R20Published

On Polymarket, Are Player-Stat Sports Markets Disputed More Often Than Moneylines?

Using a paper’s public data on resolved markets, we compared how often player-stat and moneyline sports markets had their settlement challenged.

Result status
Results
Last updated
Start date
Duration
Day 3 of 3
Plan registered
Illustration of a grid of small squares standing for markets, with just one yellow square marked by a small flag, showing that only a few markets have their settlement challenged
Image:Illustration (made by Mysports.AI)

Settlement Disputes in Polymarket Sports Markets: Key Findings

Finding so far
Across all sports, player-stat markets’ dispute rate as a multiple of moneylines’:1.00×

Sports markets in the authors’ analysis sample scheduled to end before April 2026 with an oracle record: 40 of 14,340 player-stat markets and 126 of 45,345 moneylines were disputed. Range 0.55–1.59: no difference can be seen (which does not mean they are the same).

  • Disputes are rare in sports markets: 297 of 142,644 (0.21%), about 1 in 500.
  • Dispute rates by type: moneyline 0.28%, player-stat 0.28%, spread 0.11%, totals 0.12%, futures and other 0.35% (descriptive, not tested).
  • By sport (in this sample): baseball 0.96% and cricket 0.73% are highest; basketball, hockey and American football about 0.14% (descriptive, not tested; a challenge does not mean the result was wrong).
  • Alternative calculation, basketball only: player-stat 0.27% vs moneyline 0.11%, about 2.4 times, but the range of 1.07–5.73 is wide and only 13 basketball moneylines were disputed; by our pre-set rule, alternative calculations do not change the main conclusion.
  • Of the 297 disputed sports markets, 18 (6.1%) ended void or settled 50-50; the rest ended with a winner. This is how they settled, not why they were disputed, and we did not check whether results changed after the challenge.

How to use this yourself

  1. Start with the official explanation: Polymarket’s help page “How are markets disputed?” says “Once a market is proposed for resolution it goes into a challenge period of 2 hours”—once a result is proposed, there are 2 hours to challenge it, and if no one does, it becomes final. If it is disputed, it goes to UMA for 24–48 hours of discussion and then a vote of about 48 hours, so a challenged market takes several more days to become final. Open source ↗

  2. Open the market page and find the “Rules” section: it says how the market is decided and what happens on postponement or cancellation, and the help page says the dispute link is there too. Two lines common in past sports markets: “If the game is postponed, this market will remain open until the game has been completed” means a postponed game keeps the market open until it is played; “If the game is canceled entirely, with no make-up game, this market will resolve 50-50” means a canceled game with no make-up voids the market at half price per share. Some markets say a canceled game resolves “No” instead—each market’s own rules decide.

  3. For player-stat markets, look for one more line. A common one in past markets: “If the player is listed as inactive or otherwise does not take the court at any point, the market will resolve "No"”—if the player is listed as inactive or never gets on the court, the market resolves “No”; it is not voided. Some markets say “Under” instead. Find this line to know what happens if the player sits out.

  4. Apply this page’s numbers: only about 1 in 500 sports markets is challenged; across all sports, player-stat and moneyline markets are both about 0.28% with no visible difference (ratio 1.00, range 0.55–1.59), so don’t treat “is it a player-stat market” as a signal for settlement disputes. In this sample, baseball was about 0.96% and cricket about 0.73% (descriptive only, not a formal test).

  5. A challenged market is not the same as a voided one: of the 297 challenged sports markets here, 18 (6.1%) ended void or settled 50-50 and the rest ended with a winner (we did not check whether results changed after the challenge).

These are past averages for Polymarket sports markets in the authors’ analysis sample, scheduled to end before April 2026 with an oracle record, and guarantee nothing for any market. Don’t treat market type as a signal for disputes: within basketball, player-stat markets were about 2.4 times as likely as moneylines, but the range of 1.07–5.73 is too wide to be a conclusion. A challenge does not mean the settlement was wrong. The rule lines quoted above are common wording from past markets; each market’s own rules on its page decide. This is research, not betting advice.

In plain words

Why study this?

When a settlement is challenged, the result takes several more days to become final. Knowing which kinds of markets are challenged more tells you which ones deserve a careful read of the settlement rules.

