Here is an experiment guaranteed to insult every trader's favourite subject. We removed the entry decision entirely — every position starts on a literal coin flip, long or short — and let only the exit policy differ. If entries are where the magic lives, all three coin-flip traders should end up in the same place.

Setup: a synthetic market built to behave like a real one (long stretches of noise, occasional persistent trends), 600 simulated accounts of 300 trades each, 1% risk per trade, realistic costs on every trade. Three exit policies: no stop (hold for a fixed time, hope included free); fixed 1:1 (stop at −1R, target at +1R); and cut-and-trail (stop at −1R, winners trailed so trends can pay 2R, 3R, 5R).

Coin-flip entries, three exit policies — median of 600 simulated accounts
Cut losses + trail winnersFixed 1:1 stop/targetNo stop (time exit)
$10k$20k$30k$40k$50k075150225300trades (median of 600 simulated accounts, costs included)Random entry, cut losses + trail winnersRandom entry, fixed 1:1 stop/targetRandom entry, no stop (time exit)trail exits → $52k1:1 exits → $9kno stop → $8k

Synthetic trending market, 1% risk per trade, costs included. Ordering is the finding; dollar values are simulation artifacts.

The result nobody wants to hear

Same coin. Three different traders. The no-stop policy bled to about $7 503 — random entries plus unbounded losses plus costs is a slow-motion accident. The symmetric 1:1 policy ended near $8 782: essentially breakeven minus costs, exactly as theory predicts for a coin flip. And the cut-and-trail policy — on the same random entries — grew to about $51 664, because on a market that sometimes trends, capping every loss at 1R while letting occasional winners run to many R produces positive expectancy with a win rate barely above 30%.

Before you quit entries forever: this is a synthetic market, deliberately generous with trends, and no simulation slips or gaps against you. The dollar figures are not a promise. The ordering is the finding — and it reproduces the famous random-entry experiments run by Van Tharp and Tom Basso decades ago on real data.

What this means for real trading

The asymmetry of exits is the load-bearing wall of profitability. Entries decide how often you're right; exits decide how much right and wrong cost — and the second lever is simply bigger. A mediocre entry with disciplined exits survives; a brilliant entry with undisciplined exits doesn't.

It also reframes what a good entry is even for. If coin-flip entries plus asymmetric exits roughly work, then a real edge in entries — structure that actually holds, confirmation that real size is defending a level — is pure upside stacked on a sound machine. That is the correct order of construction: risk framework first, exit asymmetry second, entry selection third. Most traders build in exactly the reverse order, and the first two floors are missing.

What the lab says

  • Exits carry more expectancy than entries. The same random entries produced ruin, breakeven, and strong profit — the exit policy was the only difference.
  • The stop is not a defense; it's the engine. Capping losses at 1R is what makes occasional multi-R winners mathematically decisive.
  • Win rate is a by-product. The profitable random trader won only ~a third of trades. Comfort with frequent small losses is a skill, not a flaw.
  • Entries are the third floor, not the foundation. Build sizing and exits first; then better entries multiply an already-working machine.
About these numbers. The charts on this page come from Monte-Carlo simulations (thousands of simulated accounts with fixed rules), not from real trading records. Simulations simplify reality — no slippage spikes, no psychology, no changing markets — so treat the comparisons as the finding, not the absolute dollar amounts.
Not financial advice. Everything on this page is educational — history, simulations, and reasoning, not recommendations. It is not a signal service and not investment advice. Trading futures and options carries a substantial risk of loss. Never risk money you cannot afford to lose.