside here and from every cut across players: no one keeps that up for a season, and some of their rounds count more inputs than the game has frames, which points to turbo hardware or a counting fault rather than a pace.

An input is any change on the controller (a new stick direction, a button pressed or released), counted from the bell to the moment a round is decided. Rates are inputs over live minutes, never an average of per-round rates. The ladder, matchup and rank-gap cuts are the input rate report's 1-in-10 sample at ranks 21–46 (802,480 player-rounds); the opponent, rating and fastest-player cuts are every regular's stats pull, which counts the rounds whose inputs have been decoded so far. Method and the rank curve: /reports/input-rate/.