for same strategy, results are totally gone negative in back testing which were showing good positive results 2 days back. back test results are totally changed since last 2 days. Previous results were correct or current results.
2 Likes
Hi @KO21XAaciZ, thanks for bringing this up. We’ve recently made optimizations to the Backtesting engine, particularly around indicator calculations and execution logic. As a result, you may notice differences in backtest results compared to earlier runs.
The exact same thing has happened to me. This is no minor change, the same strategy but there is a huge differences between old and new backtest results. My indicator-based strategy were positive and profitable after 3-4 rounds of backtest results but now backtesting the same strategy again I got the result fully changed in big nagitive.This is really concerning given that I was actually using this automated strategy in the live market daily. Now I’m totally confused and fearful about the strategy and all other backtest results because I don’t know which one is correct and whice one isn’t. If the new backtest results are correct then…
Exactly, I’m facing the same issue.
The strategy I backtested on 24 July showed an overall profit of +1488% for the selected time period. When I tested the exact same strategy today (31 July), the result changed to -218%.
Even if this is due to an optimization, I would expect some variation—not a complete reversal from an extremely profitable strategy to a heavily losing one. This doesn’t feel like an optimization; it feels like a fundamental logic change. At this point, I’m not sure whether the backtesting engine has been optimized or if something is broken completely.
This is a complete day-and-night difference. I’ve completely lost confidence in the strategy because I no longer know which result reflects reality or which backtest I should trust.
could you at least provide access to the historical options data? I’d be happy to build my own custom backtesting engine so I can verify the results independently.