I have created ema crossover need help. Then I backtest on open price … It gave 90% accuracy and then I selected close price… Then it gave me 30% … Why is this difference. I have created on 13 ema crossover on candle zero (close) > ema 10 candle zero (close)
Tradetron TeamTradetron team
The difference comes from which candle you test on. Position 0 is the current, forming candle. In a backtest, 0 on “open” can effectively use prices that were not known yet at the decision time, so the result looks much better than reality.
Use the closed candle: Position(EMA(13), -1) > Position(EMA(10), -1) (or the Cross keyword).
Treat the more conservative result as closer to real.
Check results out-of-sample (a period you did not tune on) before trusting them.
A backtest is a claim, not proof.
Answer written with AI from Tradetron docs (Oct 2026). Original replies are kept above. Spot something wrong? Reply below.