What do we want to know?

After each Polymarket market ends, someone proposes the result, and others then have 2 hours to challenge it; a challenge sends it to UMA (the third-party system that decides Polymarket results) for discussion and a vote, and the result takes several more days to become final. Player-stat markets depend on a player’s official stats, and the player may not play at all—are they challenged more often than moneyline markets, which only depend on who wins?

How do we study it?

We used public data from Mahrous and Di Pietro (about 800,000 resolved markets), keeping the 142,644 sports markets scheduled to end before April 2026 that have an oracle record (so a challenge is visible). We sorted markets into five types from the question text, and timestamped the plan and the analysis code before looking at any results. We resampled whole games 4,000 times to get the range. This is a retrospective study, not a blind test.

What have we found?

Disputes are rare in sports markets: 297 of 142,644 (0.21%), about 1 in 500. Across all sports, 40 of 14,340 player-stat markets (0.28%) and 126 of 45,345 moneylines (0.28%) were disputed, so no difference can be seen: the ratio is 1.00, and by the 95% range the player-stat rate could be as low as 0.55 times or as high as 1.59 times the moneyline rate. Of the 297 disputed markets, 18 (6.1%) ended void or settled 50-50; the rest ended with a winner.

How does this help when reading markets?

Across all sports, player-stat markets are not seen to be challenged more often than moneylines (ratio 1.00, range 0.55–1.59), so don’t treat market type as a signal for settlement disputes. In this sample, baseball (about 0.96%) and cricket (about 0.73%) markets were challenged more often, but this is descriptive only, with no formal test, and a challenge does not mean the result was wrong. Whatever the market, what happens on postponement, cancellation or a player not playing is written in the rules section of the market page—reading it first helps most.

What can’t we say yet?

This is a retrospective study, not a blind test: the paper had already published the overall dispute rate for sports markets. The two groups differ in sport: about 89% of player-stat markets are basketball, while moneylines are mostly soccer (about 42%), so the main comparison mixes in differences between sports. Within basketball only, player-stat markets were disputed about 2.4 times as often as moneylines (range 1.07–5.73), but the range is wide and only 13 basketball moneylines were disputed; by our pre-set rule, alternative calculations like this do not change the main conclusion. The main result and the basketball-only result disagree, possibly because the two groups contain different sports; this is an interpretation made after seeing the results, not a pre-planned test, so we cannot say player-stat markets are more prone to disputes. Also, a challenge does not mean the settlement was wrong, and the data cannot tell why a market was challenged; only markets with an oracle record (about two-thirds) show disputes.

Terms used on this page

Disputed
After a result is proposed, someone posts a bond to challenge it (a dispute) within the 2-hour challenge period, and the question goes to UMA token holders for discussion and a vote. In this study, at least one challenge counts. A challenge does not mean the proposed result was wrong.
Oracle record
The record of proposals, challenges and settlement left by the UMA oracle, the third-party system that decides Polymarket outcomes. Only markets with this record show whether they were challenged.
Moneyline market
Who wins a game or match, including “will it end in a draw”, and the winner of a period, set or half.
Player-stat market
Whether a player’s single-game stat (points, rebounds, assists, receiving yards, etc.) goes over or under a line, plus first touchdown and whether a player scores.
Void or 50-50 settlement
The market has no winner and each share settles at half price ($0.50). Common when a game is canceled with no make-up, postponed past a deadline or not completed, and in some markets when it ends in a tie; the rule is in each market’s rules section.
Range
A 95% interval from resampling whole games 4,000 times. We only say there is a difference when the ratio’s range does not include 1.

Detailed results

Share of sports markets whose settlement was disputed, by market type

Sports markets scheduled to end before April 2026 with an oracle record. The main result compares player-stat with moneyline; other rows are descriptive (not tested). Numbers are the percentage disputed.

Player-stat (main result)0.28 %
Moneyline (main result)0.28 %
Spread0.11 %
Totals0.12 %
Futures and other0.35 %

Main result: player-stat markets are 1.00 times moneylines, range 0.55–1.59; no difference can be seen (which does not mean they are the same). Disputes are few, so other rows are descriptive only. This is research, not betting advice.

