Prop-challenge study self-test 2026-10-14T18:04:24.756Z
[PASS] theory: target 3000, drawdown 2000: simulated pass = D / (T + D) within 3 standard errors: 40.31% vs 40.00%
[PASS] theory: target 10000, drawdown 10000: simulated pass = D / (T + D) within 3 standard errors: 50.21% vs 50.00%
[PASS] theory: target 5000, drawdown 6000: simulated pass = D / (T + D) within 3 standard errors: 54.29% vs 54.55%
[PASS] rule off: the published coin-flip matrix (12 cells) is reproduced exactly: 12 of 12 cells identical
[PASS] rule off: a published daily-limit cell is reproduced exactly: 0.208
[PASS] one big day: the target is reached on day 1 but the pass waits until the best day is at most 40% of the total (day 16): {"res":"pass","day":16,"delayed":true}
[PASS] even days: the pass comes on the day the target is reached: {"res":"pass","day":4,"delayed":false}
[PASS] a delayed pass can still fail on a loss limit: {"res":"breach","day":3,"reason":"drawdown","delayed":true}
[PASS] rule off: the same sequence passes on day 1: {"res":"pass","day":1,"delayed":false}
[PASS] independent re-implementation, U1 no edge, $250: engine 26.98% vs reference 26.66%
[PASS] independent re-implementation, U2 negative edge, $500: engine 17.18% vs reference 17.64%
[PASS] independent re-implementation, F1 phase 1 good edge, 0.5%: engine 98.27% vs reference 98.16%
[PASS] independent re-implementation, F1 phase 1 no edge, 2%, 30 days: engine 23.41% vs reference 22.93%
[PASS] independent re-implementation, F2 small edge, 1%: engine 13.14% vs reference 13.25%
[PASS] independent re-implementation, F2-C no edge, 2%: engine 6.80% vs reference 6.69%
[PASS] independent re-implementation, U1-C good edge, $125: engine 87.23% vs reference 86.83%

PASSED