Improving Your Game
How To Learn From Your Mistakes
by Phil Simborg, 2008
Phil Simborg
Like most serious players, I really want to improve my game, so like most I make a lot of use of Snowie and GnuBG to analyze matches and positions and look at my mistakes. The problem is that most of us don't really learn by looking at our mistakes.

By learn I mean internalize. Understand. Know the reasons why one play or decision is better than the others. And do you know why I know most of us don't learn from our mistakes? Because I have been one of those stupid people myself for many years.

Like many of you, when Snowie tells me I made an error, I look at it, and I look to see what the right play is, and I think about why Snowie's play is better. If I don't understand why, I roll it out to make sure Snowie is right and then try again. And if I still don't understand why, I send it to some experts and they tell me why.

Then I hope I will remember that play and what was right and why it was right so I won't make the same mistake again when the same or a similar situation comes up again.

But this is not really learning. This is memorizing. And memorizing has made me about as good as any player in the world on the opening roll, and maybe even on roll 2. And that's about as far as memorizing will take you. The problem is that there are just too many positions and situations for any human being to memorize. The only way to truly learn, and to truly improve your game, is to understand why, and you can't do that by just looking at a position and trying to figure out your mistake.

Let me give you an example and show how you truly can learn something from a mistake.

Take a look at Figure 1 below and see how you think White should play 3-2 in a money game with Black holding the cube.

Figure 1
White to play 3-2

If you made the 5 point, you made the same mistake I made. It's right to hit two. Even if you did decide to hit two, do you completely understand why that is better than making the 5 point in this situation?

If you look at the Snowie Eval (on the right), you will see that by hitting two White wins 1 percent less, but he wins 8 percent more gammons.

Here, I have to admit, is where I stopped and believed I had learned something. But in recent years, some excellent coaching and advice I was fortunate to receive from many terrific teachers (particularly Perry Gartner, Howard Ring and David Rockwell), I have learned to go deeper.

The fact that you win more gammons is not why hitting two is correct; that is the effect, or result of hitting two. You have to dig deeper to understand why you win so many more gammons. Without going into great depth on this one position, the reason why you win more gammons hitting two is that if you don't hit two he may well anchor on your two point and it would be very hard to gammon him. Also, when you hit two you have the builders in position to allow you to have a good chance to make your 5 point before he anchors there. And, even if you do get hit, Black's board is not strong, so you can probably come back in quickly and continue attacking.

Now, how do I know that? How do I learn those things without having someone simply tell them to me? What Howard Ring and Perry Gartner almost always did when I showed them a position was to run the position, look at it, think about the reasoning, and then, they tested it. They tested it by moving a few checkers around to see if that changed the situation. They changed the location of the cube, or the score of the match, to see how that changed things. And every time they made a change we learned a little more about why one decision was better than another.

Take a look at Figure 3 below. It is exactly the same as Position 1 with the exception that White's two checkers on the Midpoint are now on Black's ace point.

Figure 3
White to play 3-2

In this position, it would be a huge blunder to hit two ... far better to make the 5 point. The difference between the two positions is that with White's checkers back in Black's zone, it would be much more difficult for White to close his board before Black anchors on his 5 point. So moving those checkers proved that this was one of the important elements to consider. By moving the checkers, I truly learned something. Something that applies not just to this position, but to virtually every position where you are debating between blitzing (attacking) or making points and safetying checkers.

If you took the same position and it was a match where gammons don't matter—say double match point—there again making the 5 point would be better simply because you win more games that way and gammons don't matter.

Now, let's take a look at Figure 4. Again it is the same situation and the same roll, but here we have given Black a stronger board with no blots.

Figure 4
White to play 3-2

In this situation, hitting two would be a blunder. So one of the reasons that hitting two was better in the original situation was that even if White got hit he could easily enter from the bar. Take away that easy entry, and hitting two becomes very wrong. So again, by making adjustments to the situation, we have learned that the strength of our opponent's board is an important consideration in these kinds of situations.

I also looked at making the 5 point and then Black anchors on the 2 point on the next roll. David Rockwell showed me how to hit the "dice" button on Snowie and see what each player's best, worst, and average rolls are for any given situation. So I made the 5 point and gave Black a good roll that anchors on White's 2 point and I found that now Black only wins game 61 percent of the time and gammons about 17 percent. Quite a ways down from the 78 and 59 percent by hitting two. So by doing this, and trying a few other combination of rolls, I learned just how important it is to keep your opponent from anchoring in situations like this.

So from just this one position, I learned how important it is to prevent anchors; I learned that you have to consider the score and position of the cube; I learned that you have to take into account how strong your opponent's board is; and I learned that you have to consider your ability to close the board when you are trying to decide whether to hit two or make points. And I learned various degrees and nuances of these variables by trying out even more slight differences in checker positions and possible rolls by both sides.

Because I have truly learned something, instead of trying to memorize what is right for a single situation, I am much more likely to make correct decisions in the future. I may not remember this exact position—not everyone has Jake Jacobs's memory skills (he can tell you the color of the checkers and what he had for lunch the day the play came up!)—but because I have knowledge I can use and apply over the board, I am more likely to look at the right variables and come up with the best play through logic and reasoning. That's the difference between learning and memorizing.

Now, I am sure some of you reading this are saying to yourself, "This is a lot of work! This takes a lot of time!" To the first part, I can only say that if you think of this as work, you shouldn't do it, or you should change your attitude. To me, improving my game and learning something I didn't know before that I am sure will help me play better in the future and lower my Snowie ratings and win me more matches, money and tournaments, is not work at all. That's fun! That is very exciting!

And to those of you who say this takes a lot of time, I will argue that it actually saves a lot of time. I have thousands of positions with mistakes I have made in the past that I have put into Snowie and studied. I would venture to say that if I had started this true learning process 20 years ago, I would have a tenth of the number of positions to study. I simply would have learned enough sooner so that many of my mistakes would have been avoided. Over the years I have made many of the same errors, or same types of errors, over and over and over. And it's because I never took the time to truly learn from my mistakes.

I hope this article will help you to learn too.

Phil Simborg is a fulltime backgammon player and teacher.
You can contact Phil at: or visit his
web site:

Other articles by Phil Simborg
Other articles on improving your game

Return to:  Backgammon Galore

Other articles by Phil Simborg
Other articles on improving your game

The Backgammon
Learning Center