From: Backgammon Galore : Woolsey-Bagai Match

Kit Woolsey vs. Jeremy Bagai
Game 2

Annotated by Kit Woolsey, Jeremy Bagai, and TD-Gammon

Match to 9:   Kit Woolsey - 2,  Jeremy Bagai - 0

Kit Woolsey is playing Red.
Jeremy Bagai is playing White.
Kit wins the opening roll.

Position 1: Red to play 51.

TD: It's nice to see that you guys are finally learning to play the opening rolls correctly. I'm glad some of my teachings have sunk in.

    24/23, 13/8      +.021
    13/8, 6/5        -.006
    24/18            -.023

Position 2: White to play 52.

Kit: This play of an early 5-2 is becoming more popular as players are understanding the importance of splitting the back men quickly. Here it is probably best, since 13/11, 13/8 exposes the blot to two indirect shots.

Jeremy: It looked like this was better than 13/11, 13/8 because of the six fly shots, but those are balanced by the value of the builder when it is missed. The plays are probably equally good.

TD: Correct. The split is important, and the indirect shots are too costly.

    24/22, 13/8      -.079
    13/11, 13/8      -.098
    13/8, 6/4        -.126

Position 3: Red to play 65.

Jeremy: Anchoring is certainly better than escaping one checker.

TD: Right. Making the anchor is clear.

    24/18, 23/18      +.125     
    24/13             +.033
    24/18, 13/8       -.015

Position 4: White to play 11.

Kit: Jeremy has several choices. The best defensive play is 24/21, 22/21, which gives him an advanced anchor and we would then be in a mutual holding game. Making the defensive three point with 24/22, 6/5(2) is not as strong. Jeremy's play balances a strong offense with excellent board coverage. Even though he doesn't make an anchor, I will have a difficult time playing safely. I like his play.

Jeremy: The five point is a must. Anchoring on the 21 point would be a waste considering that Kit is not yet threatening anything offensively. Advancing to the 20 point covers more of the outfield in order to make it more difficult for Kit to bring builders down. Notice that even as early as my second roll all of Magriel's criteria argue for a bold play rather than a defensive play.

TD: I'm not exactly sure why I like 22/21, 6/5(3) a bit better than Jeremy's play -- maybe because since Jeremy is behind in the race preparing to make the anchor a bit farther back and keeping more contact while not making Kit's aces strong is the idea. At any rate, Jeremy's play is fine, certainly better than making either of the anchors.

    22/21, 6/5(3)        -.092
    22/20, 6/5(2)        -.098
    6/5(2), 6/4          -.099

Position 5: Red to play 32.

Kit: I couldn't stomach the thought of playing safe with 13/8, putting a fifth checker on the eight point, so I hit loose and went after my five point. I hate to say it, but I think the safe 13/8 might be better because Jeremy has the stronger board but I am ahead in the race, therefore I don't want a blot hitting contest.

Jeremy: 13/8 would be a mistake. My board is better, but it is only two points and Kit does have an anchor. The opening is a battle for the five points, and Kit rightly fights for his.

TD: Kit shows excellent insight in his analysis -- too bad he couldn't find it at the table. 13/8 is much better than hitting on the five point. The position thematically screams for a safe play. Kit is ahead in the race and has nobody stuck back, but Jeremy has the stronger board. Kit does not want to get involved in a blot hitting contest. He shouldn't even particularly care if Jeremy makes his five point.

    13/8            +.082
    13/11, 8/5*     -.014
    8/5*/3          -.038

Position 6: White to play 42 from the bar.

Position 7: Red to play 32 from the bar.

Kit: I keep flailing away, but this strips my eight point as well as leaving many shots. I think I should have played the more solid B/22, 13/11 and waited for a better chance to attack.

Jeremy: I'm pretty sure this is right, although it's less clear than last time because Kit strips his eight point. An interesting example of when Magriel's criteria conflict with each other. Kit has more men back and an anchor which argue for bold plays, but has the weaker board which argues for safe plays.

TD: Once again Kit's post-mortem is better than his actual play. While I am a big fan of knocking my opponent off my five point, the circumstances have to be right for it. They just aren't here. Stripping the eight point and failing to lock up the 11 point cost too much. Hitting loose is a serious error.

