🐘 Stock Market Elephant/The Graveyard Series

Archived: written before this publication started in November 2026. Some claims and numbers are older than the posts under Writing.

The graveyard series · part 5 of 6

What position sizing can and cannot do

Aug 25, 2026 · 6 min read

I Monte-Carlo'd a 50K prop eval (profit target vs trailing drawdown) thousands of times at every contract size, feeding it my strategy's real daily P&L. The result surprised me into a two-phase doctrine.

Phase 1: during the eval, sizing is a treadmill

Monte Carlo pass rate for a $50K eval at 1, 2, 3, and 5 micros — the bars barely move. Eval geometry scales with your size: bigger contracts reach the target faster and hit the trailing limit faster, in almost exact proportion.

Size	Pass rate	Median days to pass
1 micro	43.9%	123
2 micros	42.1%	44
5 micros	46.5%	11

Pass probability is flat at every size. An eval is a race between a profit target and a trailing drawdown — multiply your bets and both scale together, so the odds never move. Sizing buys speed, not probability. (Speed has value — evals bill monthly — but know what you're buying.) I also tested clever dynamic sizing against the eval. Same wall: the geometry is scale-invariant. Only a genuinely better strategy improves eval odds. Nothing else does. Anyone selling you an "eval passing" sizing trick is selling you the treadmill.

Phase 2: once funded, sizing becomes real

The funded-phase scheme that passed my tests on 1-minute data: volatility-scaled 1–5 micros (risk-matched per trade) versus flat 1 micro. More profit, better Sharpe — the drawdown grows slower than the P&L. This is the sizing that cleared the 95% deflated-Sharpe bar.

A funded account with a buffer has no target racing you. There, volatility-scaled sizing — risking a constant number of dollars instead of a constant number of contracts, so quiet days get more size and wild days get less — raised my risk-adjusted return by ~25% and nearly doubled the return-to-drawdown ratio. Same trades. Same signals. Only the when-big-when-small changed. And it passed the full overfitting gauntlet (PBO 0%).

Sizing is a wealth tool, not an eval tool. Flat and small while a trailing drawdown hunts you; volatility-scaled once you have cushion; and always decided by a written rule, never by how yesterday felt.

My rule is literally written down with a formula and a "recomputed Fridays only" clause — because the one guaranteed way to ruin good sizing is resizing after a loss.

Next: Part 6 — The plateau test, and the only clean data left

Educational content only — not financial advice; no performance promised or implied. All results are backtests on real CME MNQ data (2019–2026, ~1,900 sessions, realistic costs) and backtests differ from live trading. Futures involve substantial risk of loss.