The 2026 World Chess Championship has an unusual shape before a single move is played: the reigning world champion is the underdog in his own title defense. The market prices the challenger, Javokhir Sindarov, at 70 cents against sitting champion Gukesh Dommaraju at 29.5 cents. When we first scored this board on August 3, those numbers were 70 and 28 cents. Two weeks of chess have moved this market by about a cent and a half.
The Quick Answer
The market makes Sindarov a 70-cent favorite over Gukesh at 29.5 cents. We re-scored the board with eight AI models, none of which was shown a price, and their blended estimate came back at 70.7% Sindarov, 28.3% Gukesh. That is a near-exact match with the market, and it is not where the panel started. In the first round the seats were split almost in half, with five of the eight putting Sindarov between 58.5 and 62. What moved three of them was not new evidence but an arithmetic error one seat found in the panel's own reasoning.
The interesting part is not that the models agree with the market. It is that they got there from a different direction, and the argument they had on the way names exactly which future results would break the number. The full model-by-model board, the reasoning that changed three seats' minds, and the dated events that would move it are below.
Stay ahead of the markets.
Daily insights and expert picks on Kalshi, Polymarket, and what's moving markets.
Free forever. Unsubscribe anytime.
The Market at a Glance
| The question | Who wins the FIDE World Chess Championship 2026 |
| The players | Gukesh Dommaraju (India), reigning champion, defending the title he won in 2024; Javokhir Sindarov (Uzbekistan), challenger |
| The format | 14 classical games; first to 7.5 points wins. A 7-7 tie is decided by faster rapid and blitz tiebreaks. There is no draw-odds advantage for the champion |
| The schedule | Geneva, Switzerland, November 25 to December 15, 2026; prize fund 2.5 million dollars |
| Settlement | Resolves to whoever wins the match. The contract closes early once a title holder is declared |
| Venue | Kalshi |
| Prices below as of | August 17, 2026, from bid/ask midpoints. Prices on this page refresh as the match approaches, and the stamp here carries the reading these numbers reflect |
Both players will be 20 years old when the first game starts, which makes this the youngest classical World Championship match ever played.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $182M traded
- 2027 Pro Football Champion · $65M traded
- 2028 U.S. Presidential Election winner? · $60M traded
- 2028 Republican presidential nominee · $58M traded
- Pro Baseball Champion · $55M traded
- Florida Republican Governor nominee? · $43M traded
Every market above links to our full AI model verdict; browse them all on the OddsShopper prediction markets hub, and see how every settled call actually scored on the full graded scoreboard.
Why the Reigning Champion Is an Underdog
The August 2026 FIDE rating list is blunt about it. Sindarov sits at 2777, world number 4. Gukesh sits at 2703, world number 31. That is a 74-point gap in the challenger's favor, and in chess a gap that size between elite players is a real structural edge rather than a rounding difference.
The path each man took to Geneva points the same way. Sindarov earned the challenger's seat by winning the 2026 Candidates Tournament with 10 points from 14 games, six wins and no losses, clinching with a round to spare. It is the highest score anyone has posted in the current double round-robin format. Gukesh, meanwhile, finished last at Norway Chess and went winless at the Chennai Grand Masters in July, losing three and winning none, a result that cost him 14 rating points and dropped him behind a 15-year-old on the world junior list.
Their classical head-to-head is three games: one Sindarov win and two draws. Gukesh has never beaten Sindarov at the classical time control, though he did beat him in a rapid game in May 2026.
That rapid win is worth holding onto, because a 7-7 tie in Geneva sends the title to a four-game rapid playoff. It is also the one place where a champion's escape route ought to be, and the ratings do not offer him one: on the same August list Sindarov is world No. 6 at rapid to Gukesh's No. 22, and world No. 16 at blitz to Gukesh's No. 39. The faster the chess gets, the less the gap closes.
Read against those facts, a challenger priced above a sitting world champion stops looking strange.
What Usually Happens to Defending Champions
The base rate cuts the other way, and it is worth being precise about how far. Since the title was reunified in 2006 there have been ten matches with a defending champion on the board. The champion kept the title in eight of them. Challengers are 2-for-10.