HypothesesFor deeper reading
  • H1Player-stat and moneyline markets are disputed at different rates → no difference can be seen: 1.00 times (range 0.55–1.59). This does not mean they are the sameInsufficient evidence
  • D1Of the 297 disputed sports markets, 18 (6.1%) ended void or settled 50-50 (descriptive, not tested)Exploratory

Settlement Disputes in Polymarket Sports Markets: Research Methods

Research questions

  • How many times as often are player-stat markets (e.g. a player’s points over a line) disputed as moneyline markets?
  • What are the dispute rates for each of five market types and for each sport? (descriptive, not tested)
  • Of the disputed sports markets, how many ended void or settled 50-50? (descriptive, not tested)

Method summary

  • There is one main result: the player-stat dispute rate divided by the moneyline rate, two-sided with no assumed direction. We judge it by whether the 95% interval includes 1, which is an approximate rule.
  • Sports markets = the authors’ “Sports” topic family plus the topic the authors labeled “NBA Player Prop O/U Markets” but placed under “Other”; without it nearly all NBA player-stat markets would be excluded. This was decided before looking at results.
  • Market types come from the question text only, and the dispute field was read only after the timestamp. A 300-question manual check found no errors, and a further 150 (60 player-stat, 60 moneyline and 30 from the other three types) were also all correct; the checker also wrote the rules, so this may be optimistic.
  • Six alternative calculations were listed in advance (for example basketball only, only markets settled from official sports sources, and ending a month earlier); by our pre-set rule they are reported but do not change the main conclusion.

Data and limitations

Limitations

  • Retrospective, not blind: the paper had already published overall sports dispute rates and two template rates.
  • About 89% of player-stat markets are basketball while moneylines are mostly soccer, so the main comparison mixes in differences between sports.
  • Only markets with an oracle record (about two-thirds) show disputes; results do not represent all Polymarket sports markets.
  • A challenge does not mean the settlement was wrong, and the data cannot tell why a market was challenged.
  • The data have no actual settlement time, so we used the scheduled end plus a one-month buffer; ending a month earlier gives a similar result.
  • The manual check of market types was done by the person who wrote the rules; there was no independent second annotator.
  • No forward check: there is no comparable new public data to re-test on.

Latest progress

Latest:Main result is in: across all sports, no difference can be seen between player-stat and moneyline dispute rates (both about 0.28%, range 0.55–1.59); the plan and analysis code were timestamped first, and results were reviewed by an AI code-review tool

Next step:Study completed; there is no comparable new public data, so no forward test; we will reassess if the dataset is updated

Study timeline

  1. Confirmed the paper’s data is public under CC BY 4.0 with a dispute field; since the paper already covered sports overall, we switched to comparing types within sports
  2. Data gate passed (market counts only); plan finalized and timestamped after one round of AI-tool code review; analysis code tested on synthetic data and timestamped before running
  3. Main results in and frozen; one round of AI-tool code review (numbers unchanged; wording and scope fixed)

Update log

Version history

  1. v1.0-results

    Main result: no difference seen between player-stat and moneyline dispute rates (1.00 times, range 0.55–1.59)

  2. v1.0

    Plan finalized and timestamped (T0 = 2026-11-16 18:38 UTC); retrospective, not blind

Study record

Plan fingerprint (SHA-256)
Plain-language plan632405d80068f440f76dda088c7eae290ff6d7ba8c95bb49115dff3428d958fc
Technical appendix5a6402b8b91d927537593742dec3bcb7d8d94ae1e5d7dd8bdc8df9c509bb7074
Data gate and classification check8988a8ee342bbb3e26ab8c0bf13e6033608f836264e51b0190dda8cfd9a6d37d
Analysis code timestamp (before running)6789ae077bff912e212679739f04ec8ce85dc84478ddab7189a959a2244b604e
Timestamped into the public Bitcoin blockchain record with OpenTimestamps on Nov 17, 2026
Plan changes:14 (each recorded in the deviation log, DEVIATIONS.md; not public for now)
Code and data:Not public for now (contains third-party data sources and internal information)

This page is a research log for reference only and is not betting advice or a guarantee of profit.