1-1 in a match to 5. Black to play 22.



Why is it better to make the 11 pt with 13/11(2) 6/4(2), rather than 13/11 8/4 6/4, which leaves a spare builder on the 6pt to make the 5pt, AND doesn't strip the midpiont?

What's so great about owning the 11 pt here?

The difference between the two plays is small, but see it this way, the 11 point gives extra containment when white would want to escape his last checker, and 64 would become a worse move for white in this case. After playing 24x20 he would have to play a vulnerable 6.

I would have made the same error. It seems that your goal from this position is to switch over to an attacking game, and since white's home board doesn't have anything going for it, making a home board point and having another one slotted is better than just making a point. If white does not hit on the next roll, you're a favorite to have the 4-6 points all made, which puts you at a fairly large structural advantage. And if he hits, it's not really a big deal since you've got an anchor and he has a blot. I think it's just one of those spots where the risk/reward ratio is so small that you can play boldly.

Points are better than blots because you can't hit a point.

I guess this doesn't actually address the question.

I don't think stripping the midpoint is such a big deal when you're making the 11 point.

13/11(2) 6/4(2) has any 3, any 6, 21, 22, 42, 51, and 44 to make the 5 point. That's 28 shots.

13/11 8/4 6/4 has any 1, any 3, any 6, 22, 42, and 44, which is 31 shots to make the 5 point.

So you're getting three extra shots for leaving one extra blot (assuming I counted right). That seems like a poor trade.

Yeah, i definitely see why making the 4pt and leaving the blot on the 5pt is better than just making the 5pt-- it's just making the 11pt i'm not too sure about.

Why is the tradeoff bad? I'm already behind in the race, I have an anchor, and he has no board development. It seems going all-out for attacking that back checker makes sense when the only extra hitting number white gets from the blot on the 11pt is 64 (although this becomes a nasty anti-joker, as it is a double hit-- is that 1 shot enough to tip the scales?)

Losing 0.035 equity isn't *super* close, so I must still be missing an important principle here...

I think you have to look ahead more than one roll. For example, you may not be able to clean up the second blot coming in from the bar, and you could end up with four checkers back. It's not some sort of disaster, but you're certainly in worse shape than if you were to have only 2 or 3 checkers back. Among other things, your timing would be terrible for a backgame and you're going to have a tough time running out your back checkers. You can easily get pounded for a gammon loss (white should be shipping the cube if he hits, I think -- and I don't think you should drop it), and that's game over*.

So I think that the extra 3/36 shots that you pick up to make the 5 point (8.3%) is just not worth the loss of equity that can occur if you get 4 checkers sent back.

* One thing you might try to do is run this as just a money game. At 4-away/4-away, doubled gammons are huge and it could be a bit of a match score effect. His position is gammonish, but yours is not. This might be a significant piece of information.

So then as a general principle, leaving more than 1 blot out (whether its under direct fire or not) should be avoided, at least when the gains of doing so are slight? This serves as "downside protection" in the sense that it limits the number of my men that can be sent back if there are a bad sequence of rolls (e.g. he hits on the 4, I dance/fail to cover, he hits again)?

I think saying it that way might lead to a misinterpretation, but that's the right idea. It's correct to leave two blots in many circumstances (especially early game stuff, or situations where your back men are split but not being threatened).

You're always looking at risk/reward when you compare plays like this. If you are risking a lot to gain a little, you're probably doing it wrong.

Yes. You can see this numerically.

Suppose you are deciding between two plays:

Play A: 100% chance of being an 80% favorite

Play B: 75% chance of being a 100% favorite (win outright), 25% chance of being a 0% favorite (lose outright).

Your equity of play A is 80%, and your equity of play B is only 75%, even though you've got a win locked up 3 out of 4 times. Sometimes, there's just no reason to incur the risk because the downside is that bad. (This one is obviously contrived in a way that would never apply to backgammon, but it makes the point.)