Virtua Fighter 5 REVO · v1.10 · Character Matchup Winrates
Pai vs Sarah
Matches counted1,707
Win rate51.4%
Widest swing8.0 pts
Survive correction0
Period2026/03/26 – 2026/09/23
Where it was fought
Pai's win rate on each stage, against this matchup everywhere| Stage | Stage effect | Win % | Change | p | Matches |
|---|---|---|---|---|---|
| Rectangle | 59.4% | +8.0 | 0.014 | 207 | |
| Half fence | 54.7% | +3.3 | 0.376 | 159 | |
| Fence and open | 52.5% | +1.1 | 0.820 | 101 | |
| Full fence | 51.9% | +0.6 | 0.874 | 181 | |
| Octagon | 50.6% | -0.8 | 0.834 | 166 | |
| Low fence | 49.6% | -1.8 | 0.674 | 133 | |
| Breakable half fence | 49.4% | -2.0 | 0.592 | 164 | |
| Breakable full fence | 49.3% | -2.1 | 0.604 | 142 | |
| Open | 47.4% | -4.0 | 0.300 | 152 | |
| Single wall | 43.9% | -7.5 | 0.030 | 189 |
p < 0.05 on its ownStill there after correcting for how many were tested
Stage against stage
Pai's win rate on the row's stage, minus the same on the column'sAll ten shapes at once: chi-square 11.95 on 9 df, p 0.216
| Stage | Full | Break. full | Half | Break. half | Low | Wall | Mixed | Open | Rect | Oct |
|---|---|---|---|---|---|---|---|---|---|---|
| Full fence | 51.9%181 | +2.60.638 | -2.80.608 | +2.50.637 | +2.30.686 | +8.00.123 | -0.60.930 | +4.60.407 | -7.50.138 | +1.30.804 |
| Breakable full fence | -2.60.638 | 49.3%142 | -5.40.347 | -0.10.987 | -0.30.957 | +5.40.331 | -3.20.625 | +1.90.741 | -10.10.062 | -1.30.819 |
| Half fence | +2.80.608 | +5.40.347 | 54.7%159 | +5.30.338 | +5.10.385 | +10.80.045 | +2.20.724 | +7.30.195 | -4.70.367 | +4.10.458 |
| Breakable half fence | -2.50.637 | +0.10.987 | -5.30.338 | 49.4%164 | -0.20.968 | +5.50.304 | -3.10.626 | +2.00.719 | -10.00.054 | -1.20.826 |
| Low fence | -2.30.686 | +0.30.957 | -5.10.385 | +0.20.968 | 49.6%133 | +5.70.312 | -2.90.666 | +2.20.704 | -9.80.076 | -1.00.866 |
| Single wall | -8.00.123 | -5.40.331 | -10.80.045 | -5.50.304 | -5.70.312 | 43.9%189 | -8.60.164 | -3.50.524 | -15.50.002 | -6.70.208 |
| Fence and open | +0.60.930 | +3.20.625 | -2.20.724 | +3.10.626 | +2.90.666 | +8.60.164 | 52.5%101 | +5.10.426 | -6.90.248 | +1.90.767 |
| Open | -4.60.407 | -1.90.741 | -7.30.195 | -2.00.719 | -2.20.704 | +3.50.524 | -5.10.426 | 47.4%152 | -12.10.024 | -3.20.564 |
| Rectangle | +7.50.138 | +10.10.062 | +4.70.367 | +10.00.054 | +9.80.076 | +15.50.002 | +6.90.248 | +12.10.024 | 59.4%207 | +8.80.089 |
| Octagon | -1.30.804 | +1.30.819 | -4.10.458 | +1.20.826 | +1.00.866 | +6.70.208 | -1.90.767 | +3.20.564 | -8.80.089 | 50.6%166 |
Better on the row's stage — deeper at 5 and 10 pointsBetter on the column's stagep < 0.05 on its ownStill there after correcting for how many were tested· fewer than 80 decided matches on one side of the comparison
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.