    B/22, 13/11      -.102
    B/23, 11/8       -.174
    B/23, 8/5*       -.203

Position 8: White to play 41 from the bar.

Position 9: Red to play 61 from the bar.

Kit: Since I will have to leave a shot whatever I do, I might as well make one more try to win the fight for my five point.

Jeremy: This is clear because B/18 leaves the direct shot anyway.

TD: Correct, for all the reasons given by the players.

    B/24, 11/5*      -.273
    B/18             -.323
    B/24, 13/7       -.402

Position 10: White to play 52 from the bar.

Kit: This is probably a bit better than B/20*, 23/21, since the balance of four men on the six point and three on the eight point is better than vice versa. The reason is that you are sometimes willing to give up your eight point in order to make an inner board point, but you will never want to give up your six point.

Jeremy: Four reasonable deuces: 23/21; 20/18; 13/11; 8/6 (6/4 is not reasonable -- it leaves a triple direct shot and strips the six point). 13/11 is in fact worthless because my five point is made and my bar point is taken. The blot would simply be a target. 23/21 would be disadvantageous because it gives up coverage of Kit's inner board, allowing him to play safe behind me. 20/18 is pretty neutral. 8/6 is actually beneficial. It's usually right to have more spares on the six point than on the eight point. I'll be willing to break my eight point long before I break my six point.

TD: Close, but Jeremy's accurate judgment is correct. The spare on the six point is better than on the eight point.

    B/20*, 8/6         +.355
    B/20*, 23/21       +.341
    B/20*, 13/11       +.332

Position 11: Red to play 54 from the bar.

Kit: My play maximizes my chances to make another anchor, but I expose myself to a possible attack. The play is ok, but now I slightly prefer B/21, 23/18. The third checker on Jeremy's bar point gives me some badly needed flexibility.

Jeremy: 23/18, 21/16, and 13/8 are all reasonable 5's. I have no idea which is better but would probably make Kit's play.

TD: As Jeremy says, all the fives are reasonable. However, bringing the spare to the bar point, which Kit suggests in his commentary, is the slight favorite. Flexibility is the key here; the attack forces can wait.

    B/21, 23/18      -.358
    B/21, 13/8       -.373
    B/16             -.381

Position 12: White to play 61.

Kit: Jeremy is content to leave me alone and run one man to safety. The problem with this play is that it is now too easy for me to make another anchor, after which my defense will be solid. I think he should have tried to hit me when I am down and back, but if I should happen to roll badly from the bar he would have excellent attacking chances.

Jeremy: Again Magriel's criteria point in different directions -- I have the stronger board, but no anchor and fewer men back. Putting two in the air with 8/2*/1* might lead to a quick blitz if Kit dances, but it leaves a direct shot, starts a point I don't really want, and strips my eight point. 20/13, disengaging, is better.

TD: I hate it when such widely differing plays come out so close

Jeremy's play won, but it was too close to be sure. Btw, don't ignore the much more flexible 20/14, 6/5. That prettier play goes well with my circuits, and almost won top billing.
    20/13           +.316
    20/14, 6/5      +.312
    8/2*/1*         +.310

Position 13: Red to play 21.

Kit: I grab the most advanced anchor. Now that my defense is solid, there is every reason to slot the five point. If I am hit the extra man back may prove to be an asset; if I am not hit my game has started to develop.

Jeremy: Kit shows that he reads the same books I do: Two anchors, five men back against one -- time to start slotting.

TD: Back to good thematic backgammon. Well done, Kit.

    23/21, 6/5        -.303
    24/21             -.362
    23/21, 8/7        -.369

Position 14: White to play 54.

Jeremy: No alternative is close.

TD: Jeremy also follows the theme of the position. He properly runs and makes the safest play rather than trying to stir things up since he is well ahead in the race.

    23/14           +.288
    13/9, 6/1*      +.212
    23/18, 13/9     +.195

Position 15: Red to play 21.

Position 16: White to play 65 from the bar.

Position 17: Red to play 54.

Kit: I am short on ammunition, so I just put the checker where it belongs. If Jeremy hits, I might be able to make another anchor in his board. If he doesn't hit, I have a chance to improve. More passive plays such as 13/9, 11/6 and 13/8, 11/7 are possible, but I don't think they go to the heart of the position.

