I think we all agree that simple backtest data, no matter how impressive, can be misleading. The test period can be unrepresentative, the parameters can be over optimized, the number of parameters can be increased inappropriately to curve-fit to the data and slippage can be accounted for in an unreasonable manner. I would never advise anyone to buy a system unless it has a real time track record, verified by an independent party like FuturesTruth or at least has been tested by them and confirmed as reasonable. There is a data vendor called CSI that had a product that analyzed trading systems. As I recall the main focus was number of parameters. It basically analyzed the system and number of parameters to assess the likelihood of it being robust in the future.