Matches counted1,579
Win rate51.0%
Widest swing8.0 pts
Survive correction0
Period2026/03/26 – 2026/09/23
Where it was fought
Pai's win rate on each stage, against this matchup everywhere| Stage | Stage effect | Win % | Change | p | Matches |
|---|---|---|---|---|---|
| Rectangle | 59.0% | +8.0 | 0.015 | 200 | |
| Half fence | 55.3% | +4.3 | 0.280 | 141 | |
| Fence and open | 53.5% | +2.6 | 0.600 | 99 | |
| Full fence | 51.5% | +0.6 | 0.882 | 163 | |
| Octagon | 50.3% | -0.7 | 0.859 | 159 | |
| Breakable half fence | 49.0% | -2.0 | 0.609 | 153 | |
| Low fence | 48.5% | -2.5 | 0.548 | 130 | |
| Breakable full fence | 48.0% | -3.0 | 0.486 | 123 | |
| Open | 46.8% | -4.2 | 0.299 | 141 | |
| Single wall | 43.1% | -7.9 | 0.031 | 167 |
p < 0.05 on its ownStill there after correcting for how many were tested
Stage against stage
Pai's win rate on the row's stage, minus the same on the column'sAll ten shapes at once: chi-square 12.59 on 9 df, p 0.182
| Stage | Full | Break. full | Half | Break. half | Low | Wall | Mixed | Open | Rect | Oct |
|---|---|---|---|---|---|---|---|---|---|---|
| Full fence | 51.5%163 | +3.60.550 | -3.80.509 | +2.50.655 | +3.10.601 | +8.40.126 | -2.00.753 | +4.70.411 | -7.50.154 | +1.20.827 |
| Breakable full fence | -3.60.550 | 48.0%123 | -7.30.233 | -1.10.862 | -0.50.937 | +4.90.412 | -5.60.410 | +1.20.851 | -11.00.053 | -2.30.696 |
| Half fence | +3.80.509 | +7.30.233 | 55.3%141 | +6.30.280 | +6.90.259 | +12.20.033 | +1.80.785 | +8.50.153 | -3.70.498 | +5.00.386 |
| Breakable half fence | -2.50.655 | +1.10.862 | -6.30.280 | 49.0%153 | +0.60.925 | +5.90.290 | -4.50.484 | +2.20.705 | -10.00.062 | -1.30.819 |
| Low fence | -3.10.601 | +0.50.937 | -6.90.259 | -0.60.925 | 48.5%130 | +5.30.358 | -5.10.447 | +1.60.785 | -10.50.060 | -1.90.754 |
| Single wall | -8.40.126 | -4.90.412 | -12.20.033 | -5.90.290 | -5.30.358 | 43.1%167 | -10.40.100 | -3.70.516 | -15.90.002 | -7.20.193 |
| Fence and open | +2.00.753 | +5.60.410 | -1.80.785 | +4.50.484 | +5.10.447 | +10.40.100 | 53.5%99 | +6.70.305 | -5.50.369 | +3.20.615 |
| Open | -4.70.411 | -1.20.851 | -8.50.153 | -2.20.705 | -1.60.785 | +3.70.516 | -6.70.305 | 46.8%141 | -12.20.026 | -3.50.544 |
| Rectangle | +7.50.154 | +11.00.053 | +3.70.498 | +10.00.062 | +10.50.060 | +15.90.002 | +5.50.369 | +12.20.026 | 59.0%200 | +8.70.100 |
| Octagon | -1.20.827 | +2.30.696 | -5.00.386 | +1.30.819 | +1.90.754 | +7.20.193 | -3.20.615 | +3.50.544 | -8.70.100 | 50.3%159 |
Better on the row's stage — deeper at 5 and 10 pointsBetter on the column's stagep < 0.05 on its ownStill there after correcting for how many were tested· fewer than 80 decided matches on one side of the comparison
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.