Jeremy: Of course Kit will play bold: now his board is as strong as mine and he still has five men back and two anchors. But which bold play should he make? 13/4 starts the point that Kit wants most, but gives up hope of re-establishing his midpoint and also leaves the blot on his 11 point where it isn't doing much good since his five point is already made. There are other plays. 18/13, 11/7 re-establishes the midpoint which prevents his army from getting too divided, and starts his bar point, bringing the checker on the 11 point more into play. The cost is releasing the anchor on my bar point. 11/7, 8/3 is the ultra-aggressive play. It starts the three point (almost as good as the four point), starts the bar point -- bringing the checker on the 11 point into play, and retains both anchors. The downside is that it leaves four blots. But should Kit care? He still has two anchors. This is my play, although any of the three could be correct.

TD: Just when it looked like you guys were getting a handle on the position you got confused. Well I guess I can't blame you too much -- it's a pretty complex position which only an excellent program like myself could be expected to solve. The proper play is 18/14, 18/13, a play you never even considered. Jeremy at least suggested 18/13, 11/7, which is the right idea but carries it out the wrong way. The key here is to see that Kit's main problem is that his men aren't communicating and that he has lost outfield control. He obviously had to give up his midpoint last roll in order to hit Jeremy's fleeing blot, but now he has a chance to recover it. Kit no longer needs two anchors, since he is not restricted to playing defensively. He needs to bring all the men he can muster to cover his outer board in order to prevent Jeremy's back checker from fleeing again. This isn't immediately obvious, but if you look at what Kit's primary goal right now is, the move should stand out.

    18/14, 18/13      -.119
    21/16, 13/9       -.142
    18/13, 11/7       -.145
    13/4              -.147

Position 18: White to play 33 from the bar.

Kit: The ten point is a very big blocking point when I am camped on Jeremy's four point. This is better than B/22, 8/5, 6/3(2).

Jeremy: The obvious alternative is B/22, 8/5, 6/3(2). This makes the three point which is clearly a good thing, but I like my play better. The ten point will be a valuable landing spot for clearing the midpoint, which may end up being my biggest obstacle once my back man escapes. I think the key here is that making the three point, while nice, leads to a static position which may not play very smoothly. The stack on the midpoint has only one place to go, and I will have only one other spare checker to play with up front. My play gives me lots of spares which creates important flexibility. Also note that I'm not interested in attacking any more checkers, so making inner board points is not as high a priority as it would be otherwise. I'm pretty sure my play is correct.

TD: That's the idea. Put your men in front of his anchors, not behind them. Jeremy's play is by far the best, and making the three point is so bad it doesn't even get on my top three list.

    B/22, 13/10(3)          +.210
    B/22, 13/10(2), 8/5     +.180
    B/22, 13/10(2), 6/3     +.156

Position 19: Red to play 31.

Kit: It is easily worth the slight impurity to put Jeremy on the bar. 8/4 allows him to escape with a five or a six.

Jeremy: Better than 8/4 which would allow me my whole roll to escape.

TD: I agree. Making the three point on Jeremy's head is clearly stronger.

    6/3*, 4/3      -.155
    8/4            -.213
    8/5, 4/3*      -.303

Position 20: White to play 42 from the bar.

Jeremy: B/23, 10/6 might be better because it doesn't leave a blot on the point which Kit most wants to make. Still he only has one builder to attack there, and I like the builder distribution on my side better with my play. I'd make my play again, but could easily be wrong.

TD: Coming in on the 21 point is much better. Jeremy's play is correct.

    B/21, 10/8      +.183
    B/21, 8/6       +.178
    B/23, 10/6      +.131

Position 21: Red to play 65.

Kit: The idea behind this play is to cover the outfield so Jeremy can't bring his blot around too easily. Reasonable, but I am leaving several shots and it's not clear the gains are all that great. 24/18, 11/6 is comfortable and puts more pressure on Jeremy's back man if he is unable to move it. I now prefer that play.

