Finally! Real Backgammon Information

Contributed by Gerry Tansey

(Thanks Gerry!)

Finally, the moment we’ve all been waiting for – real news you can use from Gerry Tansey, a source you can trust!

Take it away Gerry:

Tansey Position


The match score can have a huge effect on the proper checker play.  In this position, Black is trailing 3-2 in a 5-point match and gave an aggressive double earlier in the game.  However, the tide has turned a bit, and now Black must make do with this roll of 43.

Note that since a 2-cube is in play, Black will lose the match if he loses this game.  Additionally, a gammon loss for Black carries no additional penalty.  Therefore, Black can afford to make the play that gives him the best chance of winning the game without worrying about losing a gammon.  But what is that play?

Red has quite a few threats, but the most deadly is making the 4 point and creating a five-point prime.  If Red is allowed to carry out this threat, Black will be nearly completely dead, as he will have three checkers blockaded and a front position that is nearing collapse.

The threats are so great that the best play is for Black to put Red on the bar by breaking his 6-point with the move 24/21 6/2*! After this play, Red has 9 numbers that dance, after which Black may be able to rebuild his board, make an advanced anchor, escape a checker, or hit another opposing checker.  Black has some real winning chances on Red’s 9 fans.  However, even if Black hits, all is not lost, as White may be able to establish an anchor or two in Black’s home board and generate some chances with a late hit.  Every other play is a serious error.

This type of play is known as a “banana split” (as in “You’d have to be bananas to make that play”).  In this particular position, the play would be too risky if White’s gammon losses hurt him, but there do exist other positions where the banana split is right even if the “banana splitter’s” gammon losses matter.  If your opponent’s threats are intolerable, sometimes you must take extreme measures to keep your opponent from carrying out those threats.


   1. Rollout¹    24/21 6/2*                   eq:-0.305
      Player:   30.56% (G:8.81% B:0.31%)
      Opponent: 69.44% (G:37.28% B:11.26%)
      Confidence: ±0.006 (-0.311..-0.299) – [100.0%]
      Duration: 11 minutes 39 seconds

    2. Rollout¹    24/21 12/8                   eq:-0.372 (-0.067)
      Player:   28.98% (G:5.13% B:0.24%)
      Opponent: 71.02% (G:29.66% B:4.85%)
      Confidence: ±0.005 (-0.377..-0.367) – [0.0%]
      Duration: 8 minutes 24 seconds

    3. Rollout¹    13/6                         eq:-0.375 (-0.070)
      Player:   28.39% (G:6.00% B:0.23%)
      Opponent: 71.61% (G:27.04% B:6.55%)
      Confidence: ±0.005 (-0.380..-0.370) – [0.0%]
      Duration: 7 minutes 31 seconds

    4. Rollout¹    24/21 13/9                   eq:-0.388 (-0.083)
      Player:   28.19% (G:5.08% B:0.24%)
      Opponent: 71.81% (G:31.54% B:4.88%)
      Confidence: ±0.005 (-0.393..-0.383) – [0.0%]
      Duration: 7 minutes 49 seconds

    5. Rollout¹    12/5                         eq:-0.399 (-0.095)
      Player:   27.33% (G:5.67% B:0.24%)
      Opponent: 72.67% (G:27.28% B:6.29%)
      Confidence: ±0.005 (-0.405..-0.394) – [0.0%]
      Duration: 8 minutes 17 seconds

¹  1296 Games rolled with Variance Reduction.
   Moves: 3-ply, cube decisions: XG Roller

eXtreme Gammon Version: 2.10, MET: Kazaross XG2


3 thoughts on “Finally! Real Backgammon Information

  1. I get the point about needing to win and having no gammon risk. But is 13/9 12/9 terrible because it allows Red to escape easily with (7) 5-rolls (6/5 x2, 5/5, 5/4x 2, 5/3 x2)? I assume so, but would have been curious to see the rollout. It must be so bad it wasn’t even worth a rollout? Please be gentile when responding. I’m a novice who likes to ask dumb questions!


  2. Hey, George. I like your question. The reason I didn’t roll out 13/9 12/9 is that I basically took the 5 top plays on XG’s initial evaluation and rolled them out. I didn’t think that making the 9 was *especially* bad. I just figured that it would end up being “just as bad” as all of the rest of the plays with less urgency. But I rolled it out just now, and it ends up being .096 behind the best play. I think your point about letting Red escape is a good one (probably why it is not as good as 13/6, say). But I just thought of something: if you think your opponent might be tempted to point on you with 54 and 53 after you make the 9-point, then making the 9 goes up in value (it’s clearly right to run at the score). Not by enough to sway the ultimate verdict, but something to think about.

    Note that for money, and with Black owning the cube (since Red should probably double Black out in this position for money after any sensible play), making the 9-point might be the best play. I ran a 1296 trial rollout with this play and 13/6, and the plays had a .001 difference (with 95% confidence interval of +-.008 for each play). Your play is a fine one under “normal” circumstances.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s