At the end of February, the UK took on Germany in a friendly match. There were twelve players on each team. This was the UK team: Gaz Owen, Oliver Squire, Rick Janowski, Chris Bray, Aref Alipour, Julian Fetterlein, Sebastian Wilkinson, Tim Cross, Jon Barnes, Lawrence Powell, Charles Hill and Steve Bibby. Chris Rogers was the non-playing captain with overall UK captain, Martin Barkwill, providing assistance.
Each person played two 13-pt matches so there were twenty-four points available (there were no points awarded for PR performances). After the singles matches were completed the overall score was taken forward into a 21-pt consulting doubles match and two players from each team were selected by their captains to contest the doubles.
The teams were very evenly matched, so it was quite a surprise to find the UK trailing 8-16 after the singles were completed. Starting at 8-16 down in a match to 21pts is obviously a huge disadvantage and despite Gaz Owen and Sebastian Wilkinson playing nearly flawless backgammon in the doubles, the German team triumphed by the final score of 21-12.
Can we explain away the loss so that we can learn from the experience and do better in subsequent matches? I did some statistical analysis of all twenty-five matches to try to get an understanding of our team’s performance.
The average PR performance for the two teams was: UK: 4.42 Germany: 3.99 so not a huge difference but still, it is a difference. Had points been awarded per match for PR performance as well as the actual score the outcome would not have changed significantly. The next thing to look at is the luck factor and here the Germans definitely had the edge, gaining an average of 1.0 per person.
Understanding the luck factor in XG is still something I am trying to fully understand but it is clear that it was a factor in determining the result. However, it is a universal truth that we all think we are unlucky in backgammon because we remember the bad luck but seldom the good luck.
I watched Aref Alipour’s second match where, with the cube on 8 in a complex back game, he missed a treble shot followed by two double shots, all of which would nearly certainly have won the game and match for him. I was certain that Aref was hugely unlucky but in fact his luck rating for that match was a tiny positive because of luck he had in the earlier games which, of course, I and other spectators, had forgotten.
On the other hand, Sebastian Wilkinson was truly unlucky. Over his two matches his average luck factor was – 6.76. That sort of number is virtually impossible to overcome, no matter how good you are. I watched Seb’s final match, and his opponent produced a sequence of super-jokers that swung the match in his favour.
In preparation for the doubles match myself, Chris Rogers and Tim Cross took on Gaz and Seb in a training match. Using Zoom the two teams collaborated with each other and could hear each other’s discussions. The short match took a couple of hours but the PRs for the two teams were 1.07 and 1.79. This proved a couple of things to me. Firstly, in backgammon one plus one equals more than two. Having a partner prevents oversights and creates a depth of analysis that an individual cannot achieve alone. Secondly, if you take your time you play much better. I have watched a lot of matches during lockdown and the most common fault of virtually everybody is in playing far too quickly.
This was borne out in the actual doubles match where Gaz and Seb had a PR of 0.81. Their opponent’s PR was a relatively poor 3.45 but their huge lead from the singles saw them home. You also have to accept the fact that sometimes tiny errors can create a huge swing in equity. In the Crawford Game in the doubles this was the key position:
Gaz and Seb (UK), playing Red, have a 31 to play. The choice is between 8/7, 8/5 and 7/3. Which would you choose? Gaz and Seb played 8/7, 8/5, as would I, and the German team responded with the super-joker of 66 and went on to win the game and match.
On a rollout XG has 7/3 as better by 0.012. Not an error you would normally lose sleep over but here it cost the loss of the game and therefore the match. Why is 7/3 the better play? Probably because, precisely, it does not allow the 66 joker, but I am by no means sure of that piece of analysis!
In summary, the UK played well and perhaps we were a little bit unlucky but the better team won. The team all enjoyed playing and it certainly promoted Anglo-German relations, which was one of the objectives of the match. I think we learnt a lot from our training doubles match and that perhaps points the way to a training method that should be used more often. The more often we take the time to analyse in depth the stronger we will become.
The next international match is against Romania over the weekend of 16th-18th April.
I don’t believe that the luck factor in XG is something that actually tells you who is the luckier player because it is measured in normalised equity rather than in match equity gained or lost through luck. So the effect of luck where the cube is high is massively undervalued relative to luck with a centred cube.
You can view the XG equities in units of MWC: “Show cost”. Luck factor can change based on the units, but usually only a little bit. MWC is better, strictly speaking, but the normalization essentially rescales the MWC into money-like units that we’re more familiar with…. That’s my short response. There are other problems with luck factor, mostly having to do with difficulty of decisions, about which I could go on and on.
Re diagrammed position, if you break the 8 point you give a joker 66 and a 64 to hit. If you keep the 8 point you leave only 55 as a joker, and it’s not as devastating as the 66 joker is after 8/7 8/5. There may be other small considerations, but with White’s forward position getting precarious, the short-term variations are very important.