lichess.org
Donate

Predicting the Outcome of Chess Games

RuyLopez1000

@jk_182

Cool stuff! I like how you modeled how draw rates change by level and included that as well as the first move advantage.

@jk_182 Cool stuff! I like how you modeled how draw rates change by level and included that as well as the first move advantage.
djconnel

Draw probability is modeled here: https://wismuth.com/elo/calculator.html
It results in this formula, which I derived: drawElo = 16 * exp((whiteElo + blackElo) / 2040)
The "draw Elo" is the Elo advantage of draw odds. So if you look at the difference in win probability between the true Elo and the Elo with draw odds, that gives the draw probability.

Draw probability is modeled here: https://wismuth.com/elo/calculator.html It results in this formula, which I derived: drawElo = 16 * exp((whiteElo + blackElo) / 2040) The "draw Elo" is the Elo advantage of draw odds. So if you look at the difference in win probability between the true Elo and the Elo with draw odds, that gives the draw probability.
Noflapsedited

Creative math is the best math.

Well done.

Is there a way to take into account the relative number of adult beverages consumed before the game?

That seems to affect outcomes, sometimes. Although history suggests that the not all players are affected to the same extent when playing blitz.

What a great player Tal was.

Creative math is the best math. Well done. Is there a way to take into account the relative number of adult beverages consumed before the game? That seems to affect outcomes, sometimes. Although history suggests that the not all players are affected to the same extent when playing blitz. What a great player Tal was.
joddleedited

Cool stuff! I think age and recent form are worth considering too, especially if you want to predict individual games. Ratings change a lot, and this method assumes that the numerical rating is accurate, but I think many players are frequently over- or under-rated on the basis of age (i.e. on average an older player, say above the age of 40, will perform below their rating, and a younger player will perform above their rating)

Cool stuff! I think age and recent form are worth considering too, especially if you want to predict individual games. Ratings change a lot, and this method assumes that the numerical rating is accurate, but I think many players are frequently over- or under-rated on the basis of age (i.e. on average an older player, say above the age of 40, will perform below their rating, and a younger player will perform above their rating)
MaxReed

Predicting the outcome of chess games is always interesting because even one move can completely change the game. Strong players usually have better chances, but strategy, focus, and experience are just as important. I like reading about chess and improving my skills whenever I have free time. While looking for other ways to relax outside of chess, I came across https://casinohexfrca.net/ . It has useful reviews and comparisons of different gaming platforms, so it was an interesting discovery for anyone who enjoys trying different kinds of online entertainment.

Predicting the outcome of chess games is always interesting because even one move can completely change the game. Strong players usually have better chances, but strategy, focus, and experience are just as important. I like reading about chess and improving my skills whenever I have free time. While looking for other ways to relax outside of chess, I came across https://casinohexfrca.net/ . It has useful reviews and comparisons of different gaming platforms, so it was an interesting discovery for anyone who enjoys trying different kinds of online entertainment.