BGFed.be

After two wins, Zsolt faced Maurits in the 3rd round of the 1st Brugge Backgammon Day.  The match turned out to be an interesting encounter with a shockingly spectacular Crawford game.  You can download the annotated match (*.xg).

Maurits (left) and Zsolt (right) exchange niceties before the match

With the score tied at 1-1, both players found themselves in an opening skirmish.  In such games it is mostly correct to keep hitting – Thou shallt hit in the beginning – but at a certain moment it becomes correct to anchor up and cut your losses.  When that moment has come is not easy to assess.  Here is how the game went :

Game 3

Zsolt Tasnadi (Black) needs 6                              Maurits Pino (White) needs 6

1.     32 : 24/21 13/11                                               32 : 13/10 6/4*

2.     43 : bar/21*  24/21                                         64 : bar/21 24/18

3.     53 : 8/3 6/3                                                      31 : ?

3.     —–                                                                     31 : 18/14*

Hitting catches up in the race of course.  White is only down 4 pips after the hit.  The alternative would be to anchor up  on the 18-point and play 10/9 with a less volatile position as a result.  This turns out to be a serious error here :

1. XG Roller+ 18/14* eq:-0,31
Player: 44,87% (G:9,61% B:0,37%)
Opponent: 55,13% (G:19,20% B:1,22%)

2. XG Roller+ 21/18 10/9 eq:-0,376 (-0,067)
Player: 42,62% (G:8,46% B:0,28%)
Opponent: 57,38% (G:18,47% B:0,62%)

So, the fact that Black has made the 3-point should not deter you from hitting, even though the hit leaves 4 blots.

4.     65 : bar/20 21/15*                                              32 : ?

Maurits sees himself faced with the same problem : he can keep hitting and leave 5 (!) blots against a stronger board, or he can anchor up. What would you do?

4. —–                                                                             32 : bar/23 13/10*

Hitting is still correct, but by a much smaller margin :

1. XG Roller++ Bar/23 13/10* eq:-0,434
Player: 42,65% (G:9,33% B:0,36%)
Opponent: 57,35% (G:22,47% B:1,84%)

2. XG Roller++ Bar/23 24/21 eq:-0,457 (-0,024)
Player: 40,63% (G:8,45% B:0,29%)
Opponent: 59,37% (G:18,95% B:0,78%)

Just to show how subtle changes can make hitting wrong, have a look at this variation :

How to play this 42?

The extra spare on the 8-point and the fact that you can make the golden anchor are sufficient to make hitting an error :

1. XG Roller++ Bar/23 24/20 eq:-0,395
Player: 41,59% (G:8,69% B:0,27%)
Opponent: 58,41% (G:17,48% B:0,65%)

2. XG Roller++ Bar/23 14/10* eq:-0,422 (-0,027)
Player: 42,47% (G:8,82% B:0,30%)
Opponent: 57,53% (G:20,55% B:1,39%)

5.      62 : bar/23 21/15*                                                    51 : ?

Hit-or-not?

5.     —–                                                                                51 : bar/20 21/20

Maurits intuitively knows that enough is enough.  Time to wave the white flag and fight a defensive war :

1. XG Roller++ Bar/20 21/20 eq:-0,364
Player: 43,28% (G:9,44% B:0,33%)
Opponent: 56,72% (G:20,13% B:0,88%)

2. XG Roller++ Bar/20 6/5* eq:-0,429 (-0,065)
Player: 44,12% (G:9,64% B:0,39%)
Opponent: 55,88% (G:25,82% B:2,75%)

Still,  circumstances may occur where hitting here may be a tactical option worth considering.  A world-class player might keep hitting  to complicate matters.  And, should this position come up towards the end of a match, the trailer might rightfully hit here if his oppenent is in time trouble.

6.     32 : 23/20 13/11*                                                           31 : bar/24 13/10*

7.     52 : ?

Hit or not?

Now it is Zsolt’s turn to ponder whether he should hit or not.  This position is fundamentally different, however.  Here, Black has a significant edge (58% winning chances), whereas in the previous positions, White was always a clear underdog (40-45% winning chances) no matter what he did.  This points to a consolidating play : Black should come in on the 20-point and make his 11-point.  It is not so difficult to see that the 11-point is very strong against the 20-point anchor. Moreover, White already has 5 men back, so hitting a 6th checker might lead to a complex triple-anchor backgame where each checker play might pose challenging problems.  Also, hitting gives up the golden anchor, a decision one should not take lightly.

Not hitting here yields a simple holding game with a 41-pip race advantage, low volatility and easy checker play coming up.  The fact that Zsolt made the wrong choice here is indicative of how the intuitive hit may lead one astray in certain positions.

Still, the error is far from monstrous.  Forward mobility and flexibility are important in positions where O (= opponent) has established multiple anchors.

