Updated August 25, 2026 · 12 min read · by Sam Smith. Panel verdicts were generated August 25, price-blind, from the series' own episode ledger. Kalshi prices as of 6:04 p.m. ET, August 25, 2026.
In January, on a public course in Burleson, Texas, Bryson DeChambeau told his YouTube audience what the year was going to look like. "I'll have six chances to break as many course records as I possibly can," he said, and then named the target: three. Three course records, one calendar year, one round at each course, no practice round, usually no prior knowledge of the property at all.
Five of those six chances have now aired. He has two.
That leaves one planned attempt to finish the job, and it turns out a lot of people have taken a position on how it ends. Kalshi, the federally regulated exchange that is also the named sponsor of the series this season, lists a five-rung market on the final count: will he finish 2026 with two or more, three or more, four or more, five or more, or six or more. Across those five contracts, 223,665 have changed hands, per the exchange's own volume figures. The busiest rung is not the one everybody is arguing about. It is the one that asks whether he gets to six.
That is the strange part of this board, and it is where our panel and the market end up looking at the same series and describing two different seasons. The full ladder, every model's number, and the argument that changed most of their minds are below.
Free: The Weekly PM Market Brief — the 8-model panel's graded record, the week's biggest market-vs-model gaps, and what's spiking next. One email, Sundays.
New to exchange-traded event contracts? Kalshi sports betting in four steps explains how a price in cents works before you read a board like this one.
The Five Rounds That Got Him Here
The series is called Course Record, and the format has not changed since it started in August 2024: one round, one try, and the course's standing record is the number to beat. Ties count. That matters more than it sounds like it should, because the first record of 2026 was one.
Season one ran eight episodes between August 2024 and September 2025 and produced two: a tie in the Dallas-Fort Worth area in March 2025, and one outright break in June 2025. That is the whole history of the format before this year, and it is not a flattering one. Add both seasons together and it comes to four conversions from thirteen attempts, or 30.8 percent.
The reason a tour professional converts less than a third of these is the rule he set for himself. A course record is usually posted by a very good player who has learned the property over years, on a day the wind cooperated. He gets one round, cold, on a golf course he has typically never seen, with the number already on the wall. Season two has asked him to beat 62, 63, 64 and 65 on four different kinds of golf course, from a Dallas municipal track to a 7,100-yard resort at altitude in the Canadian Rockies.
The Cedar Crest round in May is the one that shows what that costs him. He asked at the club what the record was and got told 62, not the 63 he had been given. "They told me 63," he said. "Son of a gun." A tour professional walked a Tillinghast municipal course blind, needing to go round in 62 or better on a golf course he had never seen, and shot three under.
Here is the complete 2026 ledger, taken from the channel's own Course Record Season 2 playlist rather than from video titles, which the channel A/B tests and rewrites. Outcomes come from each episode's own audio.
| Aired | Course | Record to beat | Result |
|---|---|---|---|
| Jan 14 | Southern Oaks Golf Course, Burleson TX | 63 | Tied. Shot 63 with a caddie. "Ties do count towards a course record" |
| Mar 30 | Texarkana Country Club, "Little Augusta" | not stated on camera | Missed. Shot 69 on a 6,969-yard course, a week before the Masters |
| May 11 | Cedar Crest Golf Course, Dallas | 62, held by Lee Carter | Missed by a distance. Three under on a 1927 PGA Championship host |
| Jun 15 | Tenison Park Golf Course, Dallas | 65 | Missed. Needed birdie-birdie to tie and got neither |
| Jul 27 | Silvertip Resort, Canmore, Alberta | 64 | Broken. Shot 63 at 5,000 feet. "Nine under and have the course record" |
Two from five. He said the number out loud at Silvertip, after his first record of any kind since January: "So, I'm two for five."
What The Contract Actually Counts
The settlement language is narrower than the question sounds. A contract pays out on records "broken or tied," so the Southern Oaks round counts in full. What counts is the upload, not the round: every round posted to the series during 2026 is in play regardless of when it was played, and a round played this year but posted in 2027 is not. The market expires on December 31, 2026, at 10:00 a.m. ET, and each rung can settle early the moment its threshold is hit. These are CFTC-regulated event contracts, open to traders aged 18+, and state availability varies by platform.
The Board
We put the whole ladder in front of eight AI models at once, price-blind. None of them saw a Kalshi quote. Each was handed the same fetched card, containing every episode of both seasons with the outcome in DeChambeau's own words, the record at each course where the episode stated it, the upload cadence, and the contract language, then asked to price all five rungs from a single view of the season rather than one rung at a time.
| Outcome | What it needs | Last trade | Panel |
|---|---|---|---|
| 2+ | already banked | 99¢ | 99.5% |
| 3+ | one more conversion | 59¢ | 50.8% |
| 4+ | a seventh episode, and two more conversions | 5¢ | 6.5% |
| 5+ | three more, so at least two extra episodes | 2¢ | 0.8% |
| 6+ | a perfect finish that does not exist yet | 2¢ | 0.1% |
The 3+ contract is where almost every headline lands, and it is also the one where the market is hardest to pin down. Buyers are bidding 31 cents and sellers are asking 60, a gap wide enough that the "price" depends on which side you read. The last trade printed at 59 cents, roughly a -144 moneyline. The panel's number after a revision round is 50.8 percent, which sits inside that spread rather than outside it. On a rung this thinly quoted, both readings are defensible and neither is news.
