California Dreaming

Usually in this blog I try to focus on my own mistakes. It's good for my own learning process and avoids exposing people to unwanted criticism. Occasionally though, I get permission to feature somebody else's errors and here are some from a regular opponent on Fibs. He's ranked at 1755 which will give you some idea of where he stands in the hierarchy.
His problems, like those of so many, revolve around cube use. His checker play is quite good, but he has the worst possible combination of cube errors. He doesn't double early enough, he cashes games that are too good to double and tragically, takes some passes. Our latest match lasted for 13 games, of which he won 7 but lost the match 9-7.
Here's where the points went missing.


Red is on roll, he trails 0-2 to 9 and he leads 108-118 in the race. Should he double? Unquestionably yes. There are two ways of arriving at this. The first will stand repeating yet again, it is Woolsey's Rule. It goes, "Look at the position from your opponent's point of view. if there is any doubt about the take, it must be a double". "Any doubt" is the key phrase here, it means that if you are not 100% certain that it is a take, it must be a double. Actually it is a take, but certainly if doubled here I wouldn't be 100% certain and I certainly didn't want to be cubed.

The second way to look at it is this, an acronym called PRAT. This stands for Position, Race and Threats. Traditionally an advantage in two of these areas is thought to be enough for double and take and an advantage in all three is double and pass. Actually this is a bit too crude and you need to refine this a little and ask yourself how big these advantages really are, but it will do for this position. Red's Position is a bit better as White still has a man behind a four prime, but this is mostly counter-balanced by the fact that White is at least at the edge of the prime and his own strong board will make hitting loose by Red a risky option. The Race favours Red but it isn't overwhelming; a ten pip lead in a 108 pip pure race would be a take. The Threats favour Red, threats being sequences that lose the market, e.g if Red hits and White dances, then White will no longer have a take. We can see here that all the doublets are excellent, as well as 4-3, 4-2 and 3-2, that's 12 numbers straight off. In addition all the above average rolls play safely and there are only a couple of lemons, 2-1 and 3-1. So all in all, strong double for Red, uncomfortable take for White.

This next one isn't quite so obvious, what's the correct cube action here?


Red trails 1-2 to 9, but leads in the race 108 to 132. Woolsey's Rule, are you sure that this is a take for White? If you said, "Yes I'm sure", then you are a bit wide of the mark as this is actually a pass! Positions where you are anchored on the 24 or 23pt and your opponent has escaped all his men and leads in the race are usually passes. Don't be misled by the fact that White holds the 16pt as well. It isn't going to bother Red much on the way home and those checkers aren't much use to White until they move round to complete his board. If you are doubtful about doubling this as Red (or passing as White), set up the position and play it out 50 times. You should find that White will win 23-24% of these, but 11-12 of his losses will be gammons and that's enough to make it a close pass.
Another way of looking at doubling decisions is to adopt the advice contained in Clausewitz's Rules of War; "Always do what your enemy least wants". As Red the last thing that he wants is to be doubled here isn't it? If you evaluate this rather conservatively and think that this might be a double and might not, then Clausewitz has advice for that too. "Faced with two courses of action that are approximately equal in strength, choose that which allows your enemy the opportunity to make a mistake". If you leave the cube in the middle, no chance of a mistake. If you double on the other hand, he may take when he should pass (as here) or pass when he should take. I might well have taken this.
Moreover, once you have doubled you can't then make another doubling error, whereas the enemy may make several. There are many players who hate to double you in with a 4 cube, actually the player playing Red here is one of them. You can double these players in very early indeed, because they don't get the vig associated with owning the cube! All of these guidelines point to being aggressive with the cube.
In my next post we'll look at some more positions from this highly instructive match and discuss the important concept of volatility, but in the meantime, enjoy the game!

More Blundering.

It's a good time to remind ourselves of Uncle Jake's minimal guide to good cube action for the trailing player in short matches. Memorise this. It will save you a lot of thinking time.
"Develop some sort of gammon threat and ship it in".
That's it, short and sweet and as near to infallible as you can get in bg. Here's a position that demonstrates it nicely and where I fell asleep and left the cube in the middle.


Red trails 1-2 in a match to five and trails in the race 137 -130.
This is a very clear double and a fairly easy take. What Uncle Jake means by "some sort" of gammon threat is anything that is better than you start with. About 26% of games played to a finish end in a gammon, so you start the game with a 13% chance of a gammon. Red wins about 63% of the games from here, of which 19 will be gammons, so that will do nicely. Of course it is very hard to tag those sort of numbers onto a position, although some people can do it, but like most people I just rely on my sense of smell and this has a definite whiff of gammon about it! Not only that, it is very volatile and this is a vital ingredient. 10 numbers (4-4, 3-3, 2-2, 1-1, 3-2, 3-1, 2-1) put White on the bar and complete the five prime. Red will also hit loose with all his other ones, twos and threes so it is beyond me to say why I didn't double in the match!
For money a double is an error and if Red was leading in the match it would be a big blunder, but for the trailer it's clearly right. The cube has little value for White and her own gammons are partially wasted by overage.

