| Annotated Game |
 Joe Dwek |
 Lee Genud |
Unlike the last several years, when in each case at least one relatively unknown player had struggled his way to the finals, these two participants boasted formidable reputations.
Joe Dwek has been considered Europe's best player for a decade.
In 1976, he scored his most notable tournament triumph, edging Kumars Motakahasses in the European Championship at Monte Carlo (the European Championship in those days was the precursor of the current World Championship; until 1979, the World Championship was held in the Bahamas in January). Dwek has been first or second in many other tournaments in both Europe and the Americas, compiling a record second to none in the world. In 1976 he authored a book, suggestively titled Backgammon for Profit, which offered one of the first real glimpses into the checker and cube techniques of backgammon's elite.
Lee Genud's tournament record, while not quite so formidable as Dwek's is nonetheless an impressive one. Her most recent victory was the Women's World Championship in Florida last September. She and Dwek are even in the books at one apiece, hers being Lee Genud's Backgammon Book, published in 1975.
Both had survived tough semifinal matches. Dwek had beaten Michael Cahmi of France, while Genud defeated Philip Swart of England. Curiously, this was the first match Swart ever lost! He had swept six rounds to win his first tournament as few months earlier in Amsterdam, and had won six straight in Monte Carlo. Swart is an economics student in London, and more will certainly be heard from him in the future.
Final matches in major backgammon tournaments (and this was no exception) have two salient features: the games are played at lightning speed (the thirty-five games of this match were completed in 2.5 hours) and almost every double is passed (of the thirty-six doubles in the match, only eight were accepted, although later analysis showed that nearly half were takes). Both features are, of course, a function of the nervousness of the players. Since Joe's experience was assumed to make him relatively immune to jitters, the bookmakers installed him as an 8-to-5 favorite.
As the match developed, the lead seesawed a bit, although there were no big swings. In fact, there were no 4-games in the entire match; the players were very tight. Lee took an early 7–5 lead, but Joe caught up and went ahead 10–8 by game 15. Lee then surged, sinning nine of the next twelve games to lead 19–15. The players traded points to reach 23–19, Lee's favor, through game 34 (match to 25). There followed the remarkable thirty-fifth game, a game in every respect worthy of the occasion.
Game 35
(Black trails 19–23. Match to 25.)
 Joe Dwek
|  Lee Genud
|
|
| White to play 4-2. |
| 1. | 4-2: 8/4, 6/4 |
Lee wins the opening roll, as she did in twenty-five of the thirty-five games of the match, a remarkable percentage. Curiously, it didn't do her much good! She won only thirteen of the twenty-five games where she moved first, while she won six of the ten games where Joe moved first. Most authorities think that the advantage of the opening roll conveys about an 11-to-10 advantage.
| 2. | 6-1: 13/7, 8/7 | 6-4: 24/14 |
| 3. | 4-2: 13/11*, 24/20 | 6-2: bar/23, 13/7 |
|
| Should Black double? |
| 4. | DOUBLE |
In a money game, this would be an early double and a trivially easy take. Trailing 19–23, Joe is theoretically justified in doubling slightly earlier than normal. Even so, I think his double here is, in a theoretical sense, premature.
Practically, a double here is mandatory, since many players in Lee's position would cringe and drop, frightened by their three blots and the spectre of gammon chances. Part of the art of catching up when trailing in a match is timing your initial doubles to catch your opponents at their most vulnerable point, when they are most likely to drop a perfectly takeable game.
Unfortunately for Joe, he had just doubled Lee out of two easy takes in games 33 and 34. Lee was consequently much less inclined to drop this double. Still, there was some chance, and I think Joe was correct to turn the cube.
| 4. | . . . | TAKE |
| 5. | 5-1: 24/18* | 3-1: bar/22, 6/5* |
| 6. | (dance) | 3-1: 8/5, 8/7* |
| 7. | (dance) |
|
| White to play 3-1. |
| 7. | . . . | 3-1: 23/20, 22/21 |
This is an outstanding play, perhaps the best play of this entire long and intricate game. When the play was made, the entire panel of experts commenting on the game (Gino Scalamandre, Kent Goulding, Kit Woolsey, and myself) criticized it and recommended 13/9 instead as clearly correct. To see why Lee was right and all of us were wrong, take a closer look at the position (above) after Lee rolled her 3-1.
