Hi, a quick question spurred by Yogi's comment in the thread regarding backgammon rules.

I know in chess that playing someone with an elo 100 points above yours you're about 35% to win. Does this same relationship hold in backgammon, or does the random element of the dice mess this up? Has anybody looked at win rates versus elo rankings in backgammon?

So I threaded about this two years ago, and I don't know what is the state of affairs at present, but as far as GNU goes,it was in no way able to assess complicated backgames.

I believe intermediate vs beginner has 2:1. The win ratio of champ vs intermediate is a little disappointing.

In what way is the win ratio of championship players versus advanced players "disappointing"? (I think it's just about what it should be.)

I know it is disappointing, but I don' t know exactly what it is.

Okay, I did some research, and there is difference between PR and ELO. PR is calculated through your error rate by the bot, and ELO through your match wins and losses relative to the opponents strengths.

Elo has got some weak points in that it doesn’t differentiate between close and clear match wins/losses, and that it can be cheated in several ways. Apart from that, players can have bad and good periods, and those fluctuations are ofcourse not visible in someones score.

The advantage of ELO versus PR is in my opinion that the psychological factor in using the doubling cube is not part of the PR equation. For example, a strong player's pride and resulting anxiety, might let him drop prematurely against a weak player, or could set him on tilt. And also facial expressions, the poker element, has a place in backgammon.

There is a formula for calculating your winning chances on basis of the -difference between your rating and your opponent’s ELO rating- and the length of the match.

In theory this looks very sophisticated, but how well does it work in practise. I am very curious what the outcome would be in case the different elos (beginner, intermediate, advanced, expert) would be playing 100-matches against each other.

I don't know if it should. 7-6 vs 7-2 could be only one dice roll difference, not a reflection of the strengths of the players. If it isn't part of the backgammon Elo calculation, match difference could be easily added. Since Elo is a long-term thing, the vagaries of dice will be smoothed out, as are the temporary ups and downs that we all experience. If someone is constantly crushing someone else, the Elo changes would reflect that as compared to a series of narrower wins.

I think you are right. Still, in case of two boxers with the same win rate, we are more impressed with the boxer with the knock outs. But I am not sure of the comparison.