Starting with $500 and risking 15% per trade (~$75) and averaging $3 profit per winning trade, you need ~25 winning trades to cover one losing trade (of course this is rough due to the account size fluctuation with each trade.)
This is fine if the strategy has a win:loss ratio of greater than 25:1. Have you done any back tests that indicate this sort of performance over time?
Don't get a false sense of security after 20 winning trades, up the ante and get slogged by a couple of consecutive losers, resulting in a 30% loss... been there, done that.
Also, 15% seems like a lot to risk with each trade. Most long term trading strategies recommend ~3% risk with each trade. A lot of EA's risk 15% in their testing, apply back tests to a period of time that is profitable and then use this as 'proof' in marketing material to show that the EA is capable of big gains...