But the two exceptions are not random, and this is where a base rate earns its keep. In 2013 Viswanathan Anand was the champion at 2775 and world No. 8, defending against a 2870-rated Carlsen who was world No. 1; Anand lost 6½–3½ without winning a game. In 2024 Ding Liren defended at 2728, having fallen from 2788 and from world No. 3 to No. 23 after a bad run at Tata Steel, Norway Chess and Grenke; he lost the title in the final game to Gukesh.
Both title changes share a shape: the sitting champion was the materially lower-rated player and visibly out of form, and the challenger was near the top of the rating list. That is the exact configuration on the board today, with one difference worth noticing. Ding went into his losing defense 55 rating points behind. Gukesh goes into his 74 points behind.
So the honest reading of "champions win 80% of the time" is that champions who are still among the best players in the world win almost all of them, and champions who arrive the way Gukesh is arriving have gone 0-for-2 in the modern era. This history was not in the fetched card the panel scored from, which covered ratings, form, the head-to-head and the format; the two seats that stayed lowest on Sindarov reached for a version of it anyway, arguing that a defending champion has already won a match at this exact format and his opponent has never played one.
The Form Check: The Challenger's Cold Start, and What Followed
The freshest evidence on either player is being generated right now. Sindarov is in the field at the Sinquefield Cup in St. Louis, a ten-player classical round-robin. The crosstable as fetched on August 17:
| # | Player | Score | Performance rating |
|---|---|---|---|
| 1 | Wesley So | 4.5 | 2849 |
| 2 | Maxime Vachier-Lagrave | 4.0 | 2796 |
| 3 | Rameshbabu Praggnanandhaa | 3.5 | 2800 |
| 4 | Fabiano Caruana | 3.5 | 2799 |
| 5 | Levon Aronian | 3.5 | 2756 |
| 6 | Samuel Sevian | 3.5 | 2755 |
| 7 | Javokhir Sindarov | 3.5 | 2744 |
| 8 | Vincent Keymer | 3.5 | 2741 |
| 9 | Anish Giri | 3.0 | 2699 |
| 10 | Jorden Van Foreest | 1.5 | 2513 |
A performance rating is what a player's results in that event alone are worth, and reading this table is a lesson in how fast a small sample changes its mind. Earlier in the event Sindarov sat 9th of 10 on 2.5 points with a 2683 performance rating, the second-worst in the field. He has since won, which lifted him into a six-way tie for third on 3.5 and pulled his performance rating up 61 points to 2744. One game did that.
That is the whole argument about form in a single table. His 2744 is still running about 33 points below his own official rating, so this is not a hot streak either. His live rating, which moves with every game rather than waiting for the monthly list, sits at 2768.9 and his live world ranking at number 6, narrowing the gap between the two match players from the official 74 points to 65.9.
The panel scored this board while the cold stretch was at its worst, and three seats spent their revision round arguing that a performance rating over eight or nine games is too noisy to act on. The very next result moved it 61 points. Readers can decide for themselves whether that settles the argument or proves it.
The Board
| Outcome | Context | Market | AI blend |
|---|---|---|---|
| Javokhir Sindarov | Challenger; 2777, world No. 4; 2026 Candidates winner | 70¢ | 70.7% |
| Gukesh Dommaraju | Reigning champion; 2703, world No. 31 | 29.5¢ | 28.3% |
| Hikaru Nakamura | Not in the match; withdrawal and replacement tail | 1.5¢ | 0.35% |
| Field (six others) | Caruana, Giri, Praggnanandhaa, Wei Yi, Esipenko, Bluebaum | 0.5¢ each, ~3¢ combined | ~0.8% combined |
Every Seat's Number
Each seat priced all nine outcomes from one view of the event, without being shown a price, then read the other seats' anonymized reasoning and was free to revise. The numbers below are after that revision round.
