Insane doubles

[BurnzAll]

Just played a game guy rolled 6 doubles in a row... Then non double then double. Insane, I wonder what the odds are in this think there was only 2 rolls before this and it was both 6/5

[euklid314]

I will only give a generic answer.

At a backgammon game of average length, there are approx. 10^60 different roll combinations. So if you ask after a game what the probability for exactly those dice were, the answer will always be the same: 10^(-60). The answer does not depend on the dice that were rolled in that game.

And *every* Backgammon game again one of these absurdly unlikely events will happen - because one of the 10^60 possibilities is guaranteed to occur.

[euklid314]

Just played a game guy rolled 6 doubles in a row... Then non double then double. Insane, I wonder what the odds are in this think there was only 2 rolls before this and it was both 6/5

And I looked at your game #483653112. Your opponent really had a very, very (!) lucky streak BUT when you rolled 5-3 and you resigned you still had a winning chance of 4.10% (by moving 24/16)! Even if you make the suboptimal move 6/3 6/1 your winning chances are 3.49%. The analysis is done with XG (Extreme Gammon).

It see it quite often that after a lucky series of rolls for their opponent some players assume that the opponent will have lucky rolls for the rest of the game and resign. This assumption has no foundation, however, - which is one of the fascinating aspects of Backgammon! Exactly those victories are especially sweet when the whole game was going south and then one or two rolls change the whole momentum.

Of course you will lose 23 out of 24 games in the final game state when you resigned - but why not test out if it was the special day that you will take profit from the 1 in 24 chance (i.e. 4.1% winning chance)?

Just a short reasoning why your chances of a turnaround at the resigning moment were still intact: After you use 5-3 to run with 24/16, your opponent will need to leave a shot after he rolls 6-3, 5-3, 4-3, or 3-2 (i.e. in 22% of all rolls!). If you hit him you have a decent game (almost an even game) because your 5-point home board is so strong. And even if your opponent does not leave a shot now, there are certain rolls for him like 6-1, 5-1, 4-1 that make it likely that you get a hitting chance one turn later.

[BurnzAll]

Thanks for the reply I know rarities can happen, Its just really rare, I've played thousands of games in real life and haven't ever seen anything like it. The odds were just crazy just sharing

[euklid314]

Yes, true - really rare. Nobody will deny that. Next time you play on and still win the game. Then it is the icing on the cake!

[The_Trioker]

Thanks Euklid314 for your valuable insights !
A side note enquiry : is there anyway to export a backgammon game to a format directly usable in XG (such as sgg, snowietxt,oldmoves...) ? I only have GNU Backgammon myself, BTW.

[euklid314]

Thanks Euklid314 for your valuable insights !
A side note enquiry : is there anyway to export a backgammon game to a format directly usable in XG (such as sgg, snowietxt,oldmoves...) ? I only have GNU Backgammon myself, BTW.

I don't think so. For backgammonmatchplay (which is Alpha status for years now) it was planned to implement a export functionality. I enter the moves by hand if I want to analyze.

[The_Trioker]

Thanks for the info Euklid314. I've been searching for this option in vain.