If this position occurred in a money game, and White doubled Black, my guess is that 95 out of 100 players would pass, and the other five would be making what they considered to be a bad take. Actually, whether Black should take or not depends entirely on how will conduct the position. White has two reasonable plans:
- Plan 1: (The tactical variation.) Pull men down from the midpoint and try for a quick blitz.
- Plan 2: (The positional method.) Escape the back men, in the process hitting Black's loose checker on the eleven point. Allow Black to enter but refrain from hitting loose, relying instead on a prime to win.
White doesn't have enough ammunition in place to make Plan 1 work. In fact, if White has a fourth checker sent back, Black can quickly become a favorite. If Black knew in advance that White would try Plan 1, he would have quite an easy take. Plan 2 is stronger, and would make a take by Black risky (though no outrageously bad). In the absence of any information about how White would conduct the game, a take by Black wouldn't be unreasonable here.
Lee deserves high praise for finding the correct plan under the intense pressure of the occasion, while the experts in the calm of the outer room were bamboozled.
| 8. | 4-2: bar/23 | 6-2: 20/14*, 13/11 |
|
| Black to play 3-2. |
| 9. | 3-2: bar/23, bar/22 |
By establishing an anchor, Joe has avoided the worst. He can now play a two-point game, with some chance of transposing into a 1-2 or 2-3 backgame. Since he can't be gammoned at this score, these variations are much stronger for him than they would be in a normal money game.
| 9. | . . . | 6-1: 14/7 |
|
| Black to play 3-1. |
| 10. | 3-1: 8/5, 6/5 |
Joe has a couple of weaker alternatives here. The all-out backgame play (7/4*, 6/5) is silly. Joe still has chances to go forward and should retain them as long as possible. 13/10, 23/22 is more plausible; however, the two-point game offers more long-term winning equity than the three-point game, and in the absence of gammons Joe is right to stay on the rear point. Making the five point is a simple, strong play.
|
| White to play 6-6. |
| 10. | . . . | 6-6: 21/3*, 13/7 |
Lee's play of 13/7 is better than 11/5. By keeping her checker on the eleven point, she increases her chances of eventually making the eight point.
|
| Black to play 6-2. |
| 11. | 6-2: bar/17 |
If Lee makes a five-prime, Joe wants as few checkers as possible trapped behind it. The play is not as risky as it might appear; notice that Lee has only eight numbers that both hit and make the twenty-two point.
| 11. | . . . | 4-1: 7/3, 24/23 |
|
| Black to play 1-1. |
| 12. | 1-1: 17/14*, 6/5 |
A crucial move, which merits close examination. Joe must hit with three of his ones. He then has three plays for the last one.
- Play A: 6/5 (Joe's play)
- Play B: 14/13
- Play C: 8/7
To assess these plays, we have to realize that Joe has two simultaneous and conflicting objectives. First is not to be hit. If hit, he loses most realistic chances of going forward. His second goal is extending his 3-prime to a 4-prime next turn. Obviously, maximizing his chances of extending his prime also maximizes his chances of being hit. To get a handle on the position, let's do some preliminary calculations. First, let's see how likely Joe is to be hit after each play.
- Play A: 18 hit numbers next turn (50%)
- Play B: 11 hit numbers next turn (30%)
- Play C: 7 hit numbers next turn (20%)
Next, let's see how likely it is that Joe makes a 4-point prime, assuming he isn't hit.
- Play A: 29 rolls make a 4-prime (80%)
- Play B: 20 rolls make a 4-prime (55%)
- Play C: 8 rolls make a 4-prime (22%)
To see the real chance that Joe makes a 4-prime, we have to multiply these percentages by the chance that Joe isn't hit, which gives these figures.
- Play A: Joe isn't hit, then makes 4-prime: 50 × 80 = 40%
- Play B: Joe isn't hit, then makes 4-prime: 70 × 55 = 39%
- Play C: Joe isn't hit, then makes 4-prime: 80 × 22 = 18%
Looking at this last set of figures, we can see that Joe's play essentially gambles everything on the next roll; he is even money to be hit, but if he isn't hit he is very likely to make his 4-prime. Play B is almost as effective in making a 4-prime as Joe's play, and has the advantage of being much safer. Joe, after all, isn't required to make a 4-prime on the next roll in order to win the game. As long as he isn't hit, he retains chances of extending his prime in the future. Play C, although very safe, looks too weak. I think Joe made an error here: 17/14*/13 looks, on balance, to be the best choice.
| 12. | . . . | 3-2: bar/22, 13/11* |
| 13. | (dance) | 6-5: 23/17*, 13/8 |
| 14. | 5-1: bar/24 |
|
| White to play 4-2. |
| 14. | . . . | 4-2: 17/13, 8/6 |
Hitting with 7/1* is a serious error. If Joe responds with a one, the game could become a toss-up. Lee has to maintain her 5-prime to keep control of the game.
