speachless: If the problem is that some image files are missing, then old games will look wrong when we play through them today. That doesn't mean they were affected when they were played.

Marshmud: I have the same problem as speachless. Just to clarify, let's take speachless' first example.

pgt: Look at the game ID in that message: 5710553 is in fact the game ID of game 4 in that match. So the message that you quote is not the message for the final game, it's the message for game 4.

pgt: The game you link to is game four of the match. You won that game, but your opponent won the match.

coan.net: There's an open bug report on the problem, plus a lot of previous discussion on this board: 

pgt: Well, there is a way to verify it: You go to the message you received when you finished your next Hyper Backgammon game after this one. The rating calculations there should say that your initial rating was 2264.

pgt: Hmm... that must be a bug. You should file a bug report if you haven't done it already (go to the message and then click the image of a bug in the upper right corner).

pgt: The link you provide doesn't show the rating change after this match. You should receive a message (under 'Events') that shows that your rating went up and his went down. If not, it seems to be a bug.

grenv: We can only speculate about the cause of the problem, since we didn't write the code.

If we are to speculate, one possible hypothesis could be that some unintended event prevents the dice from being generated in the normal way. Whenever this event happens, you get the same dice as the previous roll. When it doesn't happen, you get a random roll. We must assume that this hypothetical event would be something that happens at the second player's first roll - but again, this observation is only useful to someone who has access to the code.

As an aside, if this hypothesis is correct, the event seems to happen in roughly 1 out of 3 games.

Thom27: That's a great initiative! It seems clear that it is actually a bug and should be repaired as soon as possible. I wonder if it would be possible for your program to analyze a larger sample of games?

Nothingness: The point you may be missing is that you only resign the current game, not the entire match. In your example, resigning would mean that you only lead by 4-1 instead of 4-0. Accepting would mean that the current game counts for two.

grenv: Just checked my plakoto games from 2009. It's 10 cases of identical first rolls out of 27.

Another observation is that there are very few instances of both dice being different, e.g. 56, 43. I think the expected frequency of this would be 4 out of 9, but in the above sample it's only 5 out of 27. In playBunny's first sample, it is 0 out of 12. And in playBunny's second sample (the "reasonable" one), I count only 3 rolls out of 38 where both dice are apart.

Well... if I'm paranoid, at least I'm not the only one.

alanback: I just checked a small, random sample of games: the 13 games (or matches) of regular Backgammon played by alanback in 2009. Out of these 13, 6 games start with identical rolls by black and white.

I know, it's hardly conclusive. But still interesting.

I can only speak for myself, but no, I've got no statistics.

I wouldn't have suspected there could be such a problem, so I haven't especially been looking out for it. Even so, the tendency in my games has been obvious at times (several games in a row).

Now I learn that at least three other people have observed the exact same phenomenon. Sounds to me like it might be worth an investigation.

Pedro Martínez: Definitely not a coincidence - if it's anything, it's a lot of coincidences!
I noticed it too.

skipinnz: Thanks for the advice. Actually, luck is what I used in this game where I was 'caught' as a 'cheater', so I guess it's not effective if you want to do it secretly. Apparently, he believed I was using a program. I don't understand how a program would help you be lucky...

Does anyone know a way to cheat in crowded backgammon? I've just been accused of cheating, and if I'm going to be blamed for it, I want to have the pleasure of committing the crime as well.