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

BGonline.org Forums

Elo calculator relating to holding cube value

Posted By: Phil Simborg
Date: Saturday, 11 April 2015, at 1:33 p.m.

In Response To: Do I have a prayer? (Stick)

When Mochy and I first agreed to our Challenge match, out of curiosity I used XG's ELO calculator. I used the assumption that Mochy is 100 elo points better than me, about 3 PR, and put in money games and saw that XG says the better player will will .2 points per game.

I had heard from Jake that giving one side the cube with two equal players was thought to be worth about .5 points per game, so XG made me feel pretty good about my bet. Is the difference in our skill worth another .3 points per game? We'll see.

We don't believe that you can replicate on the computer what would happen with two humans playing, even if you could get the computer to play with one player 3 PR worse than the other, and that is why we are doing this experiment.

It seems odd to me that it doesn't matter to XG if the 100 PR difference is 1500 to 1600 ELO or or 1850 to 1950, it's still going to be worth .2 points per game. That might be true in a computer simulation, but I would think that the swings in play and points between a 2 and 5 PR player would be far less than between, say, a 7 and 10 PR player.

Here is my reasoning. There will be far fewer positions where a 2 and 5 PR player will disagree on a cube or checker play, as both will get obvious decision right, and both will get most of the semi-difficult decisions right, and it will only be the more complicated decisions that will separate the two players. The number of times decisions that could be deemed complicated, in a given game, will not be very many. But for the 7 and 10 PR players, the number of decisions that would be complicated for them will be much higher, and the more there are the bigger the advantage to the better player. Particularly in money games where you have far less complexity.

Just from experience, it would seem to me that from teaching hundreds of players over the years, the knowledge-difference between a 7 and 10 PR player is huge. The 10 PR player has far less understanding of recubes, and too good to double, and end game cubes, and has far fewer reference positions for blitzing and holding games and back games in his arsenal.

Both the 2 and 5 PR players have much of the same knowledge base except for very complicated cube and recube situations which get particularly complicated in Match play, but not in money play. And they both have most of the key reference positions for game types, though I am sure the 2 PR player has more and remembers them better and is better able to handle variations. And of course the 2 PR player is far better than the 5 PR player in difficult checker plays. But again, these difficult plays are fewer at those levels.

So wouldn't the points per game be higher at the higher levels? This is my thinking, and I am not sure it is right, but I put it out here to see where the flaw is in my thinking. So far, whenever I start out disagreeing with anything in XG it has been because I was missing something.

As for my Duel with Mochy, it's all set and many have committed to stay in Vegas an extra day to be there at 1PM, not only to see what happens in the 30 game match, but to hear Falafel and Carter's commentary, hear what Mochy, himself has to say about his thinking in complicated or interesting situations, and to hear my excuses and crying. To make it interesting, there is going to be a lot of money on the line and there are already many people with significant side bets, and there will be an opportunity for people to place all kinds of bets on the side. Things will move fast as we will be playing with a clock, using a new setting that adds time if a game is exceptionally long. It will be interesting to see how this new approach to timing works as it may also be good for match play.

It will be broadcast on a big screen for those in attendance, and for those who cannot make it, it will be streamed live with commentary on www.twitch.tv/backgammontv. 1PM Pacific time, Monday April 20th.

Thanks again for Howard to get us a room for Monday and the internet connections, and to one of my students (and now also a very good friend) who has underwritten this event by providing 3 Geoffrey Parker Boards for us, all the expenses for streaming and YouTube, an Erhan board for the Mochy Under25 program, and travel expenses for Mochy to come to Vegas for this event and the US Open.

And here is one last wrinkle to consider. Let's assume that Mochy's PR is 3 PR better than me. That is based on Match play. If it is true, and I believe it is, that money play is not as complicated as Match play, might it be right to assume that our PR difference for money play might not be as great? Does my 45 years of money play experience count for anything compared to Mochy's far less experience in that area? (Of course, most of my money play experience is in chouettes as I have played far less heads up play than chouette for money over the years.)

This is part of what we will find out on the 20th. I believe my biggest challenge during the match will be worrying about what Carter might be saying during the commentary. Since the day I agreed to his working with Falafel I have developed a rash and a nervous tick and I hope they are both under control by next week.

Post Response

Subject:
Message: