Virtua Fighter 5 REVO · v1.10 · Every character

Kage

Mid and high combined47,950 matchesMid ranks39,533 matchesHigh ranks8,417 matches

The same figures as a 1080×1080 card, for posting. Click it to copy the image. Full size ↗

Win rate51.5%

Standing#5 of 19 · tier A

Pick rate6.5%

Most picked#3 of 19

Matches counted47,950

Unique players392

Period2026/03/26 – 2026/09/23

Against the roster

Win rate in each matchup, over decided non-mirror matches

Opponent	Win %	Wins	Losses	Matches
Dural	55.77%	266	211	477
Wolf	54.41%	1,603	1,343	2,946
Sarah	52.95%	1,337	1,188	2,525
Shun	52.60%	1,325	1,194	2,519
Jacky	52.45%	2,507	2,273	4,780
Vanessa	52.31%	1,018	928	1,946
Brad	51.97%	1,502	1,388	2,890
Goh	51.91%	1,316	1,219	2,535
Lei Fei	51.81%	1,031	959	1,990
El Blaze	51.58%	1,026	963	1,989
Akira	51.33%	2,618	2,482	5,100
Jeffry	51.33%	1,466	1,390	2,856
Jean	50.78%	1,403	1,360	2,763
Aoi	50.68%	971	945	1,916
Lau	50.35%	1,730	1,706	3,436
Eileen	49.94%	420	421	841
Pai	49.90%	1,190	1,195	2,385
Lion	49.78%	1,497	1,510	3,007
Taka	44.61%	468	581	1,049

By shapeWalls or no wallsBy shape · roundsWalls or no walls · rounds

Stage Impact To Win Probability

Change from this character's own win rate, in points

The same figures as a 1080×1080 card, for posting. Click it to copy the image. Full size ↗

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Each matchup, stage by stage

Win rate in that matchup on that stage

Opponent	Full	Break. full	Half	Break. half	Low	Wall	Mixed	Oct	Rect	Open
Dural55.8%	46	54	63	60	57	71	50	53	43	63
Wolf54.4%	56	53	55	56	59	58	55	53	51	50
Sarah53.0%	49	52	50	50	56	54	56	52	55	56
Shun52.6%	55	55	51	52	56	54	59	46	52	53
Jacky52.5%	53	53	50	56	54	53	49	49	54	51
Vanessa52.3%	59	55	57	53	46	49	55	48	52	53
Brad52.0%	52	53	53	55	52	52	53	52	51	50
Goh51.9%	54	48	56	57	54	50	50	55	48	49
Lei Fei51.8%	55	60	55	50	56	47	53	51	46	49
El Blaze51.6%	55	49	55	51	54	48	54	53	49	47
Akira51.3%	47	51	49	51	54	50	52	55	50	54
Jeffry51.3%	54	54	50	53	45	55	52	50	51	52
Jean50.8%	51	42	53	49	55	54	53	53	50	46
Aoi50.7%	52	53	53	52	47	52	50	47	53	52
Lau50.4%	51	46	48	49	50	52	56	49	51	50
Eileen49.9%	44	43	56	49	45	51	64	58	48	52
Pai49.9%	45	47	51	49	51	55	58	48	50	48
Lion49.8%	45	55	51	51	48	45	51	52	52	48
Taka44.6%	48	65	50	40	43	33	50	45	39	44

Worse on this stageNo changeBetter on this stage

Walls Impact To Win Probability

Against this character's own rate, in points

Walls or no walls+0.0 pts behind walls (95% interval -1.1 to +1.1); within chance across the 20 characters tested, q=0.997

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Walls Impact To Win Probability

Win rate in that matchup with a wall, and without

Opponent	Walls	No walls
Dural55.8%	53	64
Wolf54.4%	54	56
Sarah53.0%	51	56
Shun52.6%	52	54
Jacky52.5%	52	53
Vanessa52.3%	54	49
Brad52.0%	53	51
Goh51.9%	53	51
Lei Fei51.8%	53	50
El Blaze51.6%	52	50
Akira51.3%	51	53
Jeffry51.3%	52	51
Jean50.8%	50	52
Aoi50.7%	52	50
Lau50.4%	49	51
Eileen49.9%	50	50
Pai49.9%	48	51
Lion49.8%	51	47
Taka44.6%	47	40

Worse on this stageNo changeBetter on this stage

Stage Impact To Round Win Probability

Change from this character's own round win rate, in points

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Each matchup, stage by stage, in rounds

Round win rate in that matchup on that stage

Opponent	Full	Break. full	Half	Break. half	Low	Wall	Mixed	Oct	Rect	Open
Wolf52.6%	53	52	53	52	55	54	54	52	53	50
Shun52.5%	54	53	52	53	54	53	56	49	53	52
Dural52.3%	46	46	57	56	52	61	50	51	47	56
Vanessa51.4%	54	53	54	51	48	51	53	49	51	52
Sarah51.4%	50	51	51	49	53	51	53	52	53	52
Brad51.2%	51	52	51	52	52	51	53	52	51	50
Jacky51.2%	51	51	50	53	51	51	49	49	52	51
Akira51.0%	49	51	51	51	52	50	52	53	50	52
Goh50.9%	53	48	54	53	52	50	51	51	49	49
Lei Fei50.6%	52	55	52	49	52	49	49	52	47	49
El Blaze50.5%	52	48	51	51	52	50	50	52	50	48
Jeffry50.5%	52	51	50	51	47	53	50	50	50	51
Jean50.5%	50	45	52	49	52	52	52	52	50	49
Lau50.3%	51	49	49	51	50	51	53	49	51	49
Aoi50.1%	51	52	52	50	48	51	50	48	51	51
Pai50.1%	49	48	50	50	50	53	52	51	49	49
Lion49.7%	47	52	50	51	48	47	52	49	50	50
Eileen49.6%	47	46	52	49	50	50	56	54	49	51
Taka46.7%	49	58	51	44	45	39	50	48	45	45

Worse on this stageNo changeBetter on this stage

Walls Impact To Round Win Probability

Against this character's own round rate, in points

Walls or no walls+0.1 pts behind walls (95% interval -0.4 to +0.7); within chance across the 20 characters tested, q=0.651

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Walls Impact To Round Win Probability

Round win rate in that matchup with a wall, and without

Opponent	Walls	No walls
Wolf52.6%	52	53
Shun52.5%	52	53
Dural52.3%	51	57
Vanessa51.4%	52	51
Sarah51.4%	51	52
Brad51.2%	51	51
Jacky51.2%	51	51
Akira51.0%	51	51
Goh50.9%	51	50
Lei Fei50.6%	51	50
El Blaze50.5%	51	50
Jeffry50.5%	51	50
Jean50.5%	50	51
Lau50.3%	50	50
Aoi50.1%	51	50
Pai50.1%	49	51
Lion49.7%	50	48
Eileen49.6%	50	50
Taka46.7%	49	43

Worse on this stageNo changeBetter on this stage

Where they are played

Share of that region's own picks, and where that ranks

Region	Share	Standing	Appearances
Asia	6.79%	4	36,660
South America	6.66%	5	7,475
North America	4.94%	9	3,744
Europe	2.63%	15	458
Middle East	0.95%	5	7
Oceania	0.00%	15	0
Not recorded	7.60%	3	2,766

The same figures as a 1080×1080 card, for posting. Click it to copy the image. Full size ↗

Win rate51.8%

Standing#3 of 19 · tier A

Pick rate6.3%

Most picked#3 of 19

Matches counted39,533

Unique players356

Period2026/03/26 – 2026/09/23

Against the roster

Win rate in each matchup, over decided non-mirror matches

Opponent	Win %	Wins	Losses	Matches
Dural	56.08%	249	195	444
Wolf	54.07%	1,381	1,173	2,554
El Blaze	53.15%	828	730	1,558
Sarah	53.11%	1,145	1,011	2,156
Lei Fei	53.07%	831	735	1,566
Jacky	52.79%	2,025	1,811	3,836
Akira	52.25%	2,193	2,004	4,197
Brad	52.22%	1,225	1,121	2,346
Shun	52.16%	1,063	975	2,038
Eileen	51.77%	293	273	566
Goh	51.71%	1,150	1,074	2,224
Jeffry	51.23%	1,316	1,253	2,569
Jean	51.09%	1,311	1,255	2,566
Aoi	51.02%	874	839	1,713
Vanessa	51.00%	740	711	1,451
Lau	50.97%	1,337	1,286	2,623
Lion	50.61%	1,209	1,180	2,389
Pai	49.97%	984	985	1,969
Taka	44.14%	339	429	768

By shapeWalls or no wallsBy shape · roundsWalls or no walls · rounds

Stage Impact To Win Probability

Change from this character's own win rate, in points

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Each matchup, stage by stage

Win rate in that matchup on that stage

Opponent	Full	Break. full	Half	Break. half	Low	Wall	Mixed	Oct	Rect	Open
Dural56.1%	46	52	64	62	59	71	50	53	41	67
Wolf54.1%	55	54	55	57	58	59	52	52	50	50
El Blaze53.1%	57	51	56	53	56	50	51	55	50	51
Sarah53.1%	49	49	50	50	55	54	57	53	56	59
Lei Fei53.1%	57	61	55	49	57	52	52	54	48	49
Jacky52.8%	52	53	50	56	53	54	50	50	54	52
Akira52.2%	49	52	49	52	55	51	54	56	50	55
Brad52.2%	53	54	53	57	55	51	51	51	50	50
Shun52.2%	57	54	48	54	54	53	58	46	53	53
Eileen51.8%	47	43	57	55	43	54	69	57	50	49
Goh51.7%	53	48	57	58	54	51	47	52	48	49
Jeffry51.2%	53	52	50	54	46	54	54	50	50	54
Jean51.1%	53	42	53	51	56	55	54	53	50	46
Aoi51.0%	54	54	53	53	47	53	51	46	53	53
Vanessa51.0%	56	55	59	52	42	47	53	45	50	54
Lau51.0%	54	47	49	51	51	53	55	50	49	49
Lion50.6%	46	56	52	52	49	44	53	52	52	48
Pai50.0%	44	46	52	48	50	60	58	48	51	47
Taka44.1%	48	58	50	41	45	27	54	49	36	44

Worse on this stageNo changeBetter on this stage

Walls Impact To Win Probability

Against this character's own rate, in points

Walls or no walls-0.3 pts behind walls (95% interval -1.5 to +0.9); within chance across the 20 characters tested, q=0.799

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Walls Impact To Win Probability

Win rate in that matchup with a wall, and without

Opponent	Walls	No walls
Dural56.1%	53	66
Wolf54.1%	54	56
El Blaze53.1%	53	52
Sarah53.1%	51	56
Lei Fei53.1%	54	53
Jacky52.8%	52	53
Akira52.2%	51	54
Brad52.2%	53	52
Shun52.2%	52	53
Eileen51.8%	52	49
Goh51.7%	53	51
Jeffry51.2%	51	51
Jean51.1%	50	52
Aoi51.0%	52	50
Vanessa51.0%	53	48
Lau51.0%	50	51
Lion50.6%	52	47
Pai50.0%	48	52
Taka44.1%	47	38

Worse on this stageNo changeBetter on this stage

Stage Impact To Round Win Probability

Change from this character's own round win rate, in points

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Each matchup, stage by stage, in rounds

Round win rate in that matchup on that stage

Opponent	Full	Break. full	Half	Break. half	Low	Wall	Mixed	Oct	Rect	Open
Dural52.5%	47	46	57	57	52	61	50	51	46	58
Wolf52.3%	53	52	52	53	54	55	53	51	52	50
Shun52.1%	55	52	50	53	53	53	56	49	53	52
Akira51.7%	50	52	51	52	52	51	53	54	50	53
Jacky51.6%	51	52	51	53	51	52	51	50	52	52
Brad51.6%	52	52	51	54	54	51	52	52	50	50
Sarah51.5%	49	50	51	49	52	51	54	52	53	53
El Blaze51.5%	53	49	52	52	53	51	49	53	51	49
Lei Fei51.2%	54	56	52	48	52	51	49	53	48	49
Goh50.9%	53	48	55	54	51	51	49	50	48	50
Vanessa50.8%	53	53	54	51	47	50	52	48	50	52
Eileen50.8%	48	46	52	52	51	51	58	53	51	49
Lau50.7%	52	49	50	52	50	51	53	49	50	49
Jean50.7%	51	45	52	50	52	53	53	52	50	49
Jeffry50.5%	51	50	50	52	47	53	51	50	50	52
Aoi50.2%	52	52	52	51	48	50	50	48	51	52
Lion50.1%	48	52	51	52	49	47	52	49	50	51
Pai50.1%	49	48	50	49	50	55	52	51	49	49
Taka46.6%	48	54	51	45	48	36	51	50	44	45

Worse on this stageNo changeBetter on this stage

Walls Impact To Round Win Probability

Against this character's own round rate, in points

Walls or no walls+0.0 pts behind walls (95% interval -0.6 to +0.6); within chance across the 20 characters tested, q=0.986

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Walls Impact To Round Win Probability

Round win rate in that matchup with a wall, and without

Opponent	Walls	No walls
Dural52.5%	51	57
Wolf52.3%	52	53
Shun52.1%	52	53
Akira51.7%	51	52
Jacky51.6%	52	52
Brad51.6%	52	52
Sarah51.5%	51	52
El Blaze51.5%	52	51
Lei Fei51.2%	52	51
Goh50.9%	51	51
Vanessa50.8%	51	50
Eileen50.8%	51	51
Lau50.7%	50	50
Jean50.7%	50	51
Jeffry50.5%	51	51
Aoi50.2%	51	50
Lion50.1%	50	49
Pai50.1%	49	51
Taka46.6%	49	43

Worse on this stageNo changeBetter on this stage

Where they are played

Share of that region's own picks, and where that ranks

Region	Share	Standing	Appearances
Asia	6.51%	5	28,463
South America	6.66%	5	7,433
North America	5.00%	8	3,574
Europe	2.32%	14	341
Middle East	0.95%	5	7
Oceania	0.00%	15	0
Not recorded	6.82%	4	2,259

The same figures as a 1080×1080 card, for posting. Click it to copy the image. Full size ↗

Win rate49.9%

Standing#9 of 19 · tier B

Pick rate8.0%

Most picked#5 of 19

Matches counted8,417

Unique players67

Period2026/03/26 – 2026/09/23

Against the roster

Win rate in each matchup, over decided non-mirror matches

Opponent	Win %	Wins	Losses	Matches
Wolf	56.63%	222	170	392
Vanessa	56.16%	278	217	495
Shun	54.47%	262	219	481
Goh	53.38%	166	145	311
Jeffry	52.26%	150	137	287
Sarah	52.03%	192	177	369
Dural	51.52%	17	16	33
Jacky	51.06%	482	462	944
Brad	50.92%	277	267	544
Pai	49.52%	206	210	416
Lau	48.34%	393	420	813
Aoi	47.78%	97	106	203
Lei Fei	47.17%	200	224	424
Akira	47.07%	425	478	903
Jean	46.70%	92	105	197
Lion	46.60%	288	330	618
Eileen	46.18%	127	148	275
El Blaze	45.94%	198	233	431
Taka	45.91%	129	152	281

By shapeWalls or no wallsBy shape · roundsWalls or no walls · rounds

Stage Impact To Win Probability

Change from this character's own win rate, in points

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Each matchup, stage by stage

Win rate in that matchup on that stage

Opponent	Full	Break. full	Half	Break. half	Low	Wall	Mixed	Oct	Rect	Open
Wolf56.6%	65	46	57	48	65	53	73	60	63	51
Vanessa56.2%	66	57	51	54	58	53	61	56	61	50
Shun54.5%	48	59	70	44	63	57	62	45	51	52
Goh53.4%	59	46	48	50	61	43	75	77	52	50
Jeffry52.3%	62	73	50	50	40	57	33	50	59	35
Sarah52.0%	51	67	47	52	62	53	52	46	47	45
Dural51.5%	50	60	60	0	0	75	·	50	60	25
Jacky51.1%	58	55	49	55	55	47	40	46	54	50
Brad50.9%	46	51	51	46	37	52	65	60	55	48
Pai49.5%	50	50	48	56	55	36	59	50	48	52
Lau48.3%	44	42	47	40	50	49	57	48	56	50
Aoi47.8%	38	36	52	41	53	52	47	54	52	46
Lei Fei47.2%	42	56	56	51	51	33	54	42	39	48
Akira47.1%	35	44	50	46	52	45	41	52	52	46
Jean46.7%	26	38	52	37	48	45	38	67	59	50
Lion46.6%	43	49	46	46	43	48	44	54	54	44
Eileen46.2%	35	43	53	38	50	48	54	59	43	56
El Blaze45.9%	50	43	48	48	49	42	61	48	42	31
Taka45.9%	50	85	48	38	40	47	33	32	48	43

Worse on this stageNo changeBetter on this stage

Walls Impact To Win Probability

Against this character's own rate, in points

Walls or no walls+1.3 pts behind walls (95% interval -1.3 to +3.9); within chance across the 20 characters tested, q=0.711

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Walls Impact To Win Probability

Win rate in that matchup with a wall, and without

Opponent	Walls	No walls
Wolf56.6%	57	56
Vanessa56.2%	57	54
Shun54.5%	52	57
Goh53.4%	56	50
Jeffry52.3%	57	45
Sarah52.0%	52	53
Dural51.5%	54	44
Jacky51.1%	53	51
Brad50.9%	51	46
Pai49.5%	50	49
Lau48.3%	46	50
Aoi47.8%	47	51
Lei Fei47.2%	48	44
Akira47.1%	47	48
Jean46.7%	46	47
Lion46.6%	48	45
Eileen46.2%	45	51
El Blaze45.9%	47	42
Taka45.9%	48	43

Worse on this stageNo changeBetter on this stage

Stage Impact To Round Win Probability

Change from this character's own round win rate, in points

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Each matchup, stage by stage, in rounds

Round win rate in that matchup on that stage

Opponent	Full	Break. full	Half	Break. half	Low	Wall	Mixed	Oct	Rect	Open
Wolf54.2%	54	49	54	51	56	53	64	58	60	50
Shun54.2%	51	57	63	51	58	55	59	49	52	53
Vanessa53.2%	58	54	52	51	54	53	57	51	55	51
Goh50.9%	56	45	48	51	54	46	62	62	51	47
Sarah50.7%	51	56	49	50	55	52	49	50	49	46
Jeffry50.4%	57	60	53	46	43	54	42	49	52	40
Dural50.4%	38	47	52	40	40	62	·	54	55	42
Pai49.9%	49	50	50	55	52	44	55	53	49	51
Brad49.7%	48	49	50	46	42	51	59	54	53	48
Jacky49.5%	52	50	49	52	51	48	43	47	52	51
Aoi49.4%	45	47	53	43	52	52	50	50	51	47
Lau49.0%	47	47	48	47	49	50	55	47	52	49
Lei Fei48.4%	44	51	53	51	52	43	49	49	44	48
Akira47.9%	43	45	51	47	52	47	44	49	49	48
Lion47.8%	43	49	48	49	45	48	49	51	50	46
Jean47.7%	37	42	52	44	50	45	44	56	53	50
Eileen47.3%	43	46	52	43	47	48	49	55	44	54
Taka47.2%	51	73	50	42	41	46	47	40	48	47
El Blaze47.1%	48	48	47	48	47	45	52	49	45	40

Worse on this stageNo changeBetter on this stage

Walls Impact To Round Win Probability

Against this character's own round rate, in points

Walls or no walls+0.7 pts behind walls (95% interval -0.6 to +2.1); within chance across the 20 characters tested, q=0.546

How significance is decided

Each gap is two win rates compared with a pooled two-proportion z test, over decided matches only.
The 95% interval is Newcombe's hybrid score method, which keeps its shape near 0% and 100% where a symmetric interval does not.
q is p corrected for how many tests were asked at once, by Benjamini–Hochberg at a 5% false discovery rate. A figure is called real on q, never on p alone.
Standard errors are widened by a measured design effect, because one player appears in many matches and those matches are not independent of each other.
Stages are assigned at random, so a gap between them is the stage's own effect and not a difference in who chose to play where.

Walls Impact To Round Win Probability

Round win rate in that matchup with a wall, and without

Opponent	Walls	No walls
Wolf54.2%	55	53
Shun54.2%	53	55
Vanessa53.2%	53	53
Goh50.9%	52	48
Sarah50.7%	51	51
Jeffry50.4%	53	47
Dural50.4%	50	50
Pai49.9%	51	49
Brad49.7%	50	47
Jacky49.5%	50	50
Aoi49.4%	49	51
Lau49.0%	48	49
Lei Fei48.4%	49	47
Akira47.9%	48	49
Lion47.8%	48	46
Jean47.7%	47	48
Eileen47.3%	47	50
Taka47.2%	49	44
El Blaze47.1%	48	45

Worse on this stageNo changeBetter on this stage

Where they are played

Share of that region's own picks, and where that ranks

Region	Share	Standing	Appearances
Asia	8.04%	4	8,197
North America	4.05%	9	170
Europe	4.25%	9	117
South America	7.42%	4	42
Not recorded	15.54%	2	507