Backgammon Lesson – 2
Beginners Please – Lesson 2
by Paul Money – July 2006
In Lesson 1, we took a look at how to play the opening roll. Logically we should now be looking at the response to the opening roll, but it isn’t that easy.
There are about 30 different reasonable ways to play the opening roll and if we multiply that by the 21 possible numbers that we can roll in reply, you can see that we would be looking at about 600 or so responses. This is without taking into account the different requirements for money play and match play.
So, what I am going to do is give you tables in forthcoming lessons with all the responses to one particular number, so that you can build up a reference library of these. These are my personal idea of a good play, but there are many instances where one or more alternates are equally good, possibly even very slightly better. However, you certainly won’t go very far wrong if you base your game on these.
A word of warning… these are not to be referred to when playing online. Using any aids to help you play online is as illegal as it is live. I agree that it doesn’t matter much when you are just playing for fun but I think that we should all be scrupulously honest at all times. When you are playing for money, cheating is of course theft. Don’t be tempted. Also, by not using aids, you will learn the game better.
Note: The moves suggested above are the #1 results of rollouts, however, sometimes other possible moves for a dice roll may have been listed as very close or as a reasonable alternate.
Responses fall into four classes. The first column is the basic response for money play or early in a match. The second column is for Double Match Point or DMP. The third and fourth columns give the best responses during the Crawford game in matches. We will come to these later on in another lesson.
Note: See Article #1 for the definition of DMP as well as other terms and conventions used in these articles, see the GammonLife Jargon section for other terms such as GammonGo (GamGo) and GammonSave (GamSave).
The 6-5 opening roll is a very good one, in that is the largest roll possible to start and also escapes a checker all the way to the mid-point. Its weakness is that it stacks another checker on the 13pt and does nothing to create any extra good numbers next turn. This means that you can play with some freedom when you reply to it, to try and grab some positional advantage while your opponent is in a rather inflexible position. Try to associate a lot of checkers on one point with inflexibility and lack of choice. It is normally something to avoid.
The usual reply to a 6-5 is to split the back checkers, to maximise contact and increase your chances of hitting something next turn. In the cases above where I choose a non-splitting play, such as for 5-2 and 2-1 for example, the splitting play is probably at least as good as my choice.
Let’s take a look at a typical one point match. The strategy for cubeless games where gammons don’t count is quite simple, “Stay Pure”. In backgammon terms this means that you try to avoid playing checkers to the lower points in your board, so that you can keep your forces to build a nice game winning prime. Hitting is a little less important than usual, with no chance to use the cube after a bad sequence for your opponent. Obtaining a high anchor, that is an anchor on the 21, 20 or 18pts is more important than usual, as it is very difficult to prime and also enables you to be more adventurous in your own prime building attempts.
Set out the checkers on your board at home to follow the play. Don’t try to learn too much from this sample game. What I am trying to do here is to get you used to following an annotated game and to give you some insights into the thinking behind the plays.
Move 1: White to play 6-3
Pip counts: White 167, Red 167
• White moves 24/18 13/10
Annotation
White opens up with a major split, the usual play with this poor roll, but do keep in mind that running with 24/15 is just about as good and will lead to a shorter and simpler game. This may suit you if you are playing a much stronger player.
Move 2: Red to play 5-5
Pip counts: White 158, Red 167
• Red moves 8/3(2) 6/1*(2)
Annotation
The game plan for one point matches is “stay pure”, i.e. avoid committing checkers to the lower points and try to build a containing prime.
Sometimes however, the dice say otherwise and you have to go with the flow. Red must attack with this 5-5 and the standard play is this 8/3(2), 6/1(2), but there is an alternate play that is rarely tried, which is 13/8(2), 6/1*(2). The five checkers stacked on the 8pt have a very awkward look, but then so does the stack on the 13pt. What this play does is to move the stack into a much more useful place, from where those checkers can make points or attack in the home board. This is known as getting men “into the zone”, much like a military battle plan these men are now in the front line and ready for action; you don’t have to waste time bringing them up from further back.
This is already quite advanced stuff, but it usefully demonstrates how little in backgammon can be taken for granted even this early in the game!
