So, I've decided to try and finally learn backgammon. A few years ago I purchased Backgammon for Winners, Backgammon for Serious Plays and 501 Positions. I read through Winners and Serious players and about 100 positions of 501 before giving up on the game due to a complete inability to find anyone who wanted to play. Once again I've decided to pursue it for the intellectual delight.


I wasn't at all sure how to play this roll. I consider myself a beginner so I would appreciate some well thought out advice as to how to analyze this spot.

Much later, from the same game:


Should I have redoubled? Why or why not?

26 from the bar is a chronic nuisance. With the pip lead, I would have played safe too, 8/2.

In the later position I would definitely redouble and expect a pass. One on the bar, 2, 3, and 4 all hit, 1 and 2 cover ... it's a real mess for white. The too good to double rollout doesn't surprise me, but I miss that often in play.

Welcome to the game!

Maybe better to not include the computer analysis with the position to get less biased responses. You can (and should) post it after you get some feedback.

The first position I would have gotten right, though I would have felt like I was guessing. My train of thought would be something like this -- For a few more shots on my side of the board, I get to do two good things - keep the builders on the 8 and 7 points in the right spots and start a very valuable point. That's probably worth the extra risk, given that White's position is in disarray.

For the second position -- Being unsure about whether to double isn't a good reason to not double. You should think of not doubling as an action just like doubling is an action. Either one can cost you equity, there's no real reason to prefer not doubling when unsure. In fact, you can only force an error from your opponent when you double, so maybe doubling should be preferred! Not doubling when unsure is a common stance among players though.

The position does look clearly too good to me. You have 1s and 2s to cover and 2s, 3s, 4s, and 5s to hit. So it's tough for things to go wrong right away and you clearly win many gammons from here. You can play on and if things start going badly you can double then. Notice if you hit and then get hit back with say a number like 5-1 that enters one checker and dances with the other, you still have a very strong position that White may not be able to take. So it's like there's limited downside to playing on for the gammon.

Peachpie and Z,

For position 1 I definitely considered B/23 11/5, but I did the math and saw that I was getting hit more than 55% of the time so I opted for the safer play with the pip lead in hand. That was basically how I eliminated 11/5 - it just seemed unnecessarily aggressive in a position I had under control.

As for the double - I doubled next roll and he passed. Doubling is something I am even weaker with than my checker play.

I'm rereading Serious Players atm and supplementing it with 501. So far I'm having fun again.

Yeah, 20/36 (~55%) is correct. But you get hit 13/36 (~36%), after 8/2 anyways, so really, you're just paying 19% extra for two pretty big positional gains. Turns out it's worth it. If White's homeboard was a bit stronger, it wouldn't be.

Reg. the first position.

11/5 seems clear to me.

Opp has 3 checkers back. With 11/5 You keep all checkers alive for the priming and can cover 1 or 2 points most of the time next roll with all those very pretty positioned spares. It's quite important to fill the gap between Your blockade and opp's anchor and it's a good time to try.

If he misses You have a shot at double-cover or single-cover + slot keep (both very strong plays), if he hits one of slotters You can enter&cover the second one, if he misses but tempo-hit Your back checker You can again enter and cover one (or maybe even hit + cover). All those are strong scenarios.

This play pays now, when it's relatively safe to do it.

8/2 is safer now, but with that much contact left and opp's weak board it's not the main consideration to me. It has the following disadvantages:
- slots the wrong point (behind opp's anchor)
- strips the 8pt., which will make it much harder to build 5&4 pts. later without a lucky small doublet
- exposes one more blot, which is never a good thing with a weaker board

In Position 1, you're focusing exclusively on what can go wrong if your opponent hits you, and you're ignoring what can go right if your opponent doesn't hit you. (Or what can go right even if he does hit you!)

This is a very common (maybe the most common) error among players starting to climb the ladder in backgammon. It leads to a super-safe style of play where you end up burning checkers (as with 8/2 in this position) and failing to make good blocking points. Here's an analogy to poker: how well would you play if you always assumed your opponent had the nuts when he bet?

Position 2 is a little more of the same, this time with a doubling decision. Instead of thinking about how your opponent could pull the game out if you roll your worst, just focus on what will happen most of the time: you'll hit a lot of blots and gammon him. That will help you decide that the position must be a monster pass, and then you can start thinking about whether you're too good or not.

One last point: everybody goes through these stages as they move on from being a beginner. Don't worry about whether this stuff seems hard or whether you're picking it up quickly enough. It is hard; backgammon's a tough game. Just keep plugging away and asking questions and the game will unfold.

Can you post the actual numbers for the cube decision in the second position? If it is reasonably close, double is probably ok against humans: the occasional wrong take will compensate for the objective loss of equity. The same goes for doubling a little too early, you will sometimes gain from a wrong pass.

Peachpie,

Cube equities in second position at XG++:
No redouble: 1.37
Redouble / Take: 2.408
Redouble / Pass: 1.00

Too good to redouble / pass.

Everyone,

Thank you for the advice. I will continue to post positions as they come up. My checker play is simply terrible.

After 19 games:
Overall Level: 16.09
Check Play: 17.38
Cube Play: 10.91

Thanks Mark.

So, redouble is -0.37. That's pretty big. But a take is worth +1.04. So if I get a take one time in four, I almost break even. I think against some human opponents, I would still double.

I'd like to play with those human opponents who would take 1 time out of 4.

I know it sounds silly but I get some really bad takes on safe harbor.

I was looking for a place to play online and I couldn't get safe harbor to work with chrome browser. What browser do you use?

Chrome / Win7. There is a small install.

Don't know if it works on other platforms.

It's a fairly good place to play. People are mostly friendly, the interface is good enough, and there are usually enough players on (in the "cove" room). Be sure to only use rooms marked with yellow dice. Red dice rooms use a bogus dice roll generator.

Ahh... Yes, i got it to work now. That was weird. I was not expecting to have to download something. Why are there bogus dice rooms? That's crazy.

I was going to join FIBS, but I'm waiting to get approved (also a weird setup).

I'm beginning to think the backgammon community is a bit nuts.

It's an interesting story. When SHG first opened, their dice roll algorithm was (unintentionally) flawed, resulting in significantly fewer doubles - about half the expectation. Eventually a couple of clever users figured this out, and demonstrated it statistically beyond a doubt. To their credit, SHG promptly accepted this and implemented a corrected algorithm which has seemed to work properly ever since.

However, some social players had gotten used to the bogus dice with fewer doubles. They felt that the "new" dice were too wild and didn't like all the extra doubles. So they threatened to leave the site in mass unless they got to keep their phony dice. SHG decided to accommodate them, marking each room with either bogus (red) or true (yellow) dice. So there you are.

If you see me on, feel free to say hi, I use the same username.