Jeremy: Very tough play. We can quickly eliminate any play which breaks either of the anchors. That leaves us with *only* 24/13; 24/18, 11/6; 24/18, 8/3; 11/5, 8/3; and 11/6, 8/2. 24/13 keeps all the checkers in play in front of my straggler and provides three builders for the bar. However, it releases coverage of my inner board which allows me to play behind his anchor in safety, doesn't bring another builder to attack the more important four point, and leaves a direct shot in the outfield. 24/18, 11/6 keeps all checkers in play, adds a builder for the four point, and leaves no shots. It also releases coverage of my inner board, and by moving from the 11 point releases coverage of Kit's outer board making it easier for my blot to survive if it escapes. 24/18, 8/3 keeps outfield coverage, but releases coverage of my inner board and takes a checker out of play on the three point. 11/5, 8/3 keeps coverage of my inner board, brings a second checker to attack the four point (remember the eight point still attacks even if stripped), but takes a checker out of play on the three point. 11/6, 8/2 does the same thing, but starts the two point instead of putting a "builder" on the three point. I think we can eliminate 24/18, 8/3. It looks wrong, and seems to have fewer advantages than any other play. 11/6, 8/2 seems to dominate 11/5, 8/3 (meaning it's better in all ways), so we can eliminate 11/5, 8/3. After that I'm not really sure. I have a feeling that the extra builder for the four point is the most important factor so I would play either 24/18, 11/6 or 11/6, 8/2, but I might be very wrong. An interesting point is that there is not a computer in the world that I would trust *at all* in rolling out this position (Sorry TD). Come to think of it, I don't think I'd give that much faith to any human either. Not that rolling it out would be a waste of time; anyone who did it might learn a lot about what's important in this position. But I certainly wouldn't trust any numbers.

TD: You don't have to apologize to me, but if you can't trust my rollouts, whose can you trust? Anyway you won't get them here; you'll have to settle for my opinions. On this one the two logical candidates, Kit's actual play and 24/18, 11/6 were in a near dead heat, with 24/18, 11/6 a slight winner.

    24/18, 11/6      -.263
    24/13            -.265
    21/10            -.310

Position 22: White to play 31.

Kit: Not much choice. If Jeremy quietly plays 8/5, 6/5 I will just go after his back man. Sending me back rips away one of my builders, and I don't have too many return shots. He can live without the midpoint; it would be a point he would have to clear eventually.

Jeremy: 6/5, 6/3 is wrong because there may be an exchange of hits immediately so I don't want blots in my inner board. If you were thinking about the position last game where I advocated not being afraid of this kind of parlay, recognize that the difference is that building my board is not even a priority for me here -- escaping is -- whereas in that position building my board was essential. 8/5, 6/5 might be right -- it gives me better distribution up front and leaves only fours to attack with plenty of return shots most likely. My play tries to do more. It clears the midpoint and puts Kit on the bar, preventing him from using his full roll to develop. But at some cost -- eight shots hit the blot on the 12 point and nine shots hit the blot on the 21 point. Granted hitting on the 12 point would give up my bar point, and hitting loose on the inside leaves returns. I guess I'd make my play again, but I really have no clue which is better.

TD: I agree. Hitting here is very valuable. Sometimes you just have to duke it out.

    13/12*, 13/10      +.161
    8/5, 6/5           +.123
    6/5, 6/3           +.059

Position 23: Red to play 31 from the bar.

Kit: No reason to step up with B/21. Keeping the checker back will make it harder for Jeremy to bear in safely. In addition, moving the spare on the eight point to the five point gives me another attacker for his blot since I am willing to give up the eight point if necessary but not give up the five point.

Jeremy: Best. Covers my inner board and deploys the new builder. If you don't see that 8/5 creates a new builder (because Kit is willing to give up the eight point to attack on the four point), go reread Magriel's section 2: Using Men Effectively. No other play is close.

TD: Kit makes a fine play. The checker back on the 24 point is a definite asset. Note that B/21 isn't even on my top three list.

    B/24, 8/5      -.254
    B/22, 8/7      -.280
    B/22, 11/10    -.290

Position 24: White to play 65.

Jeremy: This leaves fewer shots than 21/16, 12/6 (15 vs. 16), forces Kit to break an anchor to hit, and will be easier to safety if missed.

TD: You think I can't see all these things?

    21/10           +.269
    21/16, 12/6     +.248
    21/15, 6/1*     +.178

Position 25: Red to play 66.

