Why win rate lies about profit
Almost every bot is sold on its win rate — "wins 7 out of 10 trades." But win rate is only half the equation, and on its own it is close to meaningless. What matters is expectancy: how much the average winner makes versus how much the average loser costs, times how often each happens, minus the fees and slippage paid on every single trade. A bot can win 65% of the time and still bleed if its winners are small, its losers are large, or — most commonly — if it trades so often that fees quietly consume the entire edge.
That last point is the one most bot operators miss. Fees scale with the number of trades, not with profit. A round-trip taker cost of 0.11% plus a little slippage is trivial on one trade. But a bot doing 20 trades a day pays it 600 times a month; at 50 trades a day, 1,500 times. A statistical edge of a few hundredths of a percent per trade — which is realistic for high-frequency scalping — simply does not survive that drag. This is the structural reason most retail trading bots are net-negative: the strategy isn't necessarily wrong, the math of frequency × fees is.
So before trusting any bot, model it honestly. Size each trade with the position size calculator, see what a single futures trade really nets after fees on the real futures profit calculator, compare what your exchange actually charges on the exchange fees comparison, and check the long-run blow-up odds with the risk of ruin calculator.
How to use this calculator
Enter the position size your bot opens per trade (the notional, not your account), how many trades it takes per day, and its win rate. Then the two numbers that decide everything: the round-trip fee + slippage (taker fees on most exchanges are ~0.055% per side, so ~0.11% round trip — add slippage and round to about 0.12–0.20% for a realistic scalper), and the average win and average loss as a percent of the position. The big number is your honest net profit or loss per month after fees. The break-even win rate metric tells you the minimum win rate this strategy needs just to stay flat.
Common mistakes
Quoting win rate without expectancy. 60% means nothing if the losers are bigger than the winners. Ignoring fees in backtests. A zero-fee backtest is a fantasy; fees flip many "profitable" bots to losers. Forgetting slippage and funding. Real fills are worse than backtest fills, and perpetual funding adds cost on held positions. Confusing notional with account. A bot trading $1,000 positions on a $200 account is using leverage — size and risk accordingly.
FAQ
My bot was profitable in backtest — why is it losing live? Almost always fees and slippage. Backtests often use mid-price fills and ignore taker fees; live trading pays the spread, the fee on both sides, and worse fills in fast markets. Re-run the numbers here with a realistic round-trip cost and the edge often disappears.
What round-trip cost should I use? For a taker-based scalping bot on a major exchange, ~0.11% in fees plus 0.02–0.10% slippage, so 0.12–0.20% is realistic. Maker-only strategies can be far lower, but only if your orders actually fill without chasing price.
Does a higher win rate always mean more profit? No. Raising the win rate by taking profit early (small winners) while letting losers run can lower expectancy even as the win rate climbs. Profit is winners × win-size minus losers × loss-size minus fees — all four levers matter.