Fishhead -
It doesn't matter if you lose 11 in a row (except emotionally). If I hit 115 out of 200 games, and I resize after each play, the order does not matter.
So in answer to your question... Risk of ruin is 0%.
You don't even have to be perfectly accurate in your estimated edge to utilize this, as long as you don't overestimate your edge. Often, I'll have a play that I think will hit between 57-61%. I don't have enough data to narrow it down with confidence. In that case, typically use 57 (with a 95% chance of underestimating my edge).
Compare two betters who hit 115/200. Your flat 2% better is up (115 - 85 *1.1), or 21.5 units, or $430 if he had a $1k bankroll.
Your second better "underbets" using 10%. After those 200 plays, his bankroll is now: 1k * (1.091) ^ 115 / (1.1) ^ 85, roughly 5800.
There are two problems with the "Kelly" approach. First, it takes a LOT more work to accurately estimate your advantage. Any educated chucklehead can hit a decent winning rate, but he doesn't know if it is 53, 55 or higher. Once he figures out how to accurately evaluate his hit rate, he finds easy ways to increase it. But most people don't have the time or knowledge to do this.
The second problem with Kelly is that if you succeed, you are quickly hitting scalability limits, and attracting countermeasures. Your imaginary bankroll of $10k could hit 60k in one year, and you are looking to make 6k wagers. Betting the amount you want gets harder and harder, especially in the smaller sports that are more susceptible to 57%+ plays.