Kit: I guess this is best. The alternative is 18/6(2), which keeps the back man in place but loses the valuable defensive bar point. Since that bar point puts double pressure on Jeremy's checkers on the ten point I think holding it is a better idea. Since I must have my four point in all variations, my play gives me the best distribution to make it in the future.

Jeremy: Kit is forced to play at least one checker from my side of the board, and he decides that the dual anchors will generate more shots than 18/6, 18/12, 8/2. I'm pretty sure he is right.

TD: Yep. Holding both anchors is much better than playing from just the 21 point and keeping the back man back. The increased pressure on Jeremy's outfield points makes the difference.

    24/6, 8/2           -.435
    24/12, 8/2(2)       -.439
    24/12, 11/5, 8/2    -.474

Position 26 before roll: White's cube action?

Kit: I think Jeremy should double. 20 pips is a pretty hefty lead in what is likely to become a race. I have a pretty easy take of course with the combined racing and shot-hitting chances, but if he gains much on the next exchange he could easily lose his market.

Jeremy: I should certainly glance at the cube. I'm 20 pips up in a race of 118 pips, a double/pass in a straight race -- which this, of course, is not. First I have to safety that blot. Then I have to bear in against two anchors. It looks to me as though Kit may well be taking until I clear my ten point, so there's no reason to double until I'm threatening to do just that.

TD: I agree with Kit. Equity up to .435 means you don't need a whole lot of volatility to send the cube over. This is just good enough.

Position 26: White to play 21.

Position 27: Red to play 21.

Jeremy: Much, much better than making the two point which would only be correct if Kit's only hope was getting a shot on the very next roll.

TD: Of course. Making the four point is clear.

    6/4, 5/4        -.423
    11/10, 6/4      -.448
    11/9, 5/4       -.456

Position 28 before roll: White's cube action?

Kit: This is basically the same position as last turn. Now Jeremy wakes up and properly turns the cube.

Jeremy: What about now? I safetied my blot, but Kit did great things to his board. One important point is that *no roll* clears my ten point, so unless this position is already a pass I probably don't need to double. But is the position a pass? I have no idea! In the second issue of Inside Backgammon (March/April 1991), Kit Woolsey has an excellent article on the double/no double decision in which he states his fundamental law: "If there is ANY doubt in your mind as to whether a position is a take or a pass, then it is ALWAYS correct to double. To put it another way: Suppose Magriel or Sylvester or even God came up to you and said: I think this position is a pass, and I am willing to pay a point and the cube for the highest stakes you can afford. If you would not instantly say 'sit down', then you are not 100% sure, which means you must automatically double". This is a great law, and a good double.

TD: Sorry, I don't agree. In fact, I think Jeremy is very inconsistent. His equity is down a tad (from .435 to .423) since Kit made the four point), and the volatility also dropped a bit. How can it not be a double last turn yet be a double now? In fact, I think this is just short of a double. Of course playing against frail human beings who might get confused and pass trivial takes such as this maybe you are supposed to double, but I'm not programmed to think that way. I only do what is right.

Kit: Looks like a solid take. I have some shot hitting chances which double anchors often bring, and the race is not totally out of sight. The combined chances should give me the necessary equity to justify the take.

Jeremy: Hmmm. This is not a very familiar position. Kit's dual anchors are strong, but how long will he be able to hold them? He only has two checkers in the outfield to soak up his next few rolls, and then either an anchor goes or his board crunches. This is a timing consideration -- one of the more elusive of backgammon concepts. I think I would take this as well. Even if the forward anchor goes, the checkers on my ten point may become very isolated if funny things happen. Most important is that Kit's board is perfect. I gave the position to Expert Backgammon and had it roll it out 1296 times. It says that Kit loses .54 points per game, which translates to 1.08 points per game when doubled, compared with the 1 point per game Kit loses if he drops. However, EXBG plays the game to conclusion with no cube -- which is a big disadvantage to Kit who will actually be able to double me out when he hits a shot in our game. This factor, coupled with the fact that any data from EXBG must be taken very generally because it often play badly in complex positions, indicates that it's a tough decision and a good double. I take.

