Today's cushion should be able to take 10 typical losses. In 20,000 simulations per case, that rule raised one trader's pass rate from 64% to 91% without touching the strategy. The price: it took more than twice as long.
What matters is the cushion, not the account size
A 50K account doesn't let you risk $50,000. It lets you risk the distance between your balance and the drawdown floor and, if the firm has a daily loss limit, what's left of it today. The smaller of the two is today's cushion, and it changes every day: it shrinks when you lose, and with a trailing drawdown, when you win the floor moves up behind you and the cushion doesn't grow (see trailing drawdown explained).
The rule we propose is simple:
today's contracts = today's cushion ÷ (10 × your typical loss per contract), never above your normal size.
If with 4 micros your typical loss is $200 ($50 per contract) and you have a $2,000 cushion, you trade 4. If a bad streak leaves your cushion at $1,000, you trade 2. At $500, 1. Below that, none: no trading that account today.
Four ways to choose your size, compared
The same fictional trader (45% win rate, average win $300, average loss $200, two trades a day), in a 50K account with a $2,000 end-of-day trailing drawdown and no time limit:
And with two other fictional traders (pass · blown · sessions to pass):
| How size is chosen | Same trader | High win-rate scalper | One big trade a day |
|---|---|---|---|
| Fixed, twice normal | 49% · 52% · 10 | 86% · 14% · 15 | 54% · 46% · 6 |
| Size up when winning | 57% · 43% · 24 | 97% · 3% · 26 | 60% · 40% · 13 |
| Fixed, normal size | 64% · 36% · 32 | 98% · 2% · 33 | 66% · 34% · 17 |
| 10 typical losses | 91% · 9% · 74 | 100% · 0% · 34 | 99% · 1% · 88 |
Three takeaways
- Matching size to the cushion is the decision that changes the result most. The first trader goes from blowing 36% of accounts to blowing 9%, with the same strategy.
- The price is time. It takes 74 sessions instead of 32. With no time limit, that's well worth it. With a 30-day limit there's no time for it, and the decision is different.
- "Size up when you're winning" makes things worse than fixed size. With a trailing drawdown, winning doesn't give you more cushion, so more size is more risk on the same cushion.
The warning almost nobody gives
We repeated the test with a trader with no edge (40% win rate, wins $300 and loses $200: in the long run, neither wins nor loses). With fixed size they passed 27%; with the 10-typical-losses rule, 26%, taking 81 sessions instead of 35. Position size doesn't create an edge: it protects the one you have. Before fine-tuning size, check that your strategy makes money after costs.
Calculator: your size and your plan for today
Besides the size, it gives you the two numbers of the daily plan: stop for the day (the smaller of half of today's cushion and a quarter of the distance to the floor) and max risk per trade (a third of that, so the daily stop comes no earlier than the third loss). It runs in your browser; nothing is sent.
What to do with this
- Work out your typical loss: the average of your recent losing trades, at your normal size.
- Every morning, before the first trade, write down today's cushion for each account and the size it allows.
- Don't size up when winning in an account with a trailing drawdown.
- If your account has a time limit, first check whether you have time: a prudent size passes more often, but more slowly.
How it was calculated
Our own simulation with a fixed seed: 20,000 evaluations per case, independent trades, a 50K account with a $2,000 end-of-day trailing drawdown that stops rising at the starting balance plus $100, a +$3,000 target, no time limit or consistency rule. "Size up when winning" adds half a size for every $1,000 of profit, up to double. The 10-losses rule changes size in steps of 1 micro (a quarter of normal size) and never goes below 1 micro in the simulation. Fictional traders; these are nobody's results.
Simulated and hypothetical results, with the limitations that implies. Not investment advice. Firm rules change: check them on each firm's website. This article contains no affiliate links.