[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

Back Game Quiz, Seminar in Chicago, Memorial Day 2013

Posted By: Stick
Date: Monday, 23 December 2013, at 2:55 p.m.

In Response To: Back Game Quiz, Seminar in Chicago, Memorial Day 2013 (christian munk-christensen)

Paul asked me about rolling these out on higher settings forever ago when he emailed the problems to me before the seminar. Here are my answers at the time to that question and the problems fwiw. I didn't double check to make sure they are the same problems or in the same order so if they're not...oh well. Start email:

Thanks for sharing these and I'll obviously keep them to myself. As far as rollout settings I do try to push the envelope when rolling out backgames because I have seen far too many positions that 3 ply through XGR+ get wrong so cutting down on these issues is a priority. Still, when the top play is not only ahead by a lot but matches my own play and intuition, I accept it without worrying about it too much. It's only if something comes out on top which I don't understand am I really concerned. Here are my thoughts as I look at the positions. You can feel free to use them however you like. The numbering matches the problems.

  1. As far as common occurrences go in backgames this one comes up often enough. This was very easy for me. He has the timing right now to sit on the bar for a bit and then come in and not crash too much after a decent sequence. By breaking our inner board point and slotting the back of the prime we are okay with being hit and recirculating. Often this concept is overdone in backgames but here it holds true. The opponent has no other timing to hold his position. If we put together a solid 5 prime or even 6 prime and allow him to enter his forward checkers will quickly crumble to the lower points in his inner board.

  2. The first thing to decide is whether or not I want to hit. His timing if I don't hit is quite good being down ~90 pips when I'm getting close to coming home. Hitting will clearly jack up our gammon and backgammon chances while not costing us really in the win column so I hit the second checker. After that I have to decide how the rest of this game is going to play out. He has two on the bar so it'll probably take a couple turns before he even has his footing again. At that time, since I'm so close to home already, I will want to tear up my prime to my high inner board points for later flexibility. I can no longer hope to out time him or have him break so why not dismantle the prime now? Next question, how to dismantle it, 9/5 or 8/4?

    In general I prefer a spare on the 5pt to either the 6pt or 4pt. Here however we have to ask ourselves what our poor numbers will be when we begin the bear off. Bingo, fours. We want men on that point and have the opportunity to do that now. I would rate this problem as very tough and don't expect many people at all to get it right.

  3. After a bit of consideration I landed on not hitting as well. It's easy to see that the opponent will likely crash on his own if we just let him play. This play seems easy for a backgame problem too. (if there is such a thing as *easy* for backgame checker play)

  4. I missed this one. It is clear that some sort of hari kiri play is called for but I went with the conservative 12/10* 10/6 4/2* which keeps the best points and offers up what I thought ample blottage to be hit to keep the timing and not sacrifice our board completely. What my play fails to do is force the opponent to hit two men in nearly every situation. I guess this is more important and that our timing issues outstrip our issues with rebuilding the board later. I hate when things get out of order. You didn't even have my play rolled out so I added it with your settings.

  5. Again, the same type theme as #1 that I've seen a hundred times before. Breaking from a 6 prime to a 5 prime is allowed (esp. when they won't be at the front of it) whereas breaking from a 5 prime to a 4 prime generally is not. We of course would also prefer to hit the most forward checker in our opp's inner board as we can. Next!

  6. Another next for me. All too common is this type of slotting into a double or triple shot even when trying to stymie a backgame. The positional make up of the opponent sometimes matters but here there's so many numbers that don't hit combined with the weak forward position he has so even when hit we'll have nothing to worry about makes it clear.

  7. Easy. Timing, flexibility.

  8. I played 18, 7 here rather quickly. His timing is shot if we let him play so hitting is out. Our biggest problem becomes escaping the back checker so we toss it into the outfield. The only question now is which 3? I guess the duplication of aces combined with the few full prime making numbers we gain is worth playing 10 opposed to 7. I'm still counting this as 'correct' on my score sheet though.

  9. Three of them are forced and then we're left with the same theme as in two of the other problems. Next!

  10. You may not even believe me at this point but I covered a very similar problem some time ago while giving lessons. The person is not in a backgame nor should you allow them to try. What you're doing is trying to get 6 of their men on the bar and have the other 4 already on the 3pt sit there. This is a blitzing position, not a backgame position.

So I got 9/10 or 8/10 if you want to dock me for #8. I'm not sure how you score your quizzes but hopefully it's by equity given up and not by simply the # right or wrong. The issue with grading 'equity lost' style though is your rollouts need to cut out a decent amount of the variance to be used to score in such a way. I wouldn't bother with 4 ply rollouts for these. They seem pretty standard and agreeable.

Stick

Messages In This Thread

 

Post Response

Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:

 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.