This is simply incorrect. Even if playing only at TC +3 and above, you will still lose more hands than you win. That is balanced by the fact that you win more on doubles, splits, and blackjacks, so you do in fact have an edge. But if your average edge is 2%, that means over the long term you will win 2% of each bet. If you are able to bet 10-15 hands per hour, then you should expect to win approximately 20-30% of a bet per hour. That means if your bet size is $75, you will make about $15/hr with a standard deviation of perhaps $300 per hour (I'm estimating here). That requires a $9000 bankroll in order to have a 5% Risk of Ruin. You can get the same results with normal counting in a game with a SCORE of 25, and most people would consider a SCORE of 25 to be unplayable.
Your plan will be profitable in the long run, but the variance is too high for the profit you will expect to make, in my opinion. Even if you wish to wong heavily, you should still spread your bets depending on your edge. I haven't studied wonging bet spreads, but I would guess something like betting a number of units equal to the true count minus one (meaning bet 1 at TC=+2, 2 at TC=+3, etc.) would probably be fine. Your unit size here would be somewhat bigger than if you were playing all. Making an educated guess, I'd say that a bankroll of 400 units would give you an RoR of around 5% with this approach, but that's based on a lot of assumptions. We can give you more accurate numbers with more details on your game. But the bottom line is that any approach that involves flat betting, even if you only bet when you have an edge, is going to be suboptimal. If you are doing the work required to keep the count, why throw that information away when it comes time to bet?