We counted 222 Worlds matches. Seeding works, until it doesn’t

Chris Beaumont

The short answer: Higher seeds won 172 of 222 results in our Worlds round-of-64 and round-of-32 dataset, a 77% success rate. Women’s singles supplied the most early chaos, with 10 higher-seed losses in 32 round-of-64 matches, but the men’s singles exits of Ben Johns, Tama Shimabukuro and Zane Ford made the men’s bracket feel more volatile on Wednesday.

The easiest way to misunderstand a tournament upset is to remember only the names.

Ben Johns losing to Noe Khlif is a story. So is Donald Young beating Tama Shimabukuro and Luc Pham overwhelming Zane Ford. But those results acquire a different meaning when placed beside every other match rather than arranged into a dramatic montage.

So we went back through the official Worlds results and counted the higher-seed wins in men’s singles, women’s singles, men’s doubles, women’s doubles and mixed doubles across the rounds of 64 and 32.

The headline is 172 from 222: 77%. Seeding worked roughly three times out of four. The more useful answer is where it did not, and what the raw percentage cannot tell us.

This is a snapshot through Wednesday’s round of 32, before Thursday’s round of 16. It is best read as an interim audit, not a final verdict on the 2026 Worlds.

Our count, division by division

The fixed slate dataset contains 219 matches with scores and three no-score advances in which the lower seed progressed. Those three are Wong over Oncins in men’s singles, Medina Alvarez and Growcott over Martinez Vich and Acevedo in men’s doubles, and Trevino and Dunlap over Jones and Smith in women’s doubles.

That convention needs stating because a withdrawal is not a completed match. Exclude those three rows and the rate becomes 170 higher-seed wins from 219 scored matches: 78% when rounded. Two further no-score advances favoured higher seeds earlier in the supplied brackets; include every no-score advance and the figure is 174 from 224, also 78% when rounded.

The men’s doubles round-of-64 sample has 30 rows in the supplied bracket. We have not invented two missing results to make the table look tidier. These boundaries are why the count is reproducible rather than merely impressive in a headline.

Women’s singles was the real early outlier

Wednesday’s men’s results dominated attention, but women’s singles produced the least seed-friendly opening round. Ten of 32 higher seeds lost at the round-of-64 stage.

The results ranged across the draw. Natalia Simson, seeded 52, beat 13th seed Genie Bouchard. Shannon Pretorius, at 61, defeated fifth seed Chao Yi Wang 11-2, 0-11, 11-7. Samantha Parker, seeded 45, edged Victoria Dimuzio 12-10 in the deciding game. Danni-Elle Townsend, Alex Walker and Kei Sawaki also beat higher-seeded opponents.

That matters because it distributes the volatility beyond one famous match. A single upset can be an exceptional performance or a bad day. Ten in one round suggest a draw where seeding position offered less protection, even though the results cannot tell us why.

The women’s second round then split exactly the same way as the men’s: 12 higher-seed wins from 16. Milan Rane beat Hannah Blatt, Sofia Sewing beat Judit Castillo, Helena Jansen beat Sahra Dennehy and Pretorius advanced again. Yet Anna Leigh Waters, Kate Fahey, Catherine Parenteau and Kaitlyn Christian all progressed.

Pretorius is the thread connecting the two rounds. She followed her three-game defeat of Wang with an 11-7, 11-8 win over Allison Phillips. That is different from a single bolt from the blue: it is a lower seed constructing a route through consecutive rounds.

Waters, meanwhile, beat Kelly Goodnow 11-0, 11-1. The same tournament can therefore contain apparent chaos and near-total control at opposite ends of the bracket. Averages describe the field; they do not flatten those extremes.

The men’s upsets were fewer but louder

Men’s singles actually recorded the joint-highest overall seed success rate, 39 wins from 48 results or 81%. That seems counterintuitive after Johns, Shimabukuro and Ford all departed within hours.