A better play, however, is 8/4, 17/15, which increases the number of rolls that make the ace point if Joe stays out next turn. Lee's play duplicates ones, but that hardly matters in this position.
| 15. | 5-1: bar/24, 13/8 |
|
| White to play 5-1. |
| 15. | . . . | 5-1: 22/17*/16 |
A tough decision. I think Lee made a good play here. Beginner manuals caution the novice to avoid hitting when defending against a backgame, so as not to improve its potential timing. This is, of course, an oversimplification. There are many positions where a hit is clearly correct, many others where the choice is murky. Briefly, here are some of the factors that mitigate in favor of Lee's hit.
- The hit helps ensure that the game will in fact be a backgame. If Lee plays 22/16 (not hitting), Joe could send either or both of Lee's checkers back behind his own growing prime. Lee could then find herself in a chancy prime-vs-prime game, rather than a backgame.
- The hit may hurt Joe's timing if he immediately reenters. If Lee can trap six or seven men behind her prime, Joe will not have enough men left to maintain a strong forward position in case of an early hit.
- Not relevant here, but a hit increases Lee's gammon chances considerably.
Lee's play isn't clearly correct, but I like it.
| 16. | 4-2: bar/23, 13/9* | 6-4: bar/15 |
| 17. | 5-2: 9/4, 6/4 |
|
| White to play 3-1. |
| 17. | . . . | 3-1: 15/12*, 6/5 |
This hit looks right. Joe's blot represents twelve useable pips. By removing those pips from the outfield, Lee leaves Joe with very little spare timing. Unless he manages to dance for several turns, his board is likely to crunch.
| 18. | 4-1: bar/24, 7/3 | 5-4: 13/8, 12/8 |
|
| Black to play 5-3. |
| 19. | 5-3: 7/2, 6/3 |
An excellent play, much superior to making the two point. By dismantling the six point, Joe prepares to deprive himself of fives.
|
| White to play 2-2. |
| 19. | . . . | 2-2: 11/3 |
Also good. Lee is in no hurry to clear the seventeen and eighteen points. If Joe's next roll contains a six, he won't be able to free a man.
|
| Black to play 6-2. |
| 20. | 6-2: 6/4 |
Lee's play pays off. Joe meanwhile removes his last five. He now has a good chance of preserving the rest of his board until he can free one or two back checkers
|
| What are Black's odds of winning the game from here? |
Although Joe has broken his six point, he has not yet buried any checkers and is very much in the game, especially since gammons and backgammons don't count against him. To get an approximation of the real odds, I played this position to a conclusion 100 times. The results:
- 10% White won a backgammon
- 10% White won a gammon
- 20% White won a single game
- 30% Black won a single game
If Lee owned the cube in a money game, he position would be slightly too good to double, given the large gammon chances. In this situation, with gammons not counting for Lee, Joe's chances are quite reasonable: he is only a 7-to-3 underdog. In many tournament situations, Joe would have quite an easy take in this position, another example of the big difference between tournament and money backgammon
|
| White to play 5-1. |
| 20. | . . . | 5-1: 8/3, 5/4 |
I would play 8/7 with the one. Lee doesn't gain much by being hit, and 8/7 prevents her from playing a five next turn. A very close play in any case.
| 21. | 4-2: 5/1, 5/3 | 4-2: 7/3, 7/5 |
| 22. | 4-2: 5/1, 4/2 |
|
| White to play 5-2. |
| 22. | . . . | 5-2: 8/6, 5/off |
A serious error. She should play 8/3, 6/4, preparing to clear the six point. When bearing off against contact, be careful to avoid being stripped on interior points, as Lee is here on the twenty point. Such positions are much more likely to leave a series of repeated shots than smooth positions.
