Comment on Streak just got a whole lot better

I have been studying some signals and creating algos based on them. However, I am facing troubles during backtesting.
The backtests sometimes dont pick up very obvious signals that fulfill all the conditions. In a recent example, my bactest for a day didnt pick up an important trade, and when i further reduced the backtest period to just one hour of that trade, it picked it up.
The first time i ran this backtest it showed me a loss of 1.2% and when I backtested the same algo hour wise, and then again did it for the whole day, it showed me a profit of 6%. It is quite dangerous if we cant rely on backtests for accuracy

View the full comment thread »