Michael Skinner opened his account in Division 2 yesterday at the River Mill, St Neots:

Oct 7th: Michael Skinner  11  Neil Foston  0

It was one of those matches where you can’t get a toehold, and most of the games were uncomplicated. But it also featured this:

XGID=-b-B-BD-B−−−aC-aabac−−b-bB:1:-1:-1:00:0:5:0:11:10

Michael is red and leading 5-0 to 11, holding a 2 cube. He’s also put two of my checkers up and is blitzing me, quite horribly, with a three point board and lots and lots of builders. He redoubled, I tried to change the universe by complaining for a while, and then dropped.

While transcribing I saw the computer criticising the redouble, and thought, “Ah, it’s too good.” Imagine my disappointment on discovering that it was a no double by 856 stupids, and a whopping 1,219 stupids to pass! In match winning chances my enormity was 11.35% and I therefore claim the match cube blunder record.

You what?

The take point, assuming you recube immediately to 8, is around 17%. Winning chances for white in this position are around 32%.

What about the gammons?

Yes, there are a lot – 48%. But they are all killed by the recube to 8 so you can ignore them.

Is instantly recubing always right?

Technically no, but only by a tiny tiny amount. In practice if you play this position with XG defending, it will recube immediately when red makes another home board point. And if you play the position with you defending, it is very easy to make a howler with the cube – recubing immediately is the best practical option.

Hang on, how can I possibly have 32% chances?

This is the really surprising bit. Blitz positions typically have high wins and high gammons. I’m accustomed to thinking more builders for the blitzer make it much much worse. They do – but mainly in terms of gammons, rather than wins.

Because you rarely play these positions out, it can be hard to believe how easy it is for the blitzer to get his back men stuck and so for the blitz to stall. In this case the back men haven’t moved and I have a pretty decent blocking structure. More builders also means worse timing so the position is a “deep blitz” that has become more polarized – it can go very right or very wrong. Play it as the blitzer a few dozen times (using the XGID above) and you’ll find it’s both easy for the blitzer to lose, and make mistakes – so for fallible humans the defender may have better than 32% chances.

Should you always recube, whoever you’re playing?

Yes. The chances in practical terms are around 2-1. So in order for it to make sense not to take, you would need your chances from 8-5 up to be less than twice your chances at 7-0 down. That’s never going to be the case.