| 23. | 5-1: 24/18 | 4-1: 4/off, 6/5 |
| 24. | 5-5: 23/8, 18/13 |
|
| White to play 3-1. |
| 24. | . . . | 3-1: 3/off, 6/5 |
A little better is 3/off, 5/4. After this play, nine numbers leave a single blot. After Lee's play, eight numbers leave a single blot and two numbers (6-4 and 4-6) leave a disastrous double blot (and a quadruple shot!)
| 25. | 1-1: 8/6, 13/11 | 6-3: 6/off, 6/3 |
| 26. | 5-1: 11/6, 3/2 |
|
|
| 26. | . . . | 6-3: 5/off, 3/off |
| 27. | 5-4: 23/14 |
|
|
|
| White to play 4-2. |
| 27. | . . . | 4-2: 4/off, 4/2* |
An alternative here is 4/off, 5/3. It is occasionally correct to leave a blot on a higher rather than a lower point in the home board if the chances of leaving a blot on a subsequent turn are rendered much less. This isn't the case here. Lee's play is immediately safer (sixteen shots vs. twenty for the play 5/3), and also safer on the second turn (fifteen subsequent blot numbers vs. twenty-one). Lee's play has the added advantage of reducing Joe's ace-deuce backgame to a mere ace-point game.
| 28. | 6-2: bar/23*/17 |
|
|
| 28. | . . . | 5-4: bar/16 |
|
| Black to play 4-3. |
| 29. | 4-3: 24/21, 14/10 |
This might seem like an obvious play. Actually, there is a very difficult decision to be made here, and I think Joe makes a slight error. There are three moves to consider.
- Play A: 24/21, 14/10 (Joe's play, giving maximum outfield coverage)
- Play B: 24/17 (creating an indirect anchor, blocking 8's)
- Play C: 14/7 (denying Lee any direct shots in the outfield)
- Play A: 40%
- Play B: 24/139%
- Play C: 32%
| 29. | . . . | 6-4: 16/6 |
| 30. | 4-1: 24/19* | (dance) |
|
| Black to play 5-2. |
| 31. | 5-2: 19/12 |
Slotting with 10/5, 17/15 is incorrect. Joe would have only four cover numbers for the checker on the five point (6-4, 5-5, and 4-4). Joe shouldn't slot the point until he has at least a direct cover number (or the equivalent in indirect numbers).
| 31. | . . . | (dance) |
|
| Black to play 1-1. |
| 32. | 1-1: 12/9, 17/16 |
A mistake. The best play is 12/11, 17/14. Leaving the checker on the eleven point guarantees a double shot against all the rolls containing a five which do not hit (5-1, 5-2, 5-3, and 5-4). Joe's play gives a double shot against 5-1, 5-2, and 5-3, but only a single shot against 5-6.
Incidentally, playing 6/5(2) with half of the roll is a serious error. It deprives Lee of the chance to roll 5-1, exposing a crucial second blot.
| 32. | . . . | 5-1: bar/20, 3/2 |
|
| Black to play 2-1. |
| 33. | 2-1: 6/5*, 16/14 |
Excellent. Once the exclusive property of a small coterie of top masters, this type of play has become more widely known in recent years. If Lee dances, Joe may be able to close her out directly. If she enters but doesn't cover the twenty-three point, Joe may be able to send a second checker back, making him a heavy favorite. Only if she rolls one of the eight numbers that both hit and cover (6-1, 6-3, 5-1, 5-3) will Joe be worse off with this play.
| 33. | . . . | 6-1: bar/19*, 3/2 |
| 34. | 3-1: bar/24, 9/6* | 6-5: bar/19*/14 |
|
| Black to play 4-1. |
| 35. | 4-1: bar/24, 10/6 |
Joe's play is a trifle less accurate than bar/24, 14/10. By removing the blot that is directly attacked in the outfield, Joe creates slightly more hitting chances for himself next turn.
Here the play that is the most effective in creating double-blot numbers next turn is bar/21, 14/13. By blocking tens, it forces Lee to expose two checkers with 6-4 and 4-6, while the other plays only yield two blots if Lee's next roll is 5-5. However, this play is much weaker in immediate shot numbers, so I don't recommend it.
| 35. | . . . | 4-3: 14/11*/7 |
| 36. | 5-4: bar/16 |
|
| White to play 3-3. |
| 36. | . . . | 3-3: 7/4*, 5/2(3) |
No respite! With $45,000 riding on the outcome, the players are forced over and over again to make pressure-packed decisions.
