Problem of the Week #37: November 15


Cash game. Black owns the cube. Black on move.




Black to play 2-2.

The obvious answer looks like it is bar/23, 11/5**
It looks like anchoring on the 22 point would be useful but after hitting white twice, black should be able to make an anchor or at least hit that blot on the next roll.

Might not be the brightest move, but I'm tempted by the triple hit: B/23 24/22* 11/9*/7*. It leaves Black with 5 blots (sigh!), including one in his home board (re sigh!!), but with 3 men on the bar to White, that let Black some time for: 1) build an anchor (ideally the 4-pt, else the 3-pt); 2) make his 5-pt; 3) make his bar point. If Black can do all that, he just evened the game. Of course, Black could get hit on the 5-pt immediately, which would complicate matters, but it would be far from over...

What about B/23 11/9*/7*/5 ? One less hit, but making the 5-pt instead. 2 blots instead of 5. Hmm, appealing. But it gives Black less chance for a good anchor.

I must be a degen gambler, because I stick with the kamikaze move:
B/23 24/22* 11/9*/7*

Well I start with b/23 24/22*. No matter how many blots you can sweep up otherwise, you can't take the chance that something goes wrong and White makes his 3-point, because then you are cooked.

Then I'm not ready to give up on the forward game yet, so I anchor with another 24/22, then hit with 11/9*. At that point the game can develop in many different ways but I have decent equity pretty much no matter what happens. Maybe I can build a bit of a forward position and escape a runner or two. Maybe a back game develops. Maybe a holding game develops. I can handle any of those.

I don't even really ponder b/23 11/5** very much at all because despite how pretty it looks there is just no real winning plan at that point. You are caught between game plans without enough timing for a deep back game or ammunition for a forward game.

b/23 24/22(2)* 11/9*

As much as I want to advance at least one checker up to the 22 point, I think I really want to make the 5 point more. I think that with 1s, 2s, and 3s advancing my back checkers so they can see daylight, I've still got enough time with two of white's checkers on the bar that I don't need to do it just yet.

IDK I really don't think the 5-point is that big a deal in this position (heresy, I know) because Black can't really expect to win going forward yet, so there's no particular rush.

If white enters both and Black whiffs the shot at the 22-point checker things can get ugly real fast.

sry,but this is really a funny idea.

B/23 24/22* 11/9*/7*

Whether black is trying to win going forward or not, he wants to make the 5 point, so I don't think this argument really carries much weight. Furthermore, advancing to the 22 point is actually playing to win moving forward because you've advanced the anchor to where escape is possible.

If things are about to get ugly, I'd rather have the 5 point made rather than unmade. Remember that by not making it, you're also giving white 11 shots to send that checker back. That could also get ugly.

I think I like the 5 point because it's essentially a permanent asset. Unless you crunch, you're going to have that point made until you bear off.

As I said in my original post, the point of making the anchor is so that you CAN move forward if the game develops that way, using the anchor as a jumping-off point. Or you can play a back game. Or a holding game.

Meanwhile you are almost never going to get steamrolled.

If you make the 5-point, that leaves you with 3 on the Ace-point and 1 on the Deuce-point, which isn't much of anything really. Sure you MIGHT be able to get some air next roll. Or you might not. Or you might get attacked when you try to do so. Meanwhile each roll presents you with a choice between continuing forward with limited prospects, or trying to put together some kind of defensive position.

Anyway, you never get these problems wrong so I'll stop there. But I at least wanted to get my flawed thinking on the record.

The real reason I'm making the 5 point because it's the 5 point. By arguing, I'm really just trying to make myself feel better about it. I'm sort of reverse-engineering the reasoning on this one...

Many times when I play against GNU and manage to argue myself out of making the 5 point, it turns out to be wrong (and sometimes very wrong). This is especially true when GNU has checkers floating around back there.

So I'm really clinging on tight to this rule of thumb here because I completely agree that the stack on the 24 point is UGLY.

It's true that making the 5-pt will be a permanent asset and this cannot be taken lightly. I mean, you're making the 5-pt AND hitting twice, it can't be that bad, right? It might well be the best move.

