Page 2 of 2

Re: Strategy Architecture: Bi-Directional Liquidity Sweep & House-Money Scaling

Posted: Fri Nov 13, 2026 4:16 pm
by Shadow Trader
Interesting framework, and the anti-martingale side is at least the safer direction to scale in. Adding size after wins with house money is a lot more forgiving than the classic martingale that adds after losses.

A couple of things worth stress-testing before any of this goes live. The first is the series reset. Exponential scaling inside a winning series means the largest position is always the one taken after the longest run of wins, and that is also the trade most likely to give the whole series back. It helps to model what happens if the last trade of every series is a full loss, because in practice that's often the realistic case, and see whether the system is still positive.

The second is the simultaneous buy and sell readiness on XAUUSD next to the majors. Gold can move several times its usual range in seconds around data, and two-sided pending orders can both trigger in a whipsaw, leaving you with two losers rather than one winner. A filter that cancels the opposite order once one side fills, plus a blackout around high-impact releases, would remove a lot of that risk.

Lastly, with ECN pricing it's worth including commission per side in the backtest from the start. Strategies with many small trades can look very different once the round-turn cost is applied to every entry, especially across a scaling series where each step is a separate fill.