kiss: anti-martingale works for some. you win, you 'double up'. you lose, you 'double down'. whether a win/loss is 1 trade or a sequence will depend on your frequency of trading whether double up/down is literally that, or another clip is another decision. a positive expectancy is vital for this strategy however, so may not meet the op's criteria. having said that, i dont think any money management or portfolio allocation methodology will turn a lousy trader in to a profitable one. just my 4cents worth (i doubled up!)
I like Ralph Vince's Leverage Space Portfolio model. Implementing his principles requires you to know about optimization (e.g. Genetic Algorithms) and requires a bit of coding, so it may not be for everyone. What I find especially interesting is that he address geometric growth under the restriction on drawdown probability. So the optimization target becomes "give me the allocation between instruments that maximizes geometric growth but that has a maximum 5% risk of reaching a 10% drawdown within the next month". Since this relies on historic data you obviously have to live with the limitations this gives you, but I think this approach beats rules like "fractional Kelly", especially if you trade a lot of different instruments since the co-movements under various types of markets will be included (again within the quality the historic information gives). The method can also help you select between different instruments to get maximum growth in the case you have several choices.