Here's what I think about it after making probably over a million bets on various dice sites
22248573. Yeah, you read it right, I rolled over 22 million times at WB since July 24 (just in case, I'm still rolling). I think I can safely assume that during this rolling marathon I've seen quite a few outliers. And you know what? They definitely come in packs (maybe that's because of the casino's prime theme, if you know what I mean). For example, you can roll for a whole week and see nothing of interest. Then you witness 2-3 massive outliers within a single day that will give you jitters and make you sweat. You kinda come to expect them in series, and that's a real thing with your skin in the game, not something to be discarded as inconsequential
So much for the so-called randomness
I can tell you what I "noticed" playing couple of years ago on PD. I was playing manually, so I could pay attention to each bet, and I was hunting 99x payout, and it seemed to me that every time after the dice rolls over 99.00, the very next roll was below 1.00. So, I started to exploit my "knowledge", and you know what? It worked! Not every time, but enough to make me think I could do it forever, winning some money in the process. Then, after it stopped working, I "improved" my betting strategy by supposing that it was going to be not the very next roll, but that if over 99.00 was rolled then a roll below 1.00 will definitely happen within the next 10-15 rolls (and vice versa). It seemed that it was working for some time too, but then it stopped working, and after losing a significant amount I decided to abandon that strategy for good
I can explain to you what happened
And what happened is actually quite in line with the idea set forth in the OP. More specifically, you noticed a pattern, and had been exploiting it for some time. But since bets are random (allegedly), the patterns are random too (necessarily), so it shouldn't in fact have surprised you (had you been aware of this intrinsic property of genuine randomness at the time) that your "strategy" stopped working after some time as another pattern had most certainly revealed itself (which you failed to discover and take advantage of). Patterns are random, but their existence itself is not random at all. It is a feature of a truly random distribution
Taking into account that rolls on PD are provable fair, I tend to think that all of those were just coincidences. But if you think that provably fair outcomes can somehow be not random, then this topic is definitely worth further investigation
No, I'm not saying that. What I am saying, though, is that true randomness destroys itself at a higher level by always being randomly patterned (given enough sample size, of course). And, more importantly, this feature can be taken advantage of in certain circumstances (with some precautions)