| Outcome | Fable | Opus | Sonnet | ChatGPT | Gemini | GLM | Kimi | DeepSeek | Blend |
|---|---|---|---|---|---|---|---|---|---|
| Javokhir Sindarov | 62.00 | 73.00 | 71.43 | 61.00 | 75.00 | 68.00 | 70.00 | 72.00 | 70.71% |
| Gukesh Dommaraju | 37.00 | 26.00 | 28.57 | 38.00 | 23.00 | 30.00 | 28.00 | 26.50 | 28.29% |
| Fabiano Caruana | 0.30 | 0.20 | 0.00 | 0.30 | 1.00 | 0.50 | 0.50 | 0.40 | 0.35% |
| Hikaru Nakamura | 0.30 | 0.20 | 0.00 | 0.25 | 1.00 | 0.50 | 0.50 | 0.40 | 0.35% |
| Anish Giri | 0.10 | 0.30 | 0.00 | 0.20 | 0.00 | 0.40 | 0.40 | 0.20 | 0.20% |
| Rameshbabu Praggnanandhaa | 0.10 | 0.10 | 0.00 | 0.10 | 0.00 | 0.30 | 0.30 | 0.15 | 0.10% |
| Wei Yi | 0.10 | 0.10 | 0.00 | 0.07 | 0.00 | 0.20 | 0.20 | 0.15 | 0.10% |
| Andrey Esipenko | 0.05 | 0.05 | 0.00 | 0.05 | 0.00 | 0.10 | 0.05 | 0.10 | 0.05% |
| Matthias Bluebaum | 0.05 | 0.05 | 0.00 | 0.03 | 0.00 | 0.00 | 0.05 | 0.10 | 0.04% |
Every seat on this panel is graded against real market settlements — records to date: Claude Fable 84% on 508 graded calls · Claude Opus 85% on 578 graded calls · Claude Sonnet 84% on 556 graded calls · GPT 84% on 5,500 graded calls · Gemini 86% on 4,292 graded calls · GLM 81% on 2,367 graded calls · Kimi 83% on 2,345 graded calls · DeepSeek 80% on 2,392 graded calls. Recomputed daily; the full scoreboard is public.
These are model estimates, not predictions of fact and not financial advice. The blend is the seat median for each outcome, so the nine figures need not total exactly 100. Every number in this piece gets graded in public once the market settles; you can check the panel's working against the full graded scoreboard.
More live boards from the same panel: 2026 Ballon d'Or, the other big individual-honor field, where Harry Kane leads at 61.5¢ · Nobel Peace Prize 2026, a 21-name field where Sudan's Emergency Response Rooms lead at 27¢. Prices fetched August 17, 2026.
Where the Panel Changed Its Mind
The first round split the panel almost in half. Five seats put Sindarov between 58.5 and 62, reasoning that a 14-game match is draw-heavy and that draws compress a rating edge. Three seats sat between 70 and 78.5. The gap was not about evidence, since every seat read the same card. It was about how to convert a rating edge into a match probability.
Sonnet, which did not move, spelled out why it thought the low group had the mathematics backwards:
Held at 70%. The strongest opposing argument (Forecaster G's higher elite-match draw rate compressing Sindarov's edge to ~58%) doesn't survive a direct calculation: with expected 14-game score pinned at ~8.3/14 by the rating gap regardless of assumed draw rate, a higher draw rate reduces variance around a mean already above the 7.5 threshold, which helps the favorite rather than hurting him. — Claude Sonnet (84% on 556 graded calls)
[Editor's note: this seat's nine raw numbers summed to 98 rather than 100, so the table above shows them rescaled to 100 — which is why a seat that says it held at 70% appears there at 71.43%. The quote is its own wording, unaltered.]
