Bill Robertie's Blog | The Gammon Press

How Match Score Affects Early Game Play

Mar 13, 2023 | Backgammon Problems: Tournament Tips, Uncategorized

Center cube. Black on move.

Black to play 3-3

(a) in a cash game.

(b) at double match point.

(c) trailing 1-2 in a 5-point match.

(a) Cash game.

As a cash game play, this position has an interesting history. When I started playing serious backgammon in the late 1970s, the game was dominated by a group of aggressive young New Yorkers centered at the legendary Mayfair Club, who emphasized making big plays early on, to set up quick cube turns and gammons. They were bold, ingenious, highly aggressive, and very successful, and most aspiring players followed their lead.

At the time, the old-fashioned play with 3-3 after an opening split to the bar was the obvious 13/7*(2). Why not? It hits a blot, makes a good point, leaves no blots, and unstacks the midpoint. That’s a lot of good things at zero risk. The New Yorkers rejected this play and instead played 13/7* 8/5(2), in accordance with the new (at the time) emphasis on making the 5-point early on no matter what the cost.

When I saw the new play I enthusiastically switched, following my general rule of thumb for beginners: “Ape your betters”. (At least until you know enough to question their judgment.) After a few years, though, I started to have second thoughts. How bad could 13/7*(2) really be? It had a lot to recommend it: good point, no blots, takes a solid initiative. I did some analysis and just couldn’t convince myself there was anything wrong with the play, so I switched and started to play it. I got some hoots of derision, but pretty soon more and more players were switching over, and by the 1990s this simple play was the standard, and the ‘old-fashioned’ 13/7* 8/5(2) looked ‘too loose, too fancy’.

Extreme Gammon rollouts long ago confirmed that 13/7*(2) is in fact best, while 13/7* 8/5(2) is a fairly close runner-up. Also in the mix is 13/7* 6/3(2), also making an inner point, unstacking, and leaving fewer shots and blots. Prior to the bot era, no one would even have considered making the 3-point, although I think most modern players would immediately spot it as a strong alternative.

(b) Double match point.

Double match point is simply any match score where gammons don’t count for either side and the result of this game decides the match. The simplest example occurs when each side is one point from victory. Other match scores can convert into double match point (DMP) pretty easily. For instance, if both players are two away from victory, and someone doubles quickly and the other side takes, then we have a double match point situation by default.

With gammons not counting, strategy in this situation undergoes a few changes. Here’s a quick summary of the major alterations.

1. The best game plans at double match point are (a) escaping your back checkers and winning the race, and (b) building a prime and winning the race. Racing advantages create a lot of wins but not a lot of gammons, which is perfect.

2. Blitzes are a poor choice of game plan unless the dice allow no reasonable alternative; too many gammons, not enough wins.

3. Anchors are less valuable because one of their functions is to prevent being gammoned, which is no longer necessary. The strength of an anchor lies in its ability to transition to a favorable race.

4. Inner board points are still good, but not quite as good. We don’t care about their ability to create gammons but we do care about their ability to create primes.

5. Slotting to build an inner board quickly is mostly wrong.

6. Plays where you run off an anchor to avoid being primed later are commonplace.

7. Escaping the back checkers is even better than usual.

Taking all this into account, it’s not hard to see that 13/7*(2) is even better at DMP than in a cash game. It gains ground in a race while taking no risk of losing ground in the race. The point is useful as well.

(c) Trailing 1-2 in a 5-point match.

Trailing 1-2 to 5, or ‘4-away 3-away’ as it is sometimes called, is a score where gammons matter more than usual for the trailer. If the trailer wins a doubled gammon, he wins the match. If the leader wins a doubled gammon, he still wins the match but some wastage is involved (he doesn’t need the fourth point). As a result, the trailer needs to create positions where he has an elevated chance of winning a gammon, at a reasonable cost. He doesn’t want to go crazy, he just wants to take a little more risk for extra gammon chances.