How about just playing 13/3(2)? This is very feeble leaving the game about even. “When in doubt, hit!” is good advice and clearly right here.
Move 3: White to play 2-1
Pip counts: White 159, Red 147
• White moves bar/23 18/17*
Move 4: Red to play 6-6
Pip counts: White 156, Red 164
• Red cannot move
Move 5: White to play 5-4
Pip counts: White 156, Red 164
• White moves 17/8
Annotation
A nice steady play makes progress after Red’s sad 6-6 from the bar. This is better than 23/18, 17/13 because White doesn’t want to give Red good sixes to play when he enters. After the match play, his sixes are very difficult.
Move 6: Red to play 1-3
Pip counts: White 147, Red 164
• Red moves bar/22 24/23
Annotation
Red makes a nice play here, exposing three blots in return for great flexibility. Bar/24, 13/10 is very safe, but very inflexible. Three checkers on the 24pt is a notoriously bad formation, usually to be avoided. Bar/21 is quite reasonable, but that blot can be attacked by White on a point that she would really like to make. This is known as “Coming under the gun”, sometimes necessary, but not here.
Move 7: White to play 4-6
Pip counts: White 147, Red 160
• White moves 10/4 8/4
Annotation
White avoids the temptation to make the 2pt on Red’s head and just makes the 4pt. This is a good steady positional gain, entirely consistent with White’s optimal game plan. She has a solid race lead and has one man back as opposed to Red’s three, so no need for a flighty attack on the 2pt.
Move 8: Red to play 3-6
Pip counts: White 137, Red 160
• Red moves 22/16 13/10
Annotation
Red makes another excellent play here, neatly demonstrating an understanding of the priorities in this position. He must unstack the midpoint to create more chances of making a point in his outfield, something that is badly lacking at present. He could also try 24/18 with the 6, but it does give White every number to hit something which seems a bit much! 22/13 is pretty safe but does nothing to improve the position, too small as the experts say, while 13/7, 13/10 exposes two more blots (too big!).
Move 9: White to play 1-6
Pip counts: White 137, Red 151
• White moves 13/7 8/7
Move 10: Red to play 5-1
Pip counts: White 130, Red 151
• Red moves 10/5 6/5
Annotation
Red might reasonably play safe with 16/10, but he chooses the superior play of making the 5pt. There is a very useful lesson to be learned here, which is that it is almost always worth leaving a blot in the opponent’s outfield in order to make a good point in your home board, particularly true when it is the 5pt.
Move 11: White to play 5-1
Pip counts: White 130, Red 145
• White moves 13/7
Annotation
No need for anything dramatic here. White still leads and has a good solid position, while her single back checker is under no immediate threat.
Move 12: Red to play 5-3
Pip counts: White 124, Red 145
• Red moves 16/11 13/10
Annotation
This is a very difficult position to sort out. Red’s play is quite reasonable, but the computer rollout indicates that 23/15 is probably best. There are several factors at work here. The match play creates a fourth blot and allows White 8 numbers to hit from the back as she escapes. 23/15 keeps the blots down to three, escapes a checker when not hit and perhaps crucially, forces White to break her midpoint in order to hit, a significant concession. Can it really be right to leave a double shot? It can in these cubeless games, where gammons don’t count. If White hits in the outfield, she isn’t making another point in the prime that she wants, while Red has little to lose by being hit, as he is a big favourite to re-enter. The differences are not large, but the roll highlights the sort of thing that top players are considering when selecting their play.
Move 13: White to play 3-6
Pip counts: White 124, Red 137
• White moves 23/14*
Move 14: Red to play 2-2
Pip counts: White 115, Red 151
• Red moves bar/23 24/20 13/11*
Annotation
A great roll! Red enters, hits and moves a checker up into the “escape hatch”, the 20pt. He could hardly ask for more.
Move 15: White to play 2-3
Pip counts: White 126, Red 143
• White moves bar/23 7/4
Move 16: Red to play 6-2
Pip counts: White 121, Red 143
• Red moves 20/14 11/9
Annotation
Red makes a sizeable error with this play. He has a much stronger home board than White and will thus be the favourite in a blot hitting contest. He still trails in the race and as he has to leave blots anyway, so he may as well put White on the bar.
