How Quant Arena calculates it
For every EA backtest we divide the net return over the test (one year of hourly data, as a share of the $10,000 starting balance) by the maximum drawdown, the largest peak-to-trough fall of the account in percent. MetaTrader's strategy tester uses dollar amounts instead; the two give similar numbers when the drawdown happens close to the starting balance. A losing EA has a negative recovery factor.
What our EA backtests show
Across the 1,123 reliable EA backtests on four markets (hourly data from May 2025 to May 2026), the median recovery factor was about zero: the typical EA did not earn back its worst drawdown even once. 240 results (21%) reached 1 or more, and only 55 (5%) reached 3.
Even among the 562 profitable results, 322 had a recovery factor below 1. They ended in profit, but made less than they had been down at their worst point — a year that a real trader would have had to sit through.
How to read it
Read it together with the length of the test. Over one year, a recovery factor of 2–3 is solid; over ten years the same number is weak, because there was far more time to earn it.
It rests on a single number — the worst drawdown — so a strategy that has not yet met its bad market can look better than it is. The Calmar ratio uses annual return instead of total profit and is easier to compare across periods.