Monte Carlo simulation

Part of Simulation outputs in the PropSurvival glossary — defined the way the engine actually measures it.

One edge, many pathsMonte Carlo simulation

Running thousands of alternate versions of your trading — each a random sequence drawn from your statistics — against the challenge rules, and reading pass/failure frequencies as probabilities. It answers "what does my edge imply under these rules," not "what will happen."N paths drawn from your own statistics

The rule corpus behind this glossary — rules last read 2026-07-16, latest auto-recheck 2026-10-15 · verification log

Run controls2 terms

Engine control · derived by defaultSeed

The single number that fixes the pseudo-random draw, so the same seed reproduces the same run — on a given engine version. By default it is derived from the run’s own numbers (your four dials plus the account, its limits and the path count), so identical inputs always return the identical answer and the reading never flickers between two values. Advanced users can pin an explicit seed to cite one exact run, or shuffle it to watch how far the survival reading moves on a different roll of the same odds — a direct read on sampling noise. If the engine's draw sequence changes, the same seed can move, so reproducibility is always stated on this engine version rather than absolutely.same seed reproduces the same run · on this engine version · measured: re-seed and watch the spread

Engine control · 2,000 by defaultSimulation depth

How many alternate futures a run draws — 2,000 by default, or 5,000 and 25,000 when you want a steadier reading — all three free, computed on your device. More paths do not change the strategy or hand back a better answer; they narrow the error bar around the same one. Monte Carlo standard error shrinks roughly with the square root of the path count, so 2,000 is enough for a first look, 5,000 is steadier, and 25,000 steadies the last digit — and higher depth is what shrinks the seed-to-seed movement above.standard error falls with 1/sqrt(N) · 2,000 / 5,000 / 25,000 paths · measured: how many paths a number needs