TIP: Very often the best way to take advantage of a stronger home board is to hit.
Move 17: White to play 6-2
Pip counts: White 121, Red 135
• White moves 23/15*
Annotation
White’s play is probably best, but 13/7, 13/11* is almost as good, leaving fewer fly shots from the bar and moving two builders into position to make the crucial 5pt.
Move 18: Red to play 4-4
Pip counts: White 113, Red 150
• Red cannot move
Move 19: White to play 1-3
Pip counts: White 113, Red 150
• White moves 15/11*
Annotation
This is an ultra tough roll, one that I would expect a lot of experts to call wrongly. White elected to pick up a second checker, 15/11*, but sending back a fourth checker has some downside to it. Red may be able to make a second anchor somewhere and even if he can’t, he shouldn’t mind doing nothing for a couple of rolls, while he waits for a hit later on. The time wasted should enable him to keep a strong board for a number of turns to come. This introduces us to the idea of timing, one of the least understood of all backgammon concepts and one that we will return to again and again in the future. For now, let’s just note that Red has nothing to lose by being hit.
White probably also considered 8/5, 6/5 making a vital point but breaking the useful 8pt. This leaves a startling 19 return shots for White. The pro play is in fact 6/3, 4/3, making a fifth point in the prime and cutting the return shots to a more acceptable 12.
This is a Goldilocks problem. The hitting play is completely safe this turn but actually achieves very little. It is like Mother bear’s porage, too cold. Making the 5pt makes a big gain but carries a lot of risk; it is like Father bear’s porridge, too hot. Making the 3pt is a reasonable gain for a smaller risk; like Baby Bear’s porage, just right! Humans tend to gravitate towards the extremes when problem solving, but the correct answer often lies somewhere in between as it does here.
Before the advent of the bots, keen players used to roll out problems manually. The data that they gathered was not particularly accurate, because the sample size was usually too small and the skill of the players inadequate, but what they did learn was what sort of positions the differing plays lead to and how to play them. You could try that with this position and move 12 could also be very informative.
TIP: Don’t just accept what you are told. Roll out the plays for yourself, playing both sides or alternating with a friend. Understanding acquired in this way will stay in your brain much more readily than anything read or heard.
That’s plenty for this lesson. We can finish this game another time.
Homework time! Using the probability square that I gave you last time, I want you to work out the chances of entering from the bar when your opponent has one, two, three, four or five points made in her home board.
To start you off, if she only has one point made (usually the 6pt, but it doesn’t matter which), then only 6-6 fails to enter, while all the other 35 rolls do so. Your chances of entering are 35/36. Now see if you can work out the rest.
Last time, I asked you to create a chart showing your chances of hitting a blot x pips away and if you did, it should look like this. How does yours look?
| Pips away | Rolls to hit | Description of hitting rolls |
| 1 | 11 | Any 1 |
| 2 | 12 | Any 2 + 1-1 |
| 3 | 14 | Any 3 + 2-1 & 1-1 |
| 4 | 15 | Any 4 + 3-1, 2-2 & 1-1 |
| 5 | 15 | Any 5 + 4-1 & 3-2 |
| 6 | 17 | Any 6 + 5-1, 4-2, 3-3 & 2-2 |
| 7 | 6 | 6-1, 5-2 & 4-3 |
| 8 | 6 | 6-2, 5-3, 4-4 & 2-2 |
| 9 | 5 | 6-3, 5-4 & 3-3 |
| 10 | 3 | 6-4 & 5-5 |
| 11 | 2 | 6-5 |
| 12 | 3 | 6-6, 4-4 & 3-3 |
| 13 | 0 | |
| 14 | 0 | |
| 15 | 1 | 5-5 |
| 16 | 1 | 4-4 |
| 17 | 0 | |
| 18 | 1 | 6-6 |
| 19 | 0 | |
| 20 | 1 | 5-5 |
| 21 | 0 | |
| 22 | 0 | |
| 23 | 0 | |
| 24 | 1 | 6-6 |
