Woody,
Yes, they do. If the move is against you, you will learn about it real quick. We don't need proof for that because it happens to everyone many times a day. What people only suspected up to now, and now know for a fact, is that they remain silent when the line moves in your favor. They just pretend the line never changed and hope you don't notice that you got screwed.
It's really hard to catch them on this, because even if you had 100 bets rejected and they all went the wrong way (none in which the line actually got better) that still doesn't count as proof. In our example, we had a line that changed at 1:22 and they still accepted the old line at 1:30 because it was in their favor to do so. Eight minutes is a very long time. In the other direction, the bet is rejected in an instant. In this direction, who knows how long they'll accept the old, less favorable line!? It's at least 8 minutes, we know that much.
It's nice to see people finding the experiment interesting. I was debating not doing it because as long as I knew this was going on, I could just refresh often and keep a relative advantage over other bettors, but on further reflection I decided that the larger question of sportsbook fairness is much more important. I look forward to a day when sports betting has the same status as banking or insurance in terms of professionalism. If that means regulation, then so be it. If not, all the better. That would be good for all of us, myself included.
I should point out that other books like WSEX, Canbet, and sportsbook.com have a fair, symmetrical bet rejection policy ie. they reject it if the line moves no matter which way it goes. At least this is what my experience shows. There are no doubt lots of other books too, but these are the ones I know about for sure.