Matches counted128
Win rate56.2%
Widest swing-56.2 pts
Survive correction0
Period2026/03/26 – 2026/09/23
Where it was fought
Pai's win rate on each stage, against this matchup everywhere| Stage | Stage effect | Win % | Change | p | Matches |
|---|---|---|---|---|---|
| Low fence | 100.0% | +43.8 | 0.122 | 3 · | |
| Rectangle | 71.4% | +15.2 | 0.405 | 7 · | |
| Breakable full fence | 57.9% | +1.6 | 0.876 | 19 · | |
| Octagon | 57.1% | +0.9 | 0.961 | 7 · | |
| Full fence | 55.6% | -0.7 | 0.949 | 18 · | |
| Breakable half fence | 54.5% | -1.7 | 0.905 | 11 · | |
| Open | 54.5% | -1.7 | 0.905 | 11 · | |
| Half fence | 50.0% | -6.2 | 0.564 | 18 · | |
| Single wall | 50.0% | -6.2 | 0.516 | 22 · | |
| Fence and open | 0.0% | -56.2 | 0.106 | 2 · |
p < 0.05 on its ownStill there after correcting for how many were tested
Stage against stage
Pai's win rate on the row's stage, minus the same on the column'sAll ten shapes at once: chi-square 6.15 on 9 df, p 0.725
| Stage | Full | Break. full | Half | Break. half | Low | Wall | Mixed | Open | Rect | Oct |
|---|---|---|---|---|---|---|---|---|---|---|
| Full fence | 55.6%18 | -2.30.886 | +5.60.738 | +1.00.958 | -44.40.142 | +5.60.726 | +55.60.136 | +1.00.958 | -15.90.467 | -1.60.943 |
| Breakable full fence | +2.30.886 | 57.9%19 | +7.90.630 | +3.30.858 | -42.10.159 | +7.90.613 | +57.90.119 | +3.30.858 | -13.50.529 | +0.80.973 |
| Half fence | -5.60.738 | -7.90.630 | 50.0%18 | -4.50.812 | -50.00.105 | +0.01.000 | +50.00.178 | -4.50.812 | -21.40.332 | -7.10.748 |
| Breakable half fence | -1.00.958 | -3.30.858 | +4.50.812 | 54.5%11 | -45.50.145 | +4.50.805 | +54.50.155 | +0.01.000 | -16.90.474 | -2.60.914 |
| Low fence | +44.40.142 | +42.10.159 | +50.00.105 | +45.50.145 | 100.0%3 | +50.00.102 | +100.00.025 | +45.50.145 | +28.60.301 | +42.90.175 |
| Single wall | -5.60.726 | -7.90.613 | -0.01.000 | -4.50.805 | -50.00.102 | 50.0%22 | +50.00.174 | -4.50.805 | -21.40.321 | -7.10.742 |
| Fence and open | -55.60.136 | -57.90.119 | -50.00.178 | -54.50.155 | -100.00.025 | -50.00.174 | 0.0%2 | -54.50.155 | -71.40.073 | -57.10.151 |
| Open | -1.00.958 | -3.30.858 | +4.50.812 | -0.01.000 | -45.50.145 | +4.50.805 | +54.50.155 | 54.5%11 | -16.90.474 | -2.60.914 |
| Rectangle | +15.90.467 | +13.50.529 | +21.40.332 | +16.90.474 | -28.60.301 | +21.40.321 | +71.40.073 | +16.90.474 | 71.4%7 | +14.30.577 |
| Octagon | +1.60.943 | -0.80.973 | +7.10.748 | +2.60.914 | -42.90.175 | +7.10.742 | +57.10.151 | +2.60.914 | -14.30.577 | 57.1%7 |
Better on the row's stage — deeper at 5 and 10 pointsBetter on the column's stagep < 0.05 on its ownStill there after correcting for how many were tested· fewer than 80 decided matches on one side of the comparison
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.