1. XG Roller++ Bar/20 13/11 eq:+0,397
Player: 57,73% (G:18,75% B:0,91%)
Opponent: 42,27% (G:8,37% B:0,27%)

2. XG Roller++ Bar/20 11/9 eq:+0,341 (-0,056)
Player: 56,38% (G:18,28% B:0,88%)
Opponent: 43,62% (G:8,55% B:0,28%)

3. XG Roller++ Bar/23 20/15* eq:+0,331 (-0,066)
Player: 54,68% (G:22,44% B:1,72%)
Opponent: 45,32% (G:9,85% B:0,41%)

7.     —–                                                                                          43 : bar/21 24/21

8.     62 : 15/9 11/9                                                                        63 : ?

How to play this 63?

Having established a 5-4 backgame structure, it seems natural to just eject one of the back men into the outfield.  This is probably what I would have done.  Maurits, however, decides to go for Black’s barpoint by simultaneously slotting it and hitting on his 5-point.  An ingenious tactical find in a complex position!

1. XG Roller++ 24/18 8/5* eq:-0,281
Player: 46,58% (G:10,32% B:0,43%)
Opponent: 53,42% (G:23,40% B:0,84%)

2. XG Roller++ 24/15 eq:-0,304 (-0,023)
Player: 45,03% (G:9,65% B:0,37%)
Opponent: 54,97% (G:20,59% B:0,62%)

3. XG Roller++ 23/14 eq:-0,309 (-0,028)
Player: 45,10% (G:9,65% B:0,38%)
Opponent: 54,90% (G:21,01% B:0,70%)

The reader can follow the rest of this exciting game via the link provided above.  Both players held their own amid all that complexity : Zsolt scored a 4,36 PR (World Class), but was still outplayed by Maurits who even outplayed the computer : – 0,68 PR (Wow!).  Nicely done.

M

For the 10th time, Bert Van Kerckhove organized the Gent Open.  As always, the location was the Koninklijke Club Union Sandeman  in Gent.  This year, 24 players showed up to play 7 rounds of 7-point matches.  All matches were played with clock.

After 6 rounds, Michel was the only player left with 6 victories.  However, in the last match of the day, Belgian-born Nardy Pillards managed to beat Michel to draw even.  The tense match culminated in a thrilling DMP game where Nardy skillfully blitzed Michel off the board after a 52 (24/22 13/8) 55 (6/1* 6/1* 8/3* 8/3*) opening sequence.

Nardy (left) and Michel (right).

This left both players with 6 wins, but the trophy went to Michel because his opponents had collected more wins (Buchholz-system).

Michel and Bert. Also (left to right) Johan Brisaert, Hans Böhm, 2 young fans, Marc Steyvers and Peter Gelpke.

The Intermediate Division was won by Manolitto from Amsterdam.  He was one of 6 players from the Netherlands who had travelled a couple of hours to compete in Gent.  One of his opponents, Peter Allemeersch, was baffled by the fastness of Manolitto’s decision-making : “He played incredibly fast,” said Peter, “and ended the match with 11 minutes and 24 seconds left on the clock!”  Truly, an amazing feat.

Intermediate winner Manolitto.

In the picture, Manolitto poses with the colourful trophy.  All three trophies were handcrafted by Gent club member Robin Bilderbeek.

The Beginner’s flight was won by Kristoffer De Weert.  In addition to the trophy, he received Bill Robertie’s bestseller Backgammon for Serious Players as study material.

Kristoffer and Bert. On the left, trophy designer Robin.

This anniversary edition of the Gent Open was a success.  There was constant action and the playing conditions were excellent.  Congrats to organizer Bert for a job well done.

An extra word of thanks to the Dutch players who made the trip.  They gave the tournament an international flair and blended in perfectly with the predominantly Flemish crowd.

M

Liliane Baptista (left) vs Zsolt Tasnadi (right)

In backgammon, certain types of positions occur time and time again.  One of the most common types are holding games.  In a holding game, one player has made an advanced anchor (4-point or higher) and the opponent has escaped his men to the midpoint.  Now it is time for the favorite to assess all features of the position and decide on his cube action.

In their match, Zsolt and Liliane each got doubled in a 4-point holding game and struggled with the cube decision.  Here, for instance, is the position in Game 1 :

White (Zsolt) on roll.
Can he double? Can Red (Lilane) take?

First thing to consider is – as always –  the RACE.  Zsolt is only up 11 pips.  Typically, against a 4-point anchor you would need about 20 pips race advantage to have a correct double.  Of course, with his barpoint already made,  Zsolt has a broken 5-prime which compensates for the fact that he has not yet fully safetied his last checker to the midpoint.  Also, Liliane’s own forward position is far from perfect : she has not yet made a extra point in her board, but she should have no problems building a strong board soon.  Moreover, in such kinds of positions the volatility is generally low.  This means that only rolls like 66, 55 and 44 would lose Zsolt’s  market.  Therefore it is very improbable that Zsolt would make a big error by waiting one roll before cubing.

Player Winning Chances: 73,72% (G:2,76% B:0,05%)
Opponent Winning Chances: 26,28% (G:1,24% B:0,01%)

No double: +0,799 (-0,009)
Double/Take: +0,807
Double/Pass: +1,000 (+0,193)

Best Cube action: Double / Take

As eXtreme Gammon shows, this is a borderline Double, but an easy Take for the defender.  Liliane does not take a big risk in taking this.  She can only lose one extra point by taking, since the chance that she will get gammoned is less than 3%.  However, players who are not familiar with these positions might easily get spooked and give away a point.  This is what happened in the match.  Liliane passed, making a big mistake in doing so.