That argument, plus a statistical point about how noisy a short performance rating is, moved three seats a long way. Kimi went from 58 to 70, and was careful to say it was correcting its own arithmetic rather than drifting toward the group:
My round-1 tiebreak probability of 20-25% is what a near-even match produces; with a 66-74 Elo gap and realistic draw rates, exact binomial calculation puts the 7-7 path at ~6-12% and Gukesh's classical win path at ~3-10%. — Kimi (83% on 2,345 graded calls)
GLM, which went from 60 to 68, landed on the sharpest version of the point about the slump:
The live rating gap of 65.9 points already incorporates his Sinquefield Cup results, so using that gap AND separately downgrading for the slump double-counts the same evidence. Forecaster D's point that an 8-game TPR has a standard deviation of ~100 Elo is decisive: Sindarov's 2683 performance is only ~0.94 SD below his rating, which is statistically insignificant noise, not evidence of durable decline. — GLM (81% on 2,367 graded calls)
[Editor's note: the panel scored this board when Sindarov's Sinquefield performance rating was 2683. He won his next game, and it now reads 2744. Model quotes are published as submitted, so the figure in the quote is the one the seat was given.]
Not everyone was persuaded, and the holdouts are the reason this board still has a spread. ChatGPT stayed at 61 and said plainly that it thinks the high group is running a model that flatters favorites. Fable held at 62 and made the cleaner methodological objection: the disagreement is not about the evidence at all.
The panel's spread (58-78) is entirely about how much to shade raw Elo-to-match conversion, not about different evidence. The high estimates (75-78) apply the Elo expectancy almost at face value: ~0.59 per game over 14 games does mechanically produce ~75-85% raw. But I deliberately shade the effective gap. — Claude Fable (84% on 508 graded calls)
Opus, the seat that first made the variance argument, trimmed itself from 75 to 73 on a wrinkle nobody else had priced:
Gukesh's 2703 is static because he is idle and has therefore fully digested Norway and Chennai, while Sindarov's 2768.9 is still falling mid-event and has NOT finished digesting Sinquefield. That asymmetry means the live gap is the most favourable-to-Sindarov reading available. — Claude Opus (85% on 578 graded calls)
The panel's final spread runs from 61 to 75, which is wide enough to be honest about. The median landing within a point of the market is the finding, not a coincidence the piece should oversell: eight models reasoning from ratings, form and format, with no price in front of them, arrived where the traders already were.
What Would Change the Panel's Mind
Four things between now and November are checkable, dated, and would move the number in a known direction.
- How Sindarov finishes at the Sinquefield Cup (by August 20). He has rounds left and has already clawed back from 9th to a tie for third. A winning finish restores the live rating he shed and settles the noise argument in the high seats' favor. Two more losses push his live rating below 2760 and hand the seats that shaded him a second data point instead of one. Direction: toward Gukesh if it turns bad again.
- The Grand Chess Tour Finals, August 22 to 27 in St. Louis. The next classical or rapid evidence on the tour, arriving within days of the Sinquefield finish.
- The 46th Chess Olympiad, September 15 to 27, in Samarkand. Gukesh is named on India's team alongside Arjun Erigaisi and Praggnanandhaa. It is the champion's last major classical outing before Geneva and the clearest read available on whether his slump is over. A strong Olympiad is the single event most likely to compress this board. It is also being held in Sindarov's home country.
- Any withdrawal or replacement announcement. The seven names who are not playing hold about 1.2% between them, and that number exists only because a scheduled two-player match can be disrupted. A confirmed withdrawal would reprice the whole board at once.
Settlement Timeline
| Date | What happens |
|---|---|
| August 20, 2026 | Sinquefield Cup ends |
| August 22-27, 2026 | Grand Chess Tour Finals, St. Louis |
| September 15-27, 2026 | 46th Chess Olympiad, Samarkand |
| November 25, 2026 | Game 1 in Geneva |
| December 15, 2026 | Final scheduled day of the match |
| Contract expiration | December 31, 2026, closing early once a title holder is declared |
This board is re-scored when the story moves, and the panel's numbers above are as of August 17, 2026.
Dark Horses the Panel Won't Dismiss
Seven names on this board are not playing in the match. They can win the title only if Gukesh or Sindarov withdraws and is replaced. That is worth being precise about rather than waving away, because the panel did not price them identically.
- Hikaru Nakamura (1.5¢, blend 0.35%) carries the largest market bid of the non-participants, which makes sense for the most famous active player in the sport. Every seat priced him near zero. His price is a "what if something extraordinary happens" number rather than a live contender read, and at 1.5 cents it is the one non-participant the market prices materially above the panel.
