your initial assertion was "you're getting bad rolls now, but chear up, you'll get the same amount of good rolls one day"
quick and simple, here's why i disagree, politely disagree
first, according to your own way if thinking, he MAY have been getting good rolls before he got into this streak of supposedly bad rolls. in fact, this would make perfect sense because the player may "think" or "feel" he's rolling bad cus he was so used to getting good rolls.
so with your way of thinking, in that case, he would NOT have any good rolls to look forward to, right? so that was a bad assumption for you to make i think.
second, let's just say for the sake of argument (to make things simple) that a guy starts playing pool and the first 500 times he comes to the table he gets what we would call bad rolls. this 500 bad rolls, as far as the stats are concerned, will really not be inclined to equal out. the reason you think they will is that over the years, yes, his percentage of getting unlucky will go down, and down, and down..... just because the sample size is getting so big compared to the 500........ but the 500 bad rolls will always be there, slightly favoring his percentage, if only by say .001%, to unfavorable rolls as a result of that 500. statistically, there is no reason to think he will get those 500 bad rolls "back" or make up for them though..... again, i feel the reason you think this is true is just due to the nature of the sample sizes..... a guy plays pool 10 years and yes, his percentage will be close enough to 50 50 to be negligable.
and if your're argument will be "well, he got the 500 bad rolls, maybe he'll get a really good streak and even out the 500 bad ones" well, you have to consider it would be just as likely he get 500 more bad as 500 good after the initial 500 bad, haha, confusing but true.
again, what the op was talking about was an aberration, or uncanny streak of bad rolls. if we assume he did get an unusually long streak of bad rolls, there is really no reason to think he will "get these back," again, he's just as likely to get another bad streak as he is to get a streak of good ones, as you assert he will (or should).
so, going back to the initail, easier to understand example, if i flipped a heads with a coin 1 million times in a row, no matter how many times i flipped it after that, there is no reason to think that this aberration within the stats will ever disappear. the reason for all this is, what we see as close enough to 50-50 (say 49.92% to 50.08%) can and will be even scientifically significant if the sample size is large enough. in other words, that 1 million heads in a row may appear to diverge toward 50-50, but as the sample size got large enough, even very small differences in the percentages would be deemed significant.