When we research and develop trading signals, we are able to compare predictions in price movements with actual price movements realized in the market based on historical data. This, of course, doesn't necessarily require a fully-fledged backtester, but does require a historical data playback software. This component, while less complex than a backtester, still has a decent amount of complexity to it and must be accurate in synchronizing different market data sources and playing market data back with accurate timestamps and event synchronization. If the market data played back in historical research platforms is not identical to what is received in live trading strategies, then the signal predictions and performance observed in historical research is not realized in live trading and can kill the profitability of a trading strategy.