I haven't seen such a profitable sounding paper trading system since surfer described his world class PileDrivers(TM) system - "Where money goes to die."
I have "developed" over 200 systems using TradeStation Wasy Language. They all worked well on the simulator in back testing. About half of them worked in live trading with real money. Those that were profitable in live trading usually made only 30-50% of what they made in simulation. Therefore, I suggest you don't get too excited until you trade your system with real money in real markets.
It's difficult to debate a statement like that unless you tell us more about the systems you've created. There's a fine line between a really profitable system and 99% of the rubbish out there. One little flaw in your logic that you either never considered or didn't implement correctly can be all the difference. Are you sure TradeStation and/or latency isn't partly responsible for that?
Did you allow for deducting commissions, did you allow for slippage. Did you have LIB turned on? Anyone of these throws the results into doubt. Missing all three (especially LIB) makes it pure fantasy.
Steve, disregard my previous post, I meant no malice. What you say is correct, for the overwhelming majority, that is most likely the case.
Get one of those free GP engines, use your data and demand avg win/avg loss < 1 and with some code you will get hundreds of systems with high win rate. Then simulate forward and choose the ones that show high profit because there will be some that do that in that period but not afterwards. This article explains this well.