Well, I hit loose and White hit back from the bar, but the dice god gave me a 3-3, almost the best roll possible. How would you play this?


The first three are pretty easy, bar/22, 8/5*(2) and I chose 11/8 for the last one, getting another builder into the zone and slotting the back of the prime. This is a blunder. The checker on the 8pt is vulnerable to 6-2 and 5-3 from the bar and I just don't need that risk, partly because if not hit I have a great cube next turn and partly because my game plan has to be blitzing against the lone man back rather than priming. The clear best play with the last 3 is 7/4, still with three attacking checkers, as Red will break the bar to hit, plus double coverage of the outfield for White's 6-2 and 5-3 from the bar. She should step out with either of these, so note that these rolls have been converted from excellent to very bad if you play 7/4 instead of 11/8. This swing accounts for almost all of the difference between the two plays.
That was a tough play to find I think, so apart from noting that doublets often require a little more thinking time than singletons, let's move on to my third blunder of this game.

By this time I've managed to double White in and have a 3-2 to play in this very strong position. Any ideas?



The first thing to say is that 6/1* is very horrid. Red will hit White loose if she gets up to the 23 pt but no point in hitting on the ace until you have some cover numbers and can go for a close out. I sort of fiddled around at the back without any clear idea in mind and played 24/21, 22/20.
You wouldn't think that you could make a big mistake here, but this play is a blunder.
Two things to think about here. White has what is called a brittle structure, with stacks on the mid and 6pts and three checkers buried on her deuce point. She isn't going to develop this into a prime but if she can point on a Red blot or two she has a chance to get back into the game. Red has to counter that by anchoring 24/22. You will see this time and again. Counter brittle structure with a rock against which it will break. So, that leaves a 3 to play and the best 3 is 13/10, not 6/3. Why leave a shot? To gain a double shot if White throws her magic 6-1, to increase the chances of making the 8pt and to give more hitting numbers if White moves up to the 23pt. As this play also duplicates the 1 that White needs to get to the edge of the prime you it's a bit of a no-brainer, but only if you have engaged your brain in the first place. Before making my nothing play I should have asked myself the pro question, "How do I plan to win from here?" It's a good idea to mutter this under your breath from time to time.
Well, that's about it for today. Thanks for reading, let me know your thoughts on these, or indeed any positions and until the next time, enjoy the game!

Put 'Em Where You Want 'Em.

Backgammon is often at its best in tough midgame positional struggles. Here are three interesting positions that a strong player got wrong. See if you can do better.

Red leads 3-0 to 5, but trails 147 - 129 in this game.

Although Red trails in the race his stronger position is enough to make him a small favourite. He has the golden anchor and all his checkers are in play, whereas White has already buried four checkers and so has to play the rest of the game with 11 checkers. In the match, Red chose 10-4 but 20/15, 7/6 is much better even though it leaves a shot.

The theme to keep in mind for this sort of game where you have the golden anchor and you are fighting to prime White’s straggler is this. Ignore shots and put your checkers where you want them. Your anchor (and White’s crappy board) mean that you can leave shots without a serious risk of a bad sequence that gets you cubed. Of course you don’t want to take risks without some gain, but the gains after 20/15, 7/6 are substantial. You free up the spare from the anchor (almost always a useless place for a spare) and you keep two spares aimed at the 5pt.

The next position shows Red making the same sort of mistake, although I grant you that this is a tough play to find.

Red played 7/3, 7/6 giving up his very valuable bar point, but “putting your checkers where you want them” also means leaving your checkers where you want them sometimes and here, with White’s blot inboard, it’s worth keeping the bar and leaving the double shot after 15/11, 15/14. Note that this is better than 15/10, one less shot and duplicates fives and two’s to hit and cover.

And here, in a different shape is the same mistake again.

No need to labour the point. You can see for yourself that 15/14, 3/1, 2/1 is easily the best play as White now has two blots inboard.

The 20pt anchor is a great insurance policy and when your opponent has a weak board as well you can pretty much ignore shots and put your checkers where they will do most good. In these tough positional struggles it's a great guide to finding the right play.

Until the next time, enjoy the game!

The Cube's The Thing.

It's so important to get your cube action right. Most games are won by luck anyway, regardless of the checker play skills, so getting the cube to the right level (high in winning games, low in losing ones) is fundamental to success in match and money play.
Expert players carry a library of reference positions in their heads, positions where they know the equity and can compare it to what they face over the board. Here are two quite differing positions, the first commonplace, the second less so, but still essential for success.

Here's the first and it's 0-0 in a match to 7. White is on roll, leading 114-139. What's the correct cube action for both sides?