Lee makes an error here. She has already borne off seven men. At this intermediate level (between six and ten men off) checkers are more important than shots. Consequently, she should play 7/4*, 3/off(2), 5/2, leaving her with nine men off. This would make her a slight favorite even if she were hit and closed out.
Curiously, the play I recommend is also safer in terms of immediate shots. Lee's play looks safe, but it leaves Joe with seventeen return shots (all 4's 6-3, 5-3, and 3-1). Playing 7/4*, 3/off(2), 5/2 actually leaves only sixteen shots (all fours, 6-3, 3-1, and 3-3). On the second turn, Lee's play is considerably safer, with only fifteen rolls leaving a subsequent blot, as opposed to twenty-one after bearing off two men.
|
| Black to play 4-2. |
| 37. | 4-2: bar/21*/19 |
Bar/21*, 16/14 offers both more cover numbers for the six and five points (9's and 8's instead of 11's and 10's) and better coverage of Joe's outer board if Lee reenters.
| 37. | . . . | (dance) |
| 38. | 3-1: 6/5, 16/13 | (dance) |
| 39. | 4-3: 19/12 | (dance) |
|
| Black to play 6-3. |
| 40. | 6-3: 13/10, 12/6 |
Joe correctly slots. If Lee misses, he can cover with all 4's (except 4-2), 3-1, 1-1, and 6-6 (thirteen numbers).
| 40. | . . . | (dance) |
|
| Black to play 4-2. |
| 41. | 4-2: 24/18 |
Incredible irony. 2-2 is the only other noncovering cover number. After this play, Joe has four new cover numbers: 4-2, 3-3, and 2-2.
| 41. | . . . | (dance) |
|
| Black to play 6-1. |
| 42. | 6-1: 18/11 |
Bad luck for Joe. Lee stayed out five times in a row, but Joe couldn't cover.
This play is a slight inaccuracy. It leaves him with twenty-seven covers next turn, while the correct play (18/12, 10/9) yields twenty-eight covers. Joe's play also yields four double-hit numbers, versus three with the other play.
|
| White to play 6-5. |
| 42. | . . . | 6-5: bar/19*/14* |
Finally Lee reenters, and with a tremendous double hit. Joe's not dead yet, however.
|
| Black to play 6-4. |
| 43. | 6-4: bar/19, bar/21 |
Joe enters both men and stays in the game. Only 6-6, 3-3, and 6-5 bring Lee's last checker home safely
|
| White to play 5-2. |
| 43. | . . . | 5-2: 14/7 |
Amazing. Lee is forced to leave a triple shot.
| 44. | 6-3: 21/18*/12 | (dance) |
| 45. | 4-1: 10/6, 19/18 | (dance) |
| 46. | 1-1: 18/14 | (dance) |
|
| Black to play 2-1. |
| 47. | 2-1: 14/11 |
Again Joe twice fails to cover. After 14/11, he will have twenty-eight cover numbers next turn, assuming, of course, Lee stays out.
| 47. | . . . | 6-2: bar/19*/17 |
| 48. | 5-4: bar/20, 12/8* | (dance) |
| 49. | 2-2: 8/6, 20/14 | 6-2: bar/19*/17 |
| 50. | 6-2: bar/19, 11/9 | 5-2: 17/10 |
| 51. | 3-1: 19/15* | (dance) |
|
| Black to play 2-1. |
| 52. | 2-1: 15/12 |
An error that might have cost Joe Dwek the World Championship. The slot (9/6) is correct. With his outfield checkers on the fifteen and fourteen points, Joe has 9's, 8's, and 6-6 to cover — twelve numbers in all. This is the equivalent of a direct cover number and is enough to justify the slot.
| 52. | . . . | (dance) |
|
| Black to play 2-2. |
| 53. | 2-2: 14/6 |
This would have covered had Joe played 9/6 last turn.
Had he closed his board, Joe would have been about a 70–30 favorite in the game. The difference between this position and that following move 28 is that Lee has moved her checkers from the three and five points to the two and three points, thus improving her equity by about 10%.
|
| White to play 6-5. |
| 53. | . . . | 6-5: bar/19*/14 |
No more jokes. Lee reenters for the eighth (and last) time.
| 54. | 5-5: bar/20, 12/7, 9/4, 24/19 | 6-3: 14/5* |
| 55. | 6-4: bar/21, 19/13 | 4-1: 5/off |
|
|
| 56. | RESIGN | Single game |
Lee Genud wins the match, 25 to 19.