TD: Get serious, Jeremy. Equity of -.423 is EASY take. Not a tough decision at all. Sometimes you humans make mountains out of molehills. As for trusting absolute equities on Expert Backgammon rollouts, don't make me laugh. I'll admit the program plays pretty good for a non-neural network program whose parameter weightings were crafted by fallible humans rather than generated from vast experience as mine are, but while for play vs. play decisions it will usually give you a decent answer you just can't trust its absolute equities for cube decisions. Only I can do that job properly.

Position 28: White to play 32.

Kit: The right idea. Jeremy's goal is to clear his ten point as fast as possible, keeping spares on the other points in order to handle bad rolls. 6/3, 5/3 would definitely be wrong.

Jeremy: Making the three point is the wrong idea. That play removes a flexible checker from the six point and does nothing to clear my ten point. I don't care about inside points because I'm not planning on hitting any more shots -- I just want to disengage. 10/8, 5/2 might be just as good.

TD: My algorithm came up with 8/3. Admittedly this is not my strong area (maybe I should work on it some -- my boss, Gerry Tesauro, says I lose too many won games), but end games just aren't as much fun as middle games. If they like 10/5, they are probably right. At least I didn't think that making the three point was a serious contender.

    8/3           +.441
    10/5          +.427
    10/8, 6/3     +.426

Position 29: Red to play 33.

Jeremy: There goes the timing. Kit is correct not to play 8/2, 4/1(2) which would keep the five point board while preserving a six to play next roll so that his dual anchors might be able to stick around. I'm not leaving shots next roll, so Kit would be forced off an anchor soon anyway and would have lost his four point in the process. Naturally he doesn't leave his rear anchor which is the strength of his game.

TD: Of course it is proper to hold everything. Anything else would be esthetically displeasing as well as just plain wrong.

    11/2, 8/5          -.433
    21/18, 11/2        -.472
    11/2, 4/1          -.488

Position 30: White to play 51.

Kit: Looks right. In order to move from the ten point Jeremy would have to play 10/5, 6/5, leaving a very awkward position.

Jeremy: The only play which keeps a builder on the six point. Flexibility is all.

TD: I agree, though I have it pretty close. Maybe the spare on the six point isn't as important as they think, although admittedly it does look valuable.

    8/2            +.409
    8/3, 6/5       +.406
    10/5, 6/5      +.400

Position 31: Red to play 11.

Jeremy: Not 5/3, 2/1(2) or 3/1(2). Kit knows that he's leaving my bar point next roll, so he starts the next point that he wants, intending to cover it with the checkers coming around.

TD: Anything else would be just plain stupid.

    5/1             -.469
    5/3, 2/1(2)     -.503
    6/3, 5/4        -.521

Position 32: White to play 52.

Kit: Jeremy chooses to make the three point in case he gets a shot. However, this is not consistent with what his game plan should be. He is stripping his eight point and making sixes potentially very awkward. I think he should stick to the clear from the back policy and play 10/8, 10/5.

Jeremy: I see that Kit may be leaving a shot on my bar point next roll so I make an inner board point for offensive purposes. This has some merit, but I now think it is clearly wrong. My goal is to clear the ten point, not to hit blots. Stripping the eight point early opens up all sorts of scenarios where I'm forced to clear it first, leaving my ten point stranded in the outfield. 6-6, 3-3, 6-5, and 6-3 do this to me on the very next roll, while 6-4 actually leaves a double shot. Much better is 10/8, 10/5.

TD: So call me a simpleton. I believe in making points when I can. I like 8/3, 5/3. Kit and Jeremy obviously see some obscure dangers in the position. I don't. Simple plays for simple minds and computers. Of course, I wouldn't exactly bet my entire circuitry on this one.

    8/3, 5/3        +.543
    10/3            +.486
    10/5, 8/6       +.460

Position 33: Red to play 62.

Kit: This is not a matter of duplicating his threes to cover. It is simply leaving fewer shots.