White leads in the race by 25 pips, she has escaped both her back checkers and if she clears her midpoint next turn she will lose her market, so the double is absolutely clear. Should Red take? If it was just a race Red would have a clear pass, but with his anchor he stands a good chance of getting a shot during the bear-in. Is this enough to give him a take?
I tend to take a few too many of these and I should have passed this one. White has 11 rolls that clear the midpoint immediately and also has a spare 6 and two spare 5s with which she can fiddle around while waiting for a clearing number. My side of the board is important as well. I already have the 5pt (essential!) and I may be able to build a nice board ready for a shot, but may be isn't will. I have to know that a shot will win for me.

Good reference positions are those where the take/pass decision is very close, so that when you come to compare it with what you have on the board, you can say, "Any improvement and I can take, any weakening and I should pass". The above position isn't a good reference, as it is a very clear pass and the take loses 1.2 points on average. The position below is much more useful. The race deficit is still 25 pips in a shorter race (White 100-Red 125) and the take/pass decision is very close indeed.


Although White's race is a bit better, because the race is shorter, her position has weakened in several respects. Now she only has 7 numbers that clear the 13pt immediately and with only one spare 6 and one spare 5 she has less time to wait for a clearing number. Red has improved too, with a solid board already in place and the bar pt slotted. I think this one too is probably a pass, but it's very close, certainly close enough to be a reference.

This next position comes from the same match and I am on the bar trailing 2-3 to 7.



You have to look at these with a view to doubling, even when on the bar against a four (and potentially five) point board. Although White only has 13 shots all the ones play well too. Even dancing isn't fatal yet. White doesn't always cover straight away and has to dig her back men out pretty quickly even when she can make a 5pt board. Giving the cube away in these two way gammonish positions doesn't hurt too much as White can't redouble without putting the the match on the line. This is very much a trailer's double. You shouldn't try this one when level or leading, but any sort of deficit makes for a strong double here. In the match I didn't double ( a large error) and eventually lost an undoubled gammon!
As a reference position it's fairly useful as a minimum double position. Red could actually be a bit weaker than this and still have a correct cube, while at some scores White should even pass. If Red owned the cube it would be a huge redouble of course and a pass.

Well, I hope you find these useful. I'm sure that you'll see something like these every week and in the case of the 20pt anchor positions, every day! Until the next time, enjoy the game.

Two Blots Inboard, How Bad Is That?

Here's a nice checker play problem for you, which I got wrong. It's interesting not only because it's a tough play, but also because it illustrates a much wider principle. How would you play 6-4 here? Red trails 0-3 to 5 in the match, but in this game he has already cubed and leads 125 - 138 before the roll. White is the world class player and noted theorist Rick Janowski.

Three plays stand out, the double hit, 9/3*, 8/4*, the single hit 9/3*, 7/3* and the non-hit 9/4, 8/2. Nothing else is worth a look. My thinking went like this. "I lead in the race and my position is solid, apart from the fact that my anchor is detached from the rest of my checkers. Moreover, White is completely stripped in the outfield, so will find it hard to make a damaging play next turn. I don't need to give him 22 shots to possibly turn the game around when he owns his 6, 5 and 4pts." I chose 9/4, 8/2.
Actually, this is probably the best play to win the game, but it doesn't win as many gammons as the hit plays and gammons are very valuable here. I can't lose one and winning one will get me to Crawford. Of the hit plays, the double hit is clearly best in the rollouts, illustrating a valuable guide to good play. Two on the bar is better by far. It's better here because of the match score (dead cube, one way gammon threat), even though it leaves two blots and White has a strong board. My play was a large blunder costing about 0.17ppg
Another very interesting aspect of this is that if the cube were centred, the non hit is clearly correct! Why is this? It's because after the quiet play, I can usually offer an optimal cube next turn, so the hit plays with their attendant risks aren't efficient. When they work and White can't hit from the bar, I lose my market by the length of the street. When they don't work and White hits, I haven't got a cube at all.
For money, the quiet play is much more competitive, although still an error and the single hit is on a par with the double hit. This of course is because gammon losses come into play.

The next position also illustrates the strength of putting two men on the roof.


It's a position from earlier in the match. White leads 2-0 to 5 and forced plays have led to this position. White doubles from the bar. I expect that everybody will drop this like a hot brick and I certainly did. However, Snowie 3-ply will take this! Rollouts suggest that it is actually a pass, but not by much, demonstrating how strong it is to put two men up, even when you are trapped behind a prime and have two blots inboard. Here too of course the score is a factor, as White's gammons are less valuable than usual because of overage and cube ownership is also worth a lot. The pass is clearer at "normal" scores, although still not huge.
I recommend that you play this out for a few games if you find this hard to believe, you can learn a lot. Your first problem might well be how to play your next two! Any ideas? I'll give you the answer in the next post. Until then, enjoy the game!

