Martingale Cap

Martingale caps are a band-aid on a structural risk problem. The underlying martingale architecture has unbounded tail risk by design; caps reduce the visible blow-up but don't eliminate the fundamental issue.

How martingale architecture works:

• Initial position opened at standard size • If position closes at loss, next position opens at 2× previous size, expecting price reversion to recover both original loss and new position cost • If second position also loses, third position opens at 4× original size; etc. • In theory, eventual price reversion produces marginal recovery profit; in practice, sustained adverse moves exhaust capital before reversion

What martingale caps add:

• Hard limit on doubling steps (typically 4-7) • After cap is reached, the EA stops opening new positions and accepts the cumulative loss • The maximum loss is bounded at the cumulative position size at cap-step, which is much smaller than uncapped exposure

Why martingale caps don't fix the structural problem:

• **Cumulative loss at cap is large**: even with cap at 5 steps, the cumulative position size is 2^5 = 32× original size. A 32× exposure that hits its stop produces 32× the original risk-per-trade loss • **High win-rate, occasional catastrophic loss**: capped martingale strategies look profitable most of the time (95% of cycles complete with small recovery profits) but the 5% of cycles that hit cap produce losses 10-30× larger than typical winners. Aggregate expectancy can still be negative • **Streak risk amplification**: market periods with sustained directional moves (the conditions martingale strategies hate) hit cap quickly. During regime stress, capped martingale strategies experience their cap-loss multiple times in rapid succession • **Survivorship bias in marketing**: capped martingale EAs that have completed 6-12 months without hitting cap look attractive in backtests and live verification; the 5% probability event hasn't yet occurred for those products. Probability accumulates over deployment time

Editorial position: even with martingale caps, martingale-architecture EAs are excluded from safety-tier and most-profitable-tier consideration. The bounded worst case is still large enough to produce account-blow-up during regime stress, and the structural strategy doesn't generate genuine edge — recovery profits depend on continued price reversion that isn't guaranteed.

For EA buyer evaluation: any EA marketed with "martingale recovery" framing should be assumed to have unbounded tail risk regardless of vendor's claims about caps and risk management. The cap mechanism is documented mathematically; the risk it creates is also documented mathematically. Buyers who deploy these products should size deployment capital as if total loss is the realistic worst case, not the cap-bounded worst case.