Both 13/7* 6/3(2) and 13/7* 8/5(2) fit the bill here. They’re the second and third best plays in a cash game, but here the extra inner point creates more gammons at a small cost in raw winning chances. Which of the two is best? A rollout left the two plays in a dead heat. Playing 6/3(2) makes a weaker point, but as compensation it unstacks a big stack and leaves fewer shots.

Back Games: Redoubling and Taking

Feb 1, 2023 | Backgammon Problems: Back Games

Cash game. White owns the cube. White on move.

Should White double? If he does, should Black take, drop, or beaver?

In our last problem we looked at a back game where the back game player got a double shot, but still couldn’t double until he actually hit the shot.

While that’s a generally good rule for many normal back game positions, it’s not always the case. When playing a massive back game, stay alert for positions where your opponent’s game has deteriorated so much that you can double much earlier.

Back Games: When to Redouble

Jan 15, 2023 | Backgammon Problems: Back Games, Uncategorized

Cash game. White owns the cube. White on move.

Should White double? If he does, should Black take, drop, or beaver?

Positions like this are fairly commonplace when playing back games, and they’re tricky enough to lead many players astray.

White has played a 5-2 back game and has just gotten a double shot. Besides fives and twos, he can also hit with 4-1, so he has a total of 22 hits and 14 misses this turn. In the future of course, White may get even more shots.

Some players, encountering this position for the first time, might reason as follows:

> Wow, I have 22 hits! I’m already the favorite!

> And if I hit, I can easily lose my market!

> And I might get even more shots in the future!

> And I could win a gammon!!

> I double!!!

Anything wrong with this reasoning? Well, just a little bit. Let’s look more closely.

Let’s start by examining what happens when White doubles. True, he then hits 22 times out of 36, a bit more than 60%. But how many of those can he win, once Black owns the cube? Let’s give White a pretty average hitting roll, like 4-2, and see how we think he’s doing.

Position after White hits with 4-2 and plays 20/18*/14.

Although he just got hit, Black still leads by 50 pips in the race, which has to be worth something. In addition, he has some pretty good shots in this position: 6-2, 5-2, 6-1, and 1-1, all of which leave Black in a lot of trouble. Let’s be conservative and say that Black can pull the game out a little less than a quarter of the time from this position. In that case, out of White’s 22 hits, we’ll say he can win about 17 of them, losing the other five.

What about White’s 14 misses? Now he’s in real trouble, trailing by 60-70 pips in the race, with his opponent owning the cube. Black won’t redouble immediately, but he’ll be able to use the cube very effectively in the near future when his position improves a little more. At worst, he’ll at least be able to play the game out to the end. I’ll give White four of these games, but no more.

Overall, White’s chances after doubling look like this:

If he hits, he wins 17 and loses 5.

If he misses, he wins 4 and loses 10.

With the cube on 4, White wins about 21 and loses about 15, for a net of 6 games, worth 4 points each. Total, +24 points.

Now suppose White keeps the cube and rolls. Again he hits 22 and misses 14. But now his 22 hits are very strong. Black needs a great shot immediately or he’ll pass a double, and even if he gets his shot, White can hang around until the end. Let’s make White about a 20 to 2 favorite in these games.

If White misses, he still has plenty of chances in the game. Let’s give him 6 games of these 14, losing the other 8. Now, his chances after not doubling look like this:

If he hits, he wins 20 and loses 2.

If he misses, he wins 6 and loses 8.

With the cube on 2, White wins about 26 and loses about 10, for a net of 16 games, worth 2 points each. His total here is +32 points. So he does substantially better by not doubling.

This is a pretty typical result when you get a double shot in a back game. You’re a big favorite if you hold onto the cube, but just a small favorite if you double. Barring some unusual features, these positions are mostly all no double and take.

Asset or Liability?

Jan 1, 2023 | Backgammon Problems: Holding Game

Cash game. Black owns the cube. White on move.

White to play 6-2.

Extreme Gammon (XG) is the most useful tool ever invented for improving at backgammon. It’s both the strongest commercially available backgammon software, and, when used properly, the best teacher.

Getting the most out of XG’s teaching ability, however, requires a little thought and planning. In this post I’ll give an example of how using XG’s power, combined with a little critical thinking, can help you patch weak spots in your game.