Re-learning Old Lessons.

Today's position comes from a nine-point Fibs League match against Walfinho from Austria. I've already doubled on the strength of my four prime but White has anchored and is now threatening to start building a prime of his own. I have 4-2 to play. What would you do?

Red (dorbel) v. White (Walfinho), Game one, Match to 9 points. Pipcount Red 144-White 156.

My prime is front-loaded, which if you haven't come across the term means that all the spares are down the low end and of course this makes for lack of flexibility. 8/4, 6/4 makes the 4pt nicely and leaves the 8pt slotted. It leaves a shot of course, but White isn't going to hit even if he does roll a 5. This was enough for me, the play makes a valuable point, improves the spread of my builders and free slots the back of a five prime, that'll do. It's an enormous blunder.
My main priority in this position should be to get my back checkers moving, before White builds a nice prime to keep them in. I should know this without even giving it much thought, because there is a hugely powerful guide to the correct checker play in these situations and it is this.

When your opponent has a high anchor and you don't, you must respond immediately to try and equalise.

That even applies here, where White's anchor is the feeble 22pt and it is primed. Red just has to get those back checkers moving, so that they can race or anchor or hit something in the outfield as the dice dictate. Otherwise the advantage will swing slowly to White as she builds her prime and stays anchored at the edge of yours. The best play is 24/18. White will usually hit it of course, but you should get lots of returns and while she is hitting, she isn't improving her board.
I am indebted for the above rule to Paul Lamford, the British writer and teacher. I had forgotten it and needed reminding of how strong it is.
Look for Paul Lamford's books, they are very useful. 100 Backgammon Puzzles is an excellent cheap read, full of useful examples like this one and his Starting Out In Backgammon is a great beginner guide.
Until the next time, enjoy the game!

Return of the Blunder Log.

Well, the Dorbel Daily has been inactive for too long. Sometimes your brain needs a rest from analysing and studying. This is a GOOD TIP actually. Don't read about backgammon or do any studying in the run-up to a tournament. It is a well known phenomenon that learning new things has an initial bad effect on your game. This is because you will start applying them in inappropriate places. Only when you have fully grasped a new concept will it be of benefit. Think of it as remodelling a golf swing. It will pay off in the long run, but you don't want to start on it a fortnight before the Open.
So, I am going to try and post daily for a bit and these posts will be a bit of a grab bag, of blunders made by me in online play. I won't be posting those that are just oversights, partly to spare my blushes and partly because they aren't very interesting. The "Oh God, I didn't even see that I could have made a 6 prime" sort of thing isn't useful.
Here's a position from a Gamesgrid match with Chiemi. It's the first game of a 7 point match, I've been cubed and I have a 5-3 to play. What would you do?


I chose to make the board, 7/4, 6/1 and wait for a shot. I reasoned that hitting a shot would certainly save the gammon and that I might even win, but his level of thinking is much too wooly. First I should have looked at the exact number of White rolls that leave a shot next turn, which is of course all 2s except 2-2. That's 10/36 after which I hit 11/36 or 110/1296 or about 8%. White will of course take 2 off with any 2, so have 10 men off in total. So, I'll be on roll with White on the bar and holding the cube in something like the position below.


In this position I win the game just about 30% of the time holding the cube. If that seems a bit low to you, the two spares on the 3pt are the problem, for two reasons. First they give me very little choice in how I bear in the last checker and I may even have to leave a shot. The second is that I badly need to bear off as many checkers as possible before breaking my board and of course for that purpose it would be much better if they were on the high points. Move them to the 6 and 5 pts and I go up to about 36%.
So, 8% to get a shot and hit it multiplied by 30% wins after hitting comes out to about 2.4% wins if I wait. We can add a little bit for getting and hitting a shot later as well and in fact the rollout says I win 3.2% if I wait. If I had made the effort I could have come up with a figure of about 3% over the table, but I didn't which was partly laziness and partly because like everybody else I play too fast on the net. Now let's look at how many gammons I lose if I wait. None if I hit of course, but lots if I don't! When I don't get a shot and hit it, White will have somewhere between 9 and 12 checkers off and I'll be anywhere between 0% and about 70% to lose a gammon, so 35% gammons is a reasonable guess. In fact the rollout says 36%.
What if I just run the 5-3? I would have guessed 15% gammons over the board and Snowie says 14.3%, and of course I never win. So by staying I win about 3% but lose 36% gammons instead of 14%! It's not even remotely close. I should have nailed the number of my wins exactly and even if I had been way off in my estimates of gammons, I would have come up with the right answer, which is that staying is a huge blunder.
So there you go. Trying to win the game is usually good, but there comes a time to turn mother's picture to the wall and ride out! This was one of them.
Until tomorow, enjoy the game!