Set the portfolio risk budget before you size

A late start could still exhaust a small account

A one-contract moving-average campaign across six markets from 1981 to 1991 produced about $197,204 of net profit. A trader who started after 1981 still recorded a two-year loss of $5,027 after 120 signals.

An account started with less than $5,000 at the beginning of 1982 would have been wiped out within two years, even though the same rules later showed $164,343 of net profit from 1982 through 1991. Variable-shift-testing re-runs the same rules after sliding the test window so start-date luck is not mistaken for robustness.

Stacked ledgers hid the joint drawdown

Simultaneous six-market testing produced a maximum drawdown of $29,619, larger than the worst single-market drawdown. A $25,000 guess based on one market understated the capital the joint book actually needed.

Adding closed-market results after the fact hides concurrent profit and loss. A $5,000 yen gain against a simultaneous $7,000 cotton loss still cuts the shared account by $2,000. Portfolio-level-simulation is a day-by-day test that applies every market signal against one shared equity curve instead of adding separate market results after the fact.

A drawdown-limit is a maximum allowable decline from peak equity used to judge whether the account is large enough to survive the plan. In the archive, the joint $29,619 figure is the scale that a one-market guess missed.

Contract count follows the shared equity curve

A workable overlay used a $25,000 start, no more than 2% of equity risked on any one trade, and a 35% ceiling on total capital at risk across the book.

Volatility-position-sizing scales contract count from current equity, market volatility, and the distance to the stop so a single trade cannot spend more than a preset fraction of capital. Dynamic sizing from the shared equity curve can raise a later cotton signal from one contract to two when a prior yen profit lifts the 2% risk budget above that market's stop distance.

A portfolio-risk-budget is a hard ceiling on how much of current account equity may be at risk across all open positions at once. The archive overlay set that ceiling at 35% of capital.

What the two ceilings changed

With the 2% and 35% overlays, profitability rose 45%, days in the market fell from 2,632 to 1,974, and the longest drawdown shrank 56%, while peak percentage drawdown stayed about the same.

Even a 40% win rate still leaves roughly a 1-in-2,100 chance of losing more than half the account in the first 15 trades if more than 5% is risked per trade.