Jeremy: Is 18/16 the right deuce? Doesn't Kit want as much contact as possible? Don't I have a blot in my board? My instinct first told me that Kit should stay back with 18/10. The general principle is that he wants an exchange of hits, not to disengage. In other words, the extra shots that he leaves by staying back don't hurt him much because he is already behind in the race, while the prospect of extra return shots should be very welcome. On closer inspection, however, those return shots just don't come up because no roll forces me to leave them. In fact, the only bad roll I have is 6-4, in which case staying back will garner Kit an extra seven shots at the blot on the eight point. Given that upside staying back would be correct -- if there were no downside. But what about Kit's subsequent 3-3? Say goodbye to a board or an anchor. Kit's play ensures no disaster on that roll. How do you compare these two very remote sequences? Let's look at a cross-section of 1296 games. I roll a 6-4 twice in 36 games (6-4 and 4-6), and 7 times out of 36 in each of those games Kit is happy he stayed back: 7*2 = 14 games. On the other hand, Kit will be unhappy in all of the games in which he rolls 3-3 assuming I still own my ten point (I'll have cleared it if I rolled double 5's, 4's, 2's, or 1's) so that's 32 games. But don't stop there -- who's to say that crashing is as bad as hitting is good? I'll guess that hitting is twice as good as crashing is bad, so that leaves us with 28 on the upsode and 32 on the downside ... which is too close to mean anything whatsoever. Oh well. I've probably missed some crucial aspect of the position anyway. Kit? TD?

Kit: Yep, I think you've lost the forest through the trees. The really big swings occur when you hit and cover -- something you are easily willing to do, since the few indirect shots that would leave are far more than compensated for by the increased gammon possibilities, the claiming in the race, and the improved chances of clearing the outer board points safely while I am on the bar. I think this factor far outweighs anything else.

TD: Don't look at me. I have Kit's play a slight winner, but by such a small margin that it doesn't prove anything. You guys can fight this one out between you -- I don't want any part of it.

    18/16, 18/12      -.525
    18/10             -.528
    18/12, 3/1        -.576

Position 34: White to play 11.

Kit: Hitting is clearly a must -- Jeremy can't let me go since the race could get close. Having hit, he must put three men on the two point since he can't afford to be hit back. 10/9*/8, 3/2(2) is possible, but the gaps in the outer board may be too serious. I think Jeremy's play is best, although there could be some problems. He prepares to clear the back point next turn if he can; otherwise he will probably be able to clear some other point.

Jeremy: The alternatives are 8/7(2), 3/1 and 3/1(2). Clearing the eight point is wrong because it isolates the 10 point. I hit because it looked like making the nine point would help clear all the outside points, but it does allow Kit to enter on my ace or three, and actually creates a point rather than clearing one. I like my play, but I'm not sure by any means.

TD: 10/9*/8, 3/2(2) feels a bit smoother to me, but I'm far from sure. What I am sure about, however, is that not hitting would be a big blunder. This race is not gin by a long shot if Jeremy doesn't hit.

    10/9*/8,  3/2(2)    +.550
    10/9(2)*, 3/2(2)    +.542
    10/9(2)*, 8/7(2)    +.475

Position 35: Red to play 54 from the bar.

Position 36: White to play 54.

Kit: Holding the blockade with 6/1, 5/1 would leave a completely stripped position which is very dangerous. Jeremy properly clears the back point, putting some spares on the six and five points to help handle awkward rolls.

Jeremy: Obviously better than making the ace point.

TD: Not even in the same league.

    10/6, 10/5      +.549
    6/1, 5/1        +.477
    10/1            +.155

Position 37: Red to play 62.

Jeremy: better than 21/15, 7/5 simply in that it leaves more cover numbers for the ace point.

TD: Back to the technical nonsense. I like bringing the man home. Who knows or cares (I certainly don't).

    21/15, 7/5      -.543
    21/13           -.562
    21/15, 3/1      -.578

Position 38: White to play 61.

Position 39: Red to play 51.

Kit: If I play 13/7 I'll just have to give up the bar point next turn, and if I don't roll a six the ace point will remain uncovered. Making it now gives me 12 free pips to play with before something has to give way.

Jeremy: This is much better than the "clever" 13/7 which creates a prime and allows Kit to play for a second checker. Kit will not need to play for a second checker until I have six or seven men off, five or six rolls from now. The main problem is that if he makes his bar point he will just have to break it next roll, and may not be able to make the ace point.

TD: Do we need all this discussion to see the obvious -- a closed board is a closed board. These guys are just trying to sound intelligent, and they aren't doing a very good job of it.

    7/1           -.597
    13/8, 2/1     -.629
    21/15         -.636

Position 40: White to play 43.

