Part one
Winning the qualifier weekend
day one (7-2) → day two (4-1), both required
Two Swiss-style gauntlets, played back to back. Day one is the long one: race to 7 wins, and a third loss knocks you out before you get there. Day two is shorter: race to 4 wins, one loss forgiven. To win the weekend you have to clear both.
Neither gate is symmetric around a coin flip. A 50% player doesn't win a 7-2 race half the time; they win it under 9% of the time, because the race runs long enough that the gap between "roughly even" and "actually ahead" has room to show up in the record. Stack two such gauntlets together and a 50% player's odds of running the whole table fall under 2%.
Watch where the gray line (day one) and the coral line (day two) cross, around an 80% win rate. Below that, day two is the easier gate: fewer wins required means less exposure to a single bad run. Above it, day one flips and becomes the more reliable gate of the two. That's the length of the race working in the stronger player's favor: more total matches gives a real skill edge more room to assert itself over variance. Day two, being shorter, stays noisier for longer, since a single unlucky loss carries more relative weight in a race to 4 than in a race to 7.
Part two
Day two, on its own
race to 4 wins · 1 loss forgiven · 5 matches, max
Pull day two out on its own and the shape is familiar, just compressed. Race to 4 wins with one loss allowed, and the whole thing is over in at most 5 matches, the shortest race anywhere on the path to the championship.
That brevity cuts both ways. A weaker player can occasionally run a short table on a hot streak, so day two's floor is more generous than day one's: the curve sits above day one's line for most of the chart. But a strong player also gets less time to let a real edge compound; the loss cushion in a 5-match race gets used up fast. That's the other side of the crossover from the section above: day two is the stage where variance has the most to say, in both directions.
The 50-50 line falls around a 63% win rate, noticeably lower than the roughly 75% it takes to clear day one alone. If you're the stronger player walking into day two, don't assume day one's arithmetic still protects you: the race is a third as long, and the cushion runs out fast.
Appendix
The probability, spelled out
Every stage here is the same kind of object: a race to w wins, where you're allowed up to L losses along the way and a match win rate of p. The race ends the moment you hit w wins (through) or L+1 losses (out), whichever comes first.
To finish with exactly k losses (for some k from 0 to L), your last match has to be a win, and among the w+k-1 matches before it you need exactly k losses, arranged in any order. There are C(w+k-1, k) ways to arrange those losses, so:
Summing over every allowable loss count covers every way the race could end in your favor. Here's what that gives for each stage on this page:
| Stage | Race | Closed form |
|---|---|---|
| Day one | 7-2 | p⁷·[1 + 7(1-p) + 28(1-p)²] |
| Day two | 4-1 | p⁴·[1 + 4(1-p)] |
| Qualifier weekend | day one × day two | P(7,2,p) · P(4,1,p) |
For reference, the qualifier stage that feeds into Qualifier Weekend (a straight 4-0, no losses forgiven) is just P(4, 0, p) = p⁴; the L=0 case collapses the sum to a single term, since any loss ends it immediately.