Discrepancy between backtesting results and live trading

During my testing, I noticed behavioral differences between the backtesting and the live trading using the same strategy with similar settings.

What are the factors that could influence this? I can only think of the backtest engine's `speed` parameter, but I'm not sure this is a sufficient explanation for all differences noticed.

Thanks for the help and suggestions!