That contract has also moved violently, and Kalshi's daily price history shows the moves tracking the upload schedule rather than anything about golf. It traded at 13 cents from July 10 through July 20, in the dead air after the Tenison Park miss, when the season looked finished. On July 27, the day the Silvertip episode posted, it printed 62. It has spent most of August between 59 and 70.
Here is every seat, after the revision round. The last two columns are the two assumptions that actually build each ladder: how many total attempts that model expects to air in 2026, and the conversion rate it applied to a single attempt.
| Seat | 2+ | 3+ | 4+ | 5+ | 6+ | Attempts | Rate |
|---|---|---|---|---|---|---|---|
| DeepSeek | 99.9% | 55% | 1.9% | 0.2% | 0.1% | 6.6 | 35% |
| Claude Sonnet | 99.5% | 55% | 6.5% | 0.7% | 0.15% | 6.7 | 35% |
| GLM | 99.5% | 51% | 8.5% | 1.5% | 0.3% | 6.6 | 35% |
| Kimi | 99.5% | 51% | 7% | 0.6% | 0.1% | 6.7 | 35% |
| ChatGPT | 99.6% | 50.5% | 7.2% | 0.9% | 0.2% | 6.6 | 36% |
| Gemini | 99.5% | 48% | 6.5% | 0.8% | 0.1% | 6.6 | 35% |
| Claude Fable | 99.5% | 47% | 6.5% | 0.9% | 0.15% | 6.6 | 35% |
| Claude Opus | 99.5% | 46% | 5% | 0.6% | 0.1% | 6.4 | 35% |
| Panel Median | 99.5% | 50.8% | 6.5% | 0.8% | 0.1% | 6.6 | 35% |
The panel row is a seat median, not an average: the median is what the panel's own accuracy audit says to blend on, because one weak seat drags a mean around and a median does not.
Six of the eight seats land within four points of a coin flip on the 3+ rung, and that convergence is the finding rather than a shrug. The arithmetic is tight and every seat showed its work: one announced attempt remains, the conversion rate across thirteen career attempts is 30.8 percent and 40 percent across this season's five, and the seats settled on roughly 35 percent for a single try. One attempt at 35 percent, plus a modest chance that a miss produces a bonus episode, lands almost exactly on 50. The edges net out because the two forces point in opposite directions and are close to the same size: a final goal-deciding round is probably staged at a softer course, which pushes the number up, and he has already missed a reachable 65 this season, which pushes it back down.
The seats were not defaulting. On the same board they spread from 1.9 to 8.5 percent on the 4+ rung, a four-fold disagreement, which is where their real argument lives.
Every seat on this panel is graded against real market settlements — records to date: Gemini 86% on 4,760 graded calls · GLM 81% on 2,504 graded calls · Kimi 83% on 2,494 graded calls · DeepSeek 80% on 2,542 graded calls. Recomputed daily; the full scoreboard is public.
These are model estimates, not predictions of fact and not financial advice. Every number on this page gets graded in public once the market settles; the full graded scoreboard carries the running record.
More live boards from the same panel: Tour Championship first-round leader · Kalshi golf picks this week. DeChambeau's other live Kalshi contract, on winning a PGA Tour major in 2027, last traded at 25¢. Prices fetched August 25, 2026.
Where The Panel Changed Its Mind
Every one of the eight seats revised something after reading the other seven anonymously. What moved them was not the golf. It was a structural point about why extra episodes get made at all, and once one seat said it out loud, most of the panel rebuilt its ladder around it.
"I mixed a binomial over an unconditional episode-count distribution, but extra episodes are not drawn independently of the golf. A seventh attempt gets filmed mainly BECAUSE the sixth missed and he is short of the stated three-record goal, so the extension mass lands exactly in the worlds that still need a hit. Rebuilding it conditionally puts 3+ at about 46 rather than 42, while pushing 4+ DOWN, because a fourth record requires an episode that exists after the goal is already met." — Claude Opus
Read that again as a claim about the shape of the board, because that is what it is. Extra attempts and extra records are negatively correlated. He films a seventh episode if the sixth fails, which is precisely the world where he still only has two. He stops if the sixth succeeds, which is precisely the world where he has exactly three. The two things that would produce a fourth record almost never co-occur.
"I rebuilt the ladder as an explicit branching tree instead of a fixed-attempts-times-fixed-rate mixture, and found that continuation is driven mainly by MISSING, so extra attempts and extra hits are anti-correlated, not independent as my round-1 binomial assumed. That anti-correlation crushes the tail." — Claude Sonnet
Sonnet took its 4+ number from 17 percent to 6.5 on that argument alone, the largest single revision on the board. Gemini cut 11 to 6.5. GLM went the other way on the arithmetic and held its ground on the level, saying the binomial on its own stated assumptions produced 51 rather than the 55 it had guessed.
One seat declined to move on the piece of the argument that most of the others adopted. Several models had assumed a final, goal-deciding attempt would be staged at a softer golf course, and shaded their conversion rate up for it. Sonnet refused, and said why: the series already targets courses with records between 62 and 65, and there is no sign of any late-season softening in that range. The record at Tenison Park was a reachable 65, and he missed it anyway.
The Rung Where The Two Views Split
Set the 3+ argument aside, because it is a coin flip on both readings, and look at the shape of the rest of the board instead.
The market's tail is close to flat. Four-or-more trades at 5 cents, five-or-more at 2, six-or-more at 2. A ladder that barely decays like that describes a season where, if production keeps going, it keeps going for a while, and each additional record is only modestly harder to reach than the last.
The panel's tail collapses: 6.5 percent, then 0.8, then 0.1. That shape comes directly from the stopping rule the models talked themselves into. One bonus episode after a miss is easy to believe. Four more episodes, converted at a rate he has never sustained, in a season that told its audience it had six chances, is not.
The volume figures make this more than an academic disagreement. The 6+ contract is the most heavily traded rung on the entire board, 68,117 contracts, more than the 3+ rung that carries the actual story. Nearly 20,000 contracts of open interest sit on the 5+ rung. Money is parked on outcomes that would require the season to more than double its announced length and for DeChambeau to convert nearly everything he has left.
The honest caveat belongs right here: our panel is a worse forecaster than the market price. Across a large sample of disagreements of ten cents or more, the models have come out ahead only about a third of the time. Where the panel and the exchange disagree on this board, history says to expect the exchange to be right. What the panel is good for is showing you the assumption each side is making, and here the two sides are making genuinely different assumptions about whether a stated plan is a plan.
What Would Change The Panel's Mind
Every model was asked to name the one observable that would move it most. All eight named a version of the same thing, and none of them named golf: the next upload, when it lands, and whether it is framed as the season finale.
- A Sixth Course-Record Episode Posts. Its result settles the 3+ rung outright, one way or the other, and the market can close early the moment a third record lands.
- The Sixth Episode Misses And A Seventh Appears Anyway. This is the single event that would move the 4+ rung materially. Gemini set a specific tripwire: if episode six has not posted by mid-October, the calendar squeezes out any chance of a seventh and the tail should be marked down further.
- A Course-Record Episode Posts And The Course Record Is Soft. Several seats shaded conversion up for a goal-deciding round on an easier property. If a sixth episode targets a 66 or 67 at a short municipal track, the panel's 35 percent per-attempt figure reads low.
- The Season Ends Without A Sixth Attempt. Since Silvertip the channel has posted a travel-format film at Banff Springs on August 12 and a Break 50 episode on August 25. Neither is a course-record attempt. A quiet pivot away from the format entirely would settle everything above 2+ at zero.
Settlement timeline
| Date | What happens |
|---|---|
| Dec 31, 2026, 10:00 A.m. ET | Contract expiration. Rounds must be uploaded by then to count |
| Any Time Before That | A rung settles early the moment its threshold is hit |
| Mid-Sept 2026 | Next episode due on the season's observed cadence, which has run 35 to 75 days between course-record uploads |
| Mid-Oct 2026 | Gemini's stated cutoff for a seventh attempt to remain physically possible |
This board is re-scored whenever the story moves, and the story here moves on upload dates. Prices and panel numbers on this page are as of August 25, 2026.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $194M traded
- 2027 Pro Football Champion · $72M traded
- 2028 U.S. Presidential Election winner? · $62M traded
- Pro Baseball Champion · $60M traded
- 2028 Republican presidential nominee · $60M traded
- Bitcoin price at the end of 2026 · $33M 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.
- MLB Worst Record Odds: The Race Nobody Wants Is On Kalshi
- NFL Worst Record Odds 2026: The AI Panel Vs. Kalshi
- How High Will The S&P 500 Go In 2026? Odds And Verdict
- NFL Best Record Odds: The 13-4 Team Kalshi Forgot
The Bottom Line
DeChambeau is two for five with one announced attempt left, and the contract that asks whether he gets to three is a genuine coin flip on the panel's read and a wide, thinly quoted market on Kalshi's. That much both sides broadly agree on.
The disagreement is further out on the board, and it is not really about golf. It is about whether "I'll have six chances" was a plan or a piece of narration. Price the ladder as if extra episodes arrive on their own schedule and the tail stays fat. Price it as if the seventh episode only exists because the sixth one failed, which is what the panel concluded, and the tail falls off a cliff. The season's next upload is the thing that decides which of those two seasons we have been watching.
Hero illustration: OddsShopper, in the house collage style. Bryson DeChambeau photo by The White House, licensed Public domain; photos cropped, toned, and composited.
