[Dire]

I'm not sure if it's what you're implying or not, but you seem to be stating the gambler's fallacy. It doesn't even out, on average. At flip 23 where you're down 6.5 points off expectation, you should expect to be 6.5 points off expectation, on average, for the rest of eternity of flipping.

You approach your expected equity relative to your sample size, not in absolute terms. So say you're 5 flips off after 20 flips. That's a 25% difference! You should now expect to always be 5 flips off, but the percent will decrease as the sample size increases. So if you run at expectation for the next 80 flips, you'll still be down 5 flips but it will only be a 5% difference. As you increase your sample size you obviously get closer and closer to 0% difference (in other words running at expectation) but you'll always be down 5 flips.