Knowing your opponent’s tendencies can help with the cube action here.  Maybe Zsolt estimated that there was a significant chance that Liliane would pass.  In that case, his double becomes a strong bluff double.  If he did not consider Liliane’s weakness, his double could be called premature.

In Game 2, the roles were reversed.  This time Zsolt is the one defending from the 4-point anchor :

Red (Liliane) on roll.
Can she double? Can White (Zsolt) take?

Here, Liliane is up 26 pips in the RACE.  She could have turned the cube 2 rolls earlier, but decided to wait.  This turned out well for her, because Zsolt, probably thinking the open 5-point was a significant weakness in Red’s position, took the cube without thinking too long about it.  This also turns out to be a big mistake :

Player Winning Chances: 81,58% (G:2,79% B:0,02%)
Opponent Winning Chances: 18,42% (G:0,90% B:0,04%)

No double: +0,988 (-0,012)
Double/Take: +1,202 (+0,202)
Double/Pass: +1,000

Best Cube action: Double / Pass

The Americans have come up with the term double whopper for errors of this magnitude.  By accepting this cube, Zsolt sacrificed 0,202 mps in equity.

Something to consider when you get doubled is whether the opponent will be prone to making checker play mistakes in the game continuation.  In the position above, Red has a relatively easy task bringing the checkers home.  Still, she might not dare leave an indirect shot in clearing the 11-point or misplace her checkers in some other way.  Still, these errors are unlikely to be big mistakes, so they would not compensate for the 0,202 in sacrificed equity.

However, this is what happened 2 rolls later :

The reader can have no doubt as to what is the correct move : making the 5-point is hugely correct.  I am sure Liliane will play this correctly 90% of the times, but due to the tension of the big occasion, she nervously played 8/6 8/2.  Her underlying thought may have been that a point cleared, is a point not to be feared, and there are situations where this would definitely be the right play.  Here, however, this useful tenet does not apply.

1. XG Roller++ 11/5 7/5 eq:+0,661
Player: 84,29% (G:2,56% B:0,04%)
Opponent: 15,71% (G:0,61% B:0,01%)

2. XG Roller++ 8/6 8/2 eq:+0,484 (-0,177)
Player: 76,71% (G:2,59% B:0,05%)
Opponent: 23,29% (G:1,30% B:0,02%)

To make matters worse for Liliane, Zsolt ended up escaping with a 66 and reached this position some rolls later :

White (Zsolt) on roll.
Can he double? Can Red (Liliane) take?

This is a classic bear-off cube decision.  If Zsolt rolls an ace, he loses, in all other cases he wins.  This is typically something that distinguishes Intermediate players from Beginners : the ability to calculate probabilities.  What are the chances Zsolt rolls an ace? Well, it is 11/36 or 30%.  That means he is 70% favorite here and should redouble.  Liliane, who has 30% winning chances, should take.  It must be said that it might be scary for the better player to give this cube, because rolling an ace swings the score to 1-4 down in a match to 7.  However, the error in not cubing is so big (0,415 – !), that the skill difference between both players would have to be between 12 and 15 PR ( = difference in Performance Rating as calculated by eXtree Gammon) for not doubling to be correct.  This means that a world class player, who averages about 4 PR, would make an error cubing this against Liliane.  Zsolt, who played this match at Advanced level, had to bite the bullet and ship the cube.  He took a long time fretting over it, but eventually redoubled to 4.  Liliane, to her credit, accepted, which as a matter of fact is very well played.  Many players, even strong ones, get overwhelmed by superstitions and end up passing this.  Gutsy cube action by both players!

Zsolt did not roll a 1, took a 5-0 (7) lead, and ended up beating Liliane 7-0.

Hopefully, these remarks will motivate players of all levels to spend some time studying the game.  Improving your game gives greater long-term satisfaction than winning the occasional match by chance.

See you soon.

M

Scoresheet Paul vs Geert

Last Saturday, 14 March 2015, a 49-point match was played at the Royal Sandeman Club in Ghent.  Only a few spectators witnessed the match, among them Ghent player Johan Vervust, who could not believe his ears when I told him about the lenght of the match.

I have had this idea for a couple of months and eventually found Dutch player Paul Van Dijke, never shy of a real challenge, prepared to take up the gauntlet.  Play started at 1:30pm and finished just before 8:30pm, when Paul got gammoned in the last game of the match, which ultimately finished 50-33 in my favour.

The match was clocked, each player starting with 1 hour and 38 minutes in the time bank, with a 12 seconds increment each move.  Only 7 or 8 small 2-minute breaks were taken, in order to be sure we could finish the match that same day before the Sandeman closed its doors.

In the match we saw two 16-cubes, which both had to be passed (either player won one).  The match was fully recorded and the full transcription in Extreme Gammon will follow in about a week.  For the moment, only the score sheet is available (see picture above).

Geert Van der Stricht