2026-07-19 · Validated Research Team
Do Polymarket BTC 15-minute strategies work? Five tests, no validated edge yet
The short answer is not yet. Across five published Polymarket tests, none cleared the full validation framework. One continuation rule produced a promising holdout, two ideas were rejected, and two remained inconclusive because executable-fill or sample evidence was missing.
What we tested
| Strategy | Trades | Profit factor | Win rate | Verdict |
|---|---|---|---|---|
| BTC 15-minute favorite continuation | 182 | 1.212 | 73.6% | Inconclusive |
| BTC strike-filtered favorite | 8 | — | 100% | Inconclusive |
| BTC spot lead/lag | 98 | 0.974 | 68.4% | Rejected |
| Up/Down complement discount | 1,716 observations | — | 100% apparent | Inconclusive |
| BTC-to-altcoin lead/lag | 35 | 1.129 | 57.1% | Rejected |
The strongest candidate
Favorite continuation was the only rule to clear the headline profit-factor threshold. Its combined PF was 1.212 after a one-cent adverse haircut on entry and exit, and its 45-trade chronological holdout PF was 1.56. That is encouraging, but the 95% bootstrap interval for mean return still crossed zero and two-cent-per-side cost stress reduced PF to 0.752.
The right conclusion is not “73.6% wins, therefore it works.” The conclusion is that continuation deserves a frozen-rule forward test with executable order-book fills.
Why the apparent arbitrage is not yet proof
The complement-discount screen found many moments where Up and Down appeared purchasable for less than $1 combined. Historical observations cannot prove both legs were simultaneously fillable at displayed size. Until synchronized bid/ask depth confirms that, a 100% apparent win rate is a market-data observation, not executable arbitrage.
Bottom line
There is no validated Polymarket BTC strategy in this batch. The best next experiment is forward paper capture of the favorite-continuation rule with frozen thresholds, actual bid/ask fills, fees and available size. See the complete Polymarket board.
By Validated Research Team (Methodology v1.0 — 11-gate validation). Part of 11-gate validation. Backtests are not investment advice.