| |
BGonline.org Forums
Questionniare
Posted By: Barry Silliman In Response To: Questionniare (Rick Janowski)
Date: Saturday, 20 February 2016, at 11:44 p.m.
Hi Rick.
It looks like what backgammon is using is considered a "modified Elo" and I don't think it uses that K-factor at all.
e.g. here I see the K-factor you mention, this was new to me: https://en.wikipedia.org/wiki/Elo_rating_system#The_K-factor_used_by_the_USCF
Here is how I think most Elos in backgammon are calculated. Of course for servers such as GridGammon, etc. only the programmers would know for sure unless they publish the formula: http://www.bkgm.com/articles/McCool/ratings.html
Notice in this that there are multipliers given to players with lower experience points until they reach 400. (See the Player Experience section in the immediately preceding URL).
I think that both Gridgammon and backgames.org (which is where the USBGF and many other clubs store statistics) use this formula, as well as the multiplier. (I don't know this for a fact but I strongly suspect it).
A couple years ago, more or less, I received an email from a reader here suggesting that this multiplier effect in backgammon is less valuable in backgammon due to the luck involved.
I sort of agreed with that and the Elos I publish in the USBGF PrimeTime Backgammon Magazine do not use these multipliers. In a sinister and devious move, I switched the ratings from using the multipliers to not use them and I didn't tell anybody, to see if anybody would notice.
The new ratings for the leaders were significantly lower without the multiplier. In well over a year, I received a total of two inquiries as to why the numbers were lower. My conclusion is that the vast majority of people just don't care all that much about the published ratings.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.