As per the thread title, I'd be interested to hear how people have approached simulating slippage, partial fills etc. in their proprietary backtesting engines. Do you look at market depth? How do you deal with backtesting at simulated speeds much faster than real-time? Do you gather slippage/fill statistics from live trading? Have you come up with some event stream processing filters?
Look forward to the next bar and combine it with the current bar. Buy at the high of those two, sell at the low. If my strat can't make money under these conditions, I keep looking. I figure once I find a profitable strat with this kind of handicap, I will be pleasantly surprised when I go live. As of now, I'm still looking.