Kit: Jeremy is rushing things. I don't think this is a good idea. My board figures to hold out for one more roll; after that something will probably have to give. After Jeremy's play, 6-1 and 6-4 leave a direct shot. I think he should play 6/3, 5/1. Not only is this 100% safe for next roll but it is smoother and builds a board in case Jeremy needs one in the future.

Jeremy: This leaves a shot on 6-1 and 6-4 and leaves me unhappy on 4-4 while after 6/3, 5/1 all my rolls play well. Yet I still think my play is best. I want to clear the eight point, and stripping it is the way to do that. It might be right to stall with 6/3, 5/1 if Kit's board were going to crash next roll, but he has the checker on his midpoint to play with. 8/5, 6/2 is a blunder, leaving a shot on 6-6 unnecessarily.

TD: I'll go along with Kit on this one. 6/3, 5/1 seems better on all counts.

    6/3, 5/1      +.608
    6/3, 6/2      +.562
    8/1           +.521

Position 41: Red to play 22.

Position 42: White to play 33.

Kit: There is no rush to come down to two men on the six point with 8/5(2), 6/3(2). That would leave a shot on 6-2, while Jeremy's play is safe for the next roll. I think he can do even better with 8/5, 8/2, 3/0. This comes down to four men on the six point, so with no cost he can delay the evil day even longer, in the meanwhile taking men off and possibly forcing me to either leave with both men or lose my board.

Jeremy: There doesn't seem to be any reason to play 8/5(2), 6/3(2) which would leave a shot on 6-2 next roll. 8/5, 8/2, 3/0 might be just as good.

TD: My algorithm comes up with the play they say is terrible. Well, I do seem to lose a lot more of these games than I should. Better listen to them on this one. I really don't feel on firm ground here -- the objectives aren't clear to me.

    8/5(2), 6/3(2)      +.782
    8/5(2), 6/0         +.735
    8/5(2), 6/3, 5/2    +.734

Position 43: Red to play 54.

Jeremy: Kit sees that breaking his board would hurt both his racing chances and his hitting chances -- and he would probably be forced to leave next roll anyway. He correctly keeps one guy back for annoyance value. There are many two roll sequences where I leave a shot.

TD: Breaking a closed board? Not even in the picture.

    21/12           -.798
    21/16, 6/2      -.839
    21/16, 5/1      -.841

Position 44: White to play 65.

Position 45: Red to play 41.

Position 46: White to play 54.

Kit: Doesn't take anybody off, but leaves a shot only on large doubles. 5/1, 5/0 would leave a shot on 6-2 next turn as well as large doubles. Next turn is most critical; after that I probably won't be able to hold my board.

Jeremy: If the gammon were a possibility, 5/1, 5/0 might be right. Here the gammon is very unlikely, so clearing the six point is a must.

TD: My maker put a new algorithm in to handle positions such as this one, and I come up with 5/1, 5/0. Their discussions about what will happen on the next roll or two are a little too technical for me, but I guess we might as well bow down to their knowledge here -- they are supposed to be experts at this sort of thing.

Position 47: Red to play 41.

Position 48: White to play 42.

Jeremy: Not 5/3, 5/1 which leaves many shots next roll. No reason not to take the checker off.

Position 49: Red to play 54.

Kit: There really isn't much point in staying. I won't be getting a shot next roll, and after that not only will my board be crunched but Jeremy will have so many men off that I probably can't win even if I hit a shot. In the meantime, there would be the danger of a freak gammon. It isn't worth the risk.

Jeremy: Kit sees that nothing leaves a shot next roll, and runs now to guard against a freak gammon. Seems right.

TD: I run also, but I have it very close. Of course, these decisions aren't my forte.

    21/12          -.963
    6/1, 5/1       -.964
    6/1, 6/2       -.973

Position 50: White to play 21.

Position 51: Red to play 42.

Position 52: White to play 62.

Position 53: Red to play 51.

Position 54: White to play 64.

Position 55: Red to play 31.

Position 56: White to play 65.

Position 57: Red to play 62.

Match to 9:   Kit Woolsey - 2,  Jeremy Bagai - 2

Go on to next game: Game 3

Return to: Backgammon Galore : Woolsey-Bagai Match