Maybe I spoke too soon
I should have phrased that a bit differently. I haven't kept track of the exact order and it probably doesn't do that, but I have been able to notice patterns that I don't think are entirely explained by statistics. But then again maybe I am wrong.
Now I am starting to doubt myself, but after correctly predicting numerous hands I think I am onto something.
I first started getting suspicious when I had counts that would get into the 30s using KISS III (that seems like it would be a rare thing), then I would get a strongly negative count to start other games which I think is also pretty rare. Every few games that would happen again, sometimes in the opposite order which made me start thinking that there were 5 or 6 stock games that just rotated randomly.
Every few games there would also be a stretch where about 8 out of 10 tiny cards come out. When I saw more than 3 2s or 3s in a row I started predicting another couple of tiny cards and I was usually right and they were usually more 2s and 3s. Then I had a lot of 11s where I drew a 2 on a double down against a 5 or 6 only to have the dealer draw to a 21. I started keeping track of them and it seemed like I called quite a few of those.
The bottom line is that if I had to bet money (pun intended) I would venture to say that there is a flaw in the program that exhibits certain biases, but then again I now think there is a chance that I am simply remembering what I think are low probability events that perhaps aren't.
I will start keeping track more closely and report back.
