We ran the 2026 Wimbledon Gentlemen's Singles 50,000 times — a random 128-draw knockout, each match decided by an Elo win model — with Alcaraz, Rune and Musetti out. Our simulation rates Sinner ~41% to the title; the market says ~58%. We lock a calibrated ~51% ensemble and surface the simulation's signal: across seven rounds, the field is more dangerous than the bookmakers price. The companion to P-14's locked anchor — the champion will not be Alcaraz.
50,000 random 128-draw knockouts, Elo per match; Alcaraz/Rune/Musetti excluded (withdrawn). Our draw-sim: Sinner 41.2% title / 50.0% final / 59.0% SF; Djokovic 13.7%; the field 21.1%. We lock a ~51% ensemble (market-anchored) and flag the divergence: the simulation sees more upset variance than the book.
We simulate the full 128-draw 50,000 times and lock a market-anchored ensemble. Sinner is the clear favourite, but our draw-simulation rates the field meaningfully more dangerous than the bookmakers — a falsifiable divergence.
| Player | Sim P(title) | P(final) | P(SF) |
|---|---|---|---|
| Jannik Sinner | 41.2% | 50.0% | 59.0% |
| Novak Djokovic | 13.7% | 22.8% | 33.4% |
| Alexander Zverev | 6.7% | 13.4% | 22.7% |
| Daniil Medvedev | 2.9% | 7.2% | 14.0% |
| Ben Shelton | 2.6% | 6.5% | 13.1% |
| Field (116 others) | 21.1% | — | — |
Each simulated tournament shuffles the 128-player field into a bracket and plays seven rounds, every match decided by P(i beats j) = 1/(1 + 10^((Eloⱼ − Eloᵢ)/400)). Grass Elos for the named contenders are calibrated so the title distribution tracks the de-vigged market; the bulk of the draw is a bottom-heavy field (qualifiers and low seeds). The draw itself isn't public until ~26 June, so we simulate fully random draws — seed protection (which keeps the top players apart until late rounds) isn't modelled yet, so our field-danger read is, if anything, an upper bound, and once seeded the favourite's true title odds would tick up (part of why we anchor the lock toward the market).
Verified 2026-06-20: Sinner 4/7 favourite, Djokovic 5/1, Zverev 9/1. Withdrawals confirmed: Alcaraz (wrist), Rune (Achilles/knee), Musetti (thigh) — all excluded. Final Sunday 12 July. Sources: outright market + entry list.
The book makes Sinner a ~58% favourite. Our simulation, run 50,000 times, says ~41% — because a best-of-five run over seven rounds, against a field collectively worth 21%, carries more upset variance than a single headline price implies. We lock a ~51% ensemble (leaning to the market for the favourite while honouring the sim's field-danger signal). If Sinner wins comfortably, the market read scores better; if a name from the field breaks through, the simulation does.
The simulator is a committed, fixed-seed script (scripts/sim/wimbledon-2026-mens.py). We synthesise and calibrate public signals (market + Elo); we do not claim a proprietary tennis model. Scored by Brier against the official Wimbledon result. P-14 already locks the clean anchor — the champion will not be Carlos Alcaraz, who is not in the draw.
Sinner cruises — if he wins as the market expects, our ~51% ensemble is well-calibrated but our 41% sim looked timid. A field upset — Djokovic's grass mastery or a big server (Shelton/Fritz) on a hot streak takes it, vindicating the sim's higher field-danger read. Either way we score the locked distribution honestly, and P-14's anchor (not Alcaraz) resolves alongside.
RAOSCAFF locks P-18 on 2026-06-21, before the ~26 June draw. The simulator is committed for reproducibility. We score the locked title + reach-final distribution by Brier against the official Wimbledon Gentlemen's Singles result on 12 July.
Random-draw simulation (the bracket isn't public yet) — fixed seed, committed script, checkable.