The explanation is visibility, not arithmetic. Khlif beat Johns 11-8, 11-8; Young beat Shimabukuro 12-10, 11-7; Pham beat Ford 11-1, 11-3. Those matches involved a global star, the recent Hong Kong finalist and a top-10 seed respectively. They naturally command more attention than a quiet higher-seed win elsewhere.

There was also a clustering effect. Rafael Lenhard beat Dylan Frazier and Harrison Brown beat Grayson Goldin in adjacent seeding bands. Four lower-seed advances in one round are enough to redraw likely quarter-final routes, even when 12 favourites still progress.

This is where seeding analysis needs restraint. A seed is an ordering, not a promised probability. The difference between 16 and 17 is far smaller than the difference between one and 61, and a numerical upset does not automatically imply a large performance gap.

Pham’s 11-1, 11-3 score is a stronger statement about that particular day than the mere fact that No 18 beat No 10. Conversely, Lenhard’s 12-10 deciding game against Frazier tells us how narrow the margin was. Treating both as identical units called “upsets” would lose the most useful information.

Doubles shows why format matters

Men’s doubles returned 36 higher-seed wins from 46 rows, 78%. Women’s doubles delivered 26 from 32, 81%, while mixed doubles was 37 from 48, 77%.

Those rates are close enough that we should resist declaring one partnership format inherently more predictable from this sample alone. The entry structures differ, the supplied men’s doubles round-of-64 sample is smaller, and teams do not transfer seamlessly between gender doubles and mixed.

There is nevertheless a practical observation. Doubles seeds carry information about two players and their partnership, not just an individual’s broad reputation. A strong return team can neutralise an opponent’s best patterns; a familiar pairing can disguise an ordinary singles seeding through complementary roles.

The World Pickleball Ranking model published with the tournament weights gender doubles at 50%, mixed at 35% and singles at 15%. That should not be confused with a documented formula for every seed in every draw. It does, however, illustrate why composite standing and event-specific expectations must be separated.

When Johns entered men’s singles at 19, his achievements elsewhere inevitably shaped public expectations. The bracket itself had already assigned a more modest position. His defeat was an upset by seed, but not an upset from No 1 to an unheralded outsider.

Round two was not more chaotic

Another appealing theory collapses under the count. Higher seeds won 109 of 142 round-of-64 results, 77%, and 63 of 80 round-of-32 results, 79%.

The second round therefore became marginally more seed-friendly, not less. The raw number of higher-seed losses fell from 33 to 17, although the rounds have different denominators and should be compared by rate.

This is a useful reminder of selection bias in tournament coverage. Early on, dozens of matches compete for attention. By the round of 32, every match is closer to the business end and familiar names face one another. Each surprise feels larger even when the overall upset rate falls.

It also means Thursday’s last 16 should not be previewed as though randomness is now the dominant force. Johns is gone, but Christopher Haworth, Federico Staksrud, Hunter Johnson and Christian Alshon remain in the men’s singles draw. Waters, Fahey, Parenteau and Christian remain in the women’s.

What this count cannot establish

It cannot prove that seeding is “right” 77% of the time. Higher-seed wins measure agreement with an ordering, not the quality of the system that produced it. A flawed ranking can still favour the better player in most matches.

Nor can it compare divisions like-for-like without context. Entry numbers, withdrawals, partnership changes and the strength of the field below the top seeds all influence the result. One tournament is a snapshot, not a stable long-run probability.

Most importantly, the numbers do not explain technique or tactics. They cannot tell us whether a return was too short, whether a player struggled moving laterally or whether a doubles specialist had prepared enough singles patterns. Match footage and player evidence would be needed for that.

What the count can do is discipline the conversation. It prevents three famous men’s results from becoming a claim that the entire draw is broken. It directs attention towards women’s singles, where the opening round genuinely was more volatile. And it gives us a baseline to update after Thursday.

Seeding is working often enough to remain useful. It is failing often enough to keep the draw alive. The interesting question is not why favourites sometimes lose; it is which parts of the field consistently make their seeding least informative.

Through Wednesday, the answer is women’s singles in the opening round. Through the next two days, the answer may change. That is why this count should grow with the tournament rather than freeze into a convenient narrative.