Same EA settings, same inputs, but backtest results changed

 
I came across this issue, I was testing an ea for over 4 months, most of the time I just re-tested same settings and results were good, yesterday I decided to do one last test with the same settings before going live and the backtest result was bad...my question is what happened? why? Did the  broker update the historical data? All this makes me think that backtesting is so unreliable 

How to avoid this issue? Please advise, thank you
 

Do you have the strategy report html file, screenshots... before and after this issue ?


What do you mean by good and bad results ?

 
The backtest does not have all the details of a real real test, there may be in EA that you test slippage, spread and other filters that the backtest does not contain.
The backtest is based on a database of ticks that are recorded at best, but it happens that some ticks are not recorded and at the time of the backtest it is sometimes not the same result at all.
Reason: