Problem of the Week #90: January 2


Cash game, center cube.





Black to play 4-2.


Note: All ‘cash game’ problems assume the Jacoby Rule is in effect. That is, you can’t win a gammon unless the cube has been turned.

24/18.

Make an advanced anchor and duplicate fours for white.

18/16 and 13/9.

yes, 24/18

it's a real giveaway when you look for other candidate moves and can't find even one

The main problem is for me the slotted golden point. Attacking it is just to loose. Other plays let white the freedom, to build the 5-point or the bar. Then our best way to stay in the game is also grab an advanced anchor and go into a mutual holding game. So why not grab it now, making the bar? Makes a more connected position, which will be connected even if we are hit and cleans up to one blot. We passed the buck to white, who is now under pressure to equalize the position.

24/18.

after 13/9 there are a number of 2s you can play. 18/16 is the most useful of those as it dups 4s but still nowhere near as useful as having white's bar.

24/18

Maybe this is a response to leveling the problem last week. It's the most useful point in the short-medium term and it cleans up 2 of the 3 blots. Or maybe I'm missing something super-deep, but I'm just not seeing it in 8/4 6/4, any 9/5*, or any 13/9.

It is so embarrassing! In Problem 90, it’s only the second turn, and I am already completely lost.

Of course, I can play the opening. Splitting with 6-4 is relatively automatic. White's reply, however, is another matter. The major split has certainly worked out this time, but a marginally better play would have been the reverse split. See the rollouts at http://bgonline.org/, or download Nack Ballard's consise summary of replies to the opening move.

Now Black must play this 4-2. Here are a few of the options my intuition causes me to reject.

• Hit: 9/5*, 13/11 -- In the wild 1980s, I might have entered a blot-hitting contest with this or one of the following moves.
• Hit: 9/5*, 18/16 -- This move takes away White's sixes from the bar.
• Make the 4 point: 8/4, 6/4 – The four point will be irrelevant if White makes the five.

And here are the choices I like:

• Make the 9 point: 13/9, 24/22 -- Making the nine point protects Black’s most advanced blot. If he is hit on the 18 point, he’ll lose only 7 pips, and he will also have plenty of return shots.
• Make the 9 point: 13/9, 18/16 -- This move leaves fewer shots for White.
• Make the 18 point: 24/18 -- This move duplicates White’s fours.

As I said above, I don’t know what play is best, but since I have to choose something, I'll follow my gut, and make the anchor. This also cleans up two of Black's three blots, and leaves the fewest number of shots for White.

My solution: 24/18.

For the Record
I am so often wrong that I like to post my record in these messages. It's kind of a truth-in-advertising thing. Grunch: I have been answering these problems without the use of a bot, and before checking the excellent solutions of others, since Problem 28. My record at this writing is 50%.

After this roll Black's ahead 9 pips in the race (assuming I counted correctly) Probably best to grab the bar point although I'm tempted to grab the 4 as well...love those inner board points. 18/16 13/9 seems like a waste of a good roll. One other play I may make although probably wrong is 13/11 9/5* putting black in the air and giving me 1, 3 and 6's to cover or hit back with. I might play that against a timid opponent but most likely the bar is best.

24/18

Lets first go back to problem #6 there we learned the principle: structure supports blots.
That principle is useful here and leads to 24/18. Furthermore white would like to use a 4 for making his 5-point rather than hitting loose.

Problem #22 also gives us some criteria for chosing between hitting and making a point.
1. Strong inner board, no, dont hit
2. Good point to make, yes, dont hit
3. Make a point while unstacking, no, hit
4. Break anchor to hit, no, hit
5. Vulnerable, exposed blot, yes, hit

Here it seems to be 3 vs 2 for the hit 9x5, criterium 5. however is somewhat double-edged.
Hitting is more vulnerable than not hitting, so it might be a “dont hit”as well.

Problem #52 also gives some extra criteria:

1. A big disadvantage in the race favors hitting: neutral
2. Gaining a lot of ground in the race favors hitting, no, dont hit
3. Strong threats favors defense, yes, dont hit make bar-point.

These extra criteria argue for a dont hit.

My play is 24/18.

Grunchaments....

1.Run and make bar (24-18)
2.Cover 9 and slot the 4 (13-9,6-4)
3.Hit inside (9-5*,something)
4.Make the 4 point.

I think running is too static this early in the game....just hit inside or make a point. I think it is too early to anchor up out there.

I like 9-5* and something. Puts a guy on the bar, slots the 5 , takes tempos etc.

damn....what nits in here....really?

I like those early game problems.

I narrowed down to 4 plays, in no particular order:

A) 24/18 --> making the defensive bar point
B) 13/9 24/22 --> making the 9-pt
C) 8/4 6/4 --> making the 4-pt
D) 9/5* 24/22 --> hitting loose on the 5-pt

I discarded 13/9 13/11 because I thought it was an inferior variation of B.

- I think I now discard B because the 9-pt isn't that useful if White makes our 5-pt.
- Choice A makes a good anchor but seems otherwise too passive.
- C makes a nice inner point and duplicates 4s (hit or make the golden pt for White).

The question is, how important is it in the early game to prevent White from making our 5-pt? I think it's important enough to fight for it. White will almost surely hit back (I think only 6-6 doesn't hit back!), but is it better if we don't hit?

While I'm thinking, I just thought about another play that wasn't even in my initial candidates: 9/5* 18/16. Sure we won't have a direct return shot if hit on the 16-pt, but it duplicates the 4s and negates all his hitting 6s. That's my pick.

9/5* 18/16

I felt the same. I think that the plays right after the opening moves (let's say moves #3-4 up to #8-10) are often not trivial at all. We can make errors and it will impact the remaining of the game, so it's tough sometimes.

I see 2 possibilities:

a) 9/5* 13/11
b) 24/18

In this situation, it doesn't seem like it is so important to make the anchor on the 18 point. I believe it pays to be aggressive and hit on the 5 point. If black can cover on the 5 point, he should be in fairly good shape.

So would you write this position in nactation: 64S-43S-42 ?

I like 9/5* 13/11. If you hit the 5 you gotta go 13/11.

You might want to explain what "nactation" is, since I suspect most readers here have never heard of it.

Fairly sure that is right...

I've seen time and time again anchoring on the bar once you've slotted it is almost always right. I'll go with it here.

24/18

Im gonna do 9 - 5 and 13 - 11 again the cube is middling,
so i don't mind trying a little action play

Yeah! You got it. Nactation is the notation developed by Nack Ballard to describe early moves. You can get a summary of its usage in the citation I gave above:

Nack Ballard's concise summary of replies to the opening move.

Here's the more complete description found on Nack's web site:

http://www.nackbg.com/nactation.pdf

13/9 13/11

Nack Ballard, one of the "Hall Of Fame" players, introduced a notation for backgammon, which is not only a short notation like in chess, but also a a description of the action. For example, 43S stands for split a runner to the golden point and the 3 down from the midpoint. If you would play the reverse split, playing the 3 with your runner, it would be 43Z (a Z is a reversed S, isn't it funny). I'm trying by now to get familiar with it, because it started not ready and with a fixed set of rules to "nactate". It is more an organic way, and there are still possibilities to develop it and improvement. You can look here: http://www.nackbg.com/nactation.pdf
At time, if you look at bgonline.org, you can see the backgammon opening transforming into chess mode, with cranked up sample size and 4ply rollouts.
Nack Ballard and Paul Weaver started this with "Backgammon Openings. Book A". If you have a new and fast computer, you can join for rolling out moves. Because XG supports up to 64 cores and uses every gran of CPU power, it is for this purposes now state of the art.
At least, if you are thinking, this is an insane development, look at http://www.thebackgammonlearningcent...-a-true-giant/
"Ed was highly rated on the KG Rating list for 10 years and even reached the No. 1 spot in 1996. Ed stopped competing in 1995 for a few years. During that time, he spent hours every day playing against the bots and studying the game. He returned to tournament play in 2000 and began cashing in several events. "
Ed O'Laughlin said in this interview: "I set up challenging positions, like back games and prime vs. prime, and play both sides until I am extremely confident of the cube and checker strategy for both sides. I develop reference positions, roll them out extensively, save them, and study them. I believe reference positions are critical, as what looks to be “logical” over the board is often wrong! For example, one pretty much accepted strategy is to double a back game defender when you threaten to clear the third point in front of them. Through extensive study and rollouts I found this is not necessarily true, and I have different rules to apply at different scores and different pip counts and configurations. I don’t think anyone can figure these things out intuitively over the board…this knowledge can only come from experience and study." If you play casual, it might look weired, but nowadays it is state of the art for top notch backgammon. Even if you don' want to make money, this approach is a better waste of time then sitting in couchpotato mode, watching tv and eating fast food :-).

9/5* 13/11
bloody morning

Can't believe I didn't even see that move, chosen by several readers. Looks like a worthy candidate. Possibly the best of the hitting choices.

Make the bar then go from there. 24/18