- Fabiano Caruana (0.5¢, blend 0.35%) is tied with Nakamura for the highest rating on the board at 2792, above both match players, and is playing well at the Sinquefield Cup. None of that gets him into a match he did not qualify for. Gemini was the one seat to give him and Nakamura a full 1%, on the reasoning that a replacement would plausibly be drawn from the highest-rated available players.
- Anish Giri (0.5¢, blend 0.20%) and Rameshbabu Praggnanandhaa (0.5¢, blend 0.10%) are elite players who lost their path to this match in the Candidates. They hold the exchange's minimum tick.
- Wei Yi, Andrey Esipenko and Matthias Bluebaum round out the tail at a blended 0.05% to 0.10% each. Sonnet priced all seven at a flat zero, which is the most aggressive read on the board and the one worth disagreeing with: a scheduled match five months out is not a certainty.
The Traps: Numbers That Can Mislead
- The 74-point rating gap is already stale. It comes from the August monthly list. The live gap is 65.9 and still moving, because one of the two players is mid-tournament and the other is idle. Quoting the official gap today overstates the challenger's current edge.
- A performance rating over a handful of games is noisy. Three seats made the same point independently: the standard deviation on an 8-to-9-game performance rating is roughly 100 rating points, so even a 94-point shortfall sits inside one standard deviation. This board proved the point in real time, with one win moving that number 61 points. It is the freshest evidence available and it is also the weakest.
- Both players are in poor form, not one. Gukesh finished last at Norway Chess and went winless in Chennai. Reading the challenger's cold stretch without weighing the champion's bad season gets the comparison wrong in the champion's favor.
- The thin end of the board is thinner than it looks. Six of the seven non-participant contracts sit at a 0-to-1-cent bid/ask, with Nakamura the exception at 1 to 2 cents. A one-cent quote on a market with no meaningful two-way interest is a minimum tick, not a considered price.
How to Read These Prices
A price in cents is the market's rough estimate of the chance an outcome happens, so 70 cents is a shade under a 2.5-to-1 favorite and 29.5 cents is a little worse than 1-in-3. In sportsbook terms the challenger is around -233 and the champion around +239. These contracts are event contracts on a CFTC-regulated exchange, not sportsbook bets, and availability varies by state.
Two timing points matter on a market that does not settle until December. A position here ties up capital for four months with no interim payout, and a single decisive game in Geneva can move the price hard in either direction. The seven non-participant contracts also carry wide relative spreads, so the quoted midpoint on those is a weaker signal than it looks.
More on this: MAC Championship Odds: Ohio Never Has To Play The Favorite · Who Wins The 2027 NBA Title? Kalshi Championship Odds · Kalshi Tour Championship Round 1 Scores: Morikawa 4-For-4 · Tour Championship Top 10 Finishers Kalshi Odds: Final Round · World Series Odds: Lindy Vs. The Machine
The Bottom Line
The market and the models are looking at the same match and, after eight price-blind seats argued with each other, they have arrived at the same number. The challenger is favored because on ratings, on his path here, and on a classical head-to-head the champion has never won, Javokhir Sindarov is the stronger player right now.
What is worth watching is how little either side of this board moves on news. The challenger spent a week playing the worst chess he has played since he qualified, and the price did not budge; he then won a game that pulled his tournament performance rating up 61 points, and it did not budge for that either. The panel's read is that both facts are noise inside one standard deviation and that the live rating already absorbs them. That is a testable claim, and the Olympiad in September tests it properly, on the champion this time.
If this method is your thing, our 2026 Ballon d'Or model verdict and our Nobel Peace Prize 2026 verdict score their fields the same price-blind way, and our guide to how prediction markets work is the natural next read.
Prices on this page are Kalshi's book. If you also trade on Polymarket, code OS4 gets new users a $50 trading bonus on a $10 deposit — affiliate link; terms as stated by Polymarket; 18+, availability varies by state.