What I liked about the triple hit is, although things get complicated if White hits on the 5-pt (or even enters right away with a double), it has a lot of upside if White misses on the 5-pt (even more if he rolls a 6 and can enter only one). We can then make the 5-pt with any 1, 2, 3 or 6, or even 4-4. 5-5 and 5-4 would be bad rolls, but there aren't many of those.

If you even considered some form of the hara-kiri play involving 6/4 (as I did), it's a sure sign you've been reading too many books by Barclay Cooke (as I have)! Although Black trails by a ton (190 to 122), and sooner or later may find himself playing a backgame, it's way too early to commit to one now. For one thing, he has only one anchor point. Moreover, this 22 gives him a fair chance to win going forward.

This is one of those positions where Black does best by preserving both forward and backward options. He should reject any play that doesn't put at least two White checkers up on the roof.

• Play A, Make the Five Point: Bar/23, 11/9*/7*/5
  Black's safest move leaves White with only a few doubles to regain any balance, but also leaves a pile of checkers buried on the twenty-four point. During the next couple of turns, Black figures to be fighting for front position, and may not get back to making an advanced anchor until it's too late. This play also leaves an awkward distribution of builders up front, with the blot on the eleven point bearing only on outfield points.
  
  
• Play B, Make an Advanced Anchor: Bar/23, 24/22*(2), 11/9*
  Making the twenty-two point gives Black better balance in the back board, but leaves White any five, along with doubles, to swing the momentum his way. The twenty-two point works well for Black in both forward and backward options. In forward games, it gives him a launching pad to leap over White's block, and in any backgame Black might play from here, it gives him one of his best anchors.
  
  
• Play C, Triple Hit: Bar/23, 24/22*, 11/9*/7*
  If White misses the return shots, the triple hit gives Black his strongest attacking options. His outfield checkers are nicely diversified, and he's got a man in place to escape with any six. If his attack fails, Black may still be able to anchor up and form some sort of backgame.
  
  
• Play D, Make the Nine Point: Bar/23, 24/22*, 11/9*(2)
  This play gives Black the double hit, but is probably weaker than the other alternatives. If Black wants to make a point somewhere, either the anchor on the twenty-two point, or the five point in his inner board is a better choice.

If not for the stack back on the twenty-four point, Play A would be Black's best. Of the others, Play B gives Black a balanced, two-way game plan more than Play C or D.

My solution: Bar/23, 24/22*(2), 11/9*.

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. 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 55% correct.
Correct: 28(a), 29, 30, 32, 35, 36
Incorrect: 28(b), 31, 33(a), 33(b), 34

I'm ranking these plays in the following best-to-worst order:

• Play B - Make an Advanced Anchor
• Play C - Triple Hit
• Play A - Make the Five Point
• Play D - Make the Nine Point

I'm with Pineapple888 on one thing today: I always hate it when Aaron W. comes out on the other side on these problems! It's usually a bad sign for me.

Black is in horrible shape but just rolled an uberjoker. So what is going on now?

Breaking a point isn't really the right idea most of the time but after B/23 24/22*(2) 23/21 Black hasn't really gained anything. Without any other good 2s to play Black might as well hit two. There doesn't seem to be any immediate need to anchor.

After the triple hit what has been gained? White will usually enter two checkers and Black will hope to make his 5 point if White hasn't rolled a 5, at which point he will have a somewhat improved prime and timing. Black loses a ton of ground if White rolls a 5 though. He definitely gains if White rolls a 6.

After making the 5 point what usually happens? White usually comes in with one and relatively often comes in with two. Black's aces are somewhat duped but most rolls improve his position as 20 shots hit and 33 & 44 aren't too bad.

I'd just make the 5 point. Getting hit with a 5 just loses way too much ground.

I just make the 5 point...

I run real good so I am going to pick up the blot on the 22 on the next roll anyway.

I think it's highly debatable whether GNU values backgames correctly. I wouldn't trust either its recommendations or its rollouts in this type of position.

I'd be shocked if it liked anything besides making the 5-point. That doesn't mean doing so is correct.

Of course, it doesn't mean I'm correct either. So who knows.

This isn't exactly a "back game".

I'll back GNU evals or ROs against any human in this position. Snowie or XG ROs are comparable I'm sure.