My BJ simulation needs help

I made a blackjack program in java to simulate playing 1000000s of hands. The problem is after all the hands are played the house ends up with very little edge (.001% or so). The variations I put in were... 8 decks, hit on soft 17, double down any first two cards, no double after split, split up to 4 times, BJ pays 3:2. I know that this should give the house around .8% edge. Can anyone think of what I could be doing wrong? thanks.

I am not familar with java programming but I do know that if your random number generator isn't up-to-par then it could cause some skewing in either direction. Are you using the computer's built-in random number generator for your programming?

I think Norm solved this problem in his software but I have no idea what he did nor do I think it would ever be disclosed.

I am using the built in generator which should be sufficient. Tonight at the casino I just discovered that after splitting aces one can only hit once (unless if get another ace then resplit), i put this in the program but it changed the edge very little. I think it is other stuff like this that I have missed.

The most likely reason for an error that large is that certain types of hands are not being handled correctly. If you are using a debugger, you might set breakpoints and step through the program, watching what it's actually doing at each critical phase. (If you try that, I'd suggest pre-arranging the deck to cause specific situations -- splits, doubles, soft Aces, dealer or player naturals or both, etc. -- so you don't have to step through a lot of ordinary hands to see those.) Or, you might find it easier to log the results of each round to a file, showing the actual cards in each of the completed hands along with the result determined by the program and the current bank. I suspect you'll find that an entire category of hands (or maybe bets) is not being handled correctly.