PriceWireA 60% win rate can still lose money
Accuracy leaves out the size of your losses and the cost of being right. Work through the arithmetic yourself.
PriceWire Research · 2026-09-22 · Prepared with AI assistance; calculations checked programmatically.
A claim that a strategy wins six times out of ten sounds incomplete to us. How much does it win? How much does it lose? What does each attempt cost?
The example below is synthetic. It is an arithmetic demonstration, not a backtest, a broker fee estimate or a claim about Market Lab.
Ten experiments, one uncomfortable result
| Component | Calculation | Result |
|---|---|---|
| Six wins | 6 × 100 | +600 |
| Four losses | 4 × 150 | −600 |
| Ten round trips | 10 × 10 | −100 |
| Combined | 600 − 600 − 100 | −100 |
The observed win rate is 60%. The total result is negative. Before costs, the wins and losses cancel; costs make the difference.
Four inputs belong together
For a repeated, fixed-size experiment, the average result is the win probability multiplied by the average gain, minus the loss probability multiplied by the average loss, minus round-trip cost.
Expected result = p × gain − (1 − p) × loss − cost
With these inputs, breaking even requires a 64% win rate: (150 + 10) / (100 + 150). This assumes the stated gain, loss and cost remain representative. It does not establish that a 64% hit rate is achievable.
Change the assumptions in the free calculator.
What this calculation cannot tell you
It does not estimate drawdowns, rare losses, compounding, dependence between trades or the cost of scaling up. A profitable-looking average can coexist with an unacceptable path of losses. Position sizing and portfolio risk require separate analysis.
When evaluating a research claim, ask for all four inputs and how each was measured. Accuracy is one input to the calculation.