BGonline.org Forums

History of cancelgammon and why it didn't catch on

Posted By: MK
Date: Tuesday, 28 January 2025, at 1:33 a.m.

Ideas for gamblegammon variants, that could have been more aptly named "rerollgammon" or "redogammon", have been discussed in RGB for decades and I participated in some of them. By doing some recent searches I learned a few new things that I didn't know before.

Apparently, the first written mention of "rerolling" was in a book titled "Goren's Modern Backgammon", which gave credit to Richard Frey for the idea, all the way back in 1974.

Here is a RGB thread from 1997, showing that it was already being discussed back then in RGB, (before I had discovered RGB, FIBS and Jellyfish in 1998).

https://groups.google.com/g/rec.games.backgammon/c/5TfNwDy3PNI/m/Ry3VTFQIRgcJ

Unlike how it was wrongly played in recent tournaments, even the earliest suggestions in RGB proposed rerolling both the opponent's lucky rolls and the player's own unlucky rolls, as logic would dictate.

Searching for "cancelgammon" only finds one "bkgm" thread from 2004: Cancelgammon strategy

https://bkgm.com/rgb/rgb.cgi?view+1151

Which proposes a weirdly different starting position, totally unnecessary and as lame as other starting position based variants like "nackgammon", but includes other suggestions that apparently were later implemented in some tournament events and also some interesting comments about how it complicates the game.

And one "bgonline" thread from 2008: What is Cancelgammon?

https://www.bgonline.org/forums/webbbs_config.pl?read=24451

Which was about the first side event in Madison Wisconsin.

In one post, Tom Keith wrote: "That brings up the question, what would be worth more -- owning the doubling cube, or owning the cancel marker?"

It was interesting to see that he had made the same comment that I have made in one of my recent posts here on the subject.

Knowing that the cube magnifies luck, whenever a variant to reduce luck in gamblegammon is suggested, I perceive it as an "anti-cube" strategy.

Canceling a lucky roll in traditional backgammon would have little effect and mostly in the late stages. In a game where doubles are allowed as opening rolls, players aren't as worried about luck and especially early luck since all games are played out to the end.

In gamblegammon it has a much bigger effect, perhaps nearly as much as the cube itself, by not really canceling a lucky roll but by canceling a cube action that would arise from it, (or conversely instigating a cube action). It can complicate the game enough to make it much harder to master and can reduce luck enough to make it much less exciting for most people. I think this is one reason why it never caught on.

The other reason is that there has never been a bot that could play cancelgammon.

In fact, that is the only reason why I never played it even though I would have liked to.

Since I was first exposed to it 25+ years ago, I never played gamblegammon against a human in real life. Early on, I played on FIBS for a few years against assumedly human players but aside from that I exclusively played against online and offline bots.

Personally, I never tried to play like the bots but I see that everyone else does, to the point of caring more about playing with a low PR than about winning the game. I think even, or should I say especially, "giants" never got into cancelgammon because they couldn't consult bots to figure out the correct cancel actions and thus couldn't go beyond making "wild-ass guesses". And not knowing whose "wild-ass guesses" were better/best, would kill the spirit of competition.

I will end this post by declaring that we urgently need cancelgammon playing bots! (Which will be the subject of my next post.)