Classicists League Table

[raphtown]

Hi all,

The league table has been updated to include the Elo style rating system. Please note ATM if you have not completed a game you will be rated 1500 however you will rank below anyone who has completed at least one game. (this can be changed if agreed upon)

You can view the League Table here: League Table

Also a list of completed games and results (and the order of their completion which is important for calculating new ratings accurately) is here: Completed Games

Suggestions of what information to display or not to display are welcome.

Uh I don't see the ELO ratings in the link you provided.

[InterMPC]

Sorry Raphtown,

Thanks for highlighting the error, I continued editing the spreadsheet to automate some of stats and accidentally changed the ELO rating column. I have fixed this now.

Mapmaker, Games Played, Wins and Loses are now automated in the data tab. You just need to accurately complete the results tab, and then the name column and ELO rating column in the data tab. I'll see if I can automate the ELO rating change for you.

Anyone else see any errors just let me know!

[InterMPC]

In looking at the table linked in the first post, it seems to me that it is in dire need of an application of the KISS principle.

Quite simply, there is no point in having a category or column for 6 or 7 way draws. These do not happen ever in a game that has been properly completed. Such a result would equate to a consensus among all players of "we don't want to play anymore so we are all giving up simultaneously". At that point, we might as well "unrank" the game entirely.
The League table will look much better with fewer columns.

Ceebs,

I agree in principle that there should not be a 6 or 7 way draw. The way the Elo system currently works is you would probably lose points if you agreed to such a draw, therefore discouraging players from agreeing to one. I also agree that the table would look better with less columns.

As of the last few months the in game mechanics actually allow for 6 and 7 way draws to happen and is my sole reason for including them in the table. I can change this quite easily if that is what the group wants.

[raphtown]

Looking at the initial scores, I think it might be wise to increase the scaling so that losses and wins make a bigger difference in scores. What does everyone think? Should a solo against equals only yield 25 points?

[InterMPC]

If every one agrees, it is easy to change. The k-factor controls the scaling. So we could increase it from 30 to whatever we feel is appropriate.

[Waterice man]

A couple of questions;

Will forum games count on the scoring (1900 is a Classicists game)?
Assuming the answer to the above is 'yes,' what would the weighting be if a forum game was run with more (or less) than seven players?

[InterMPC]

Good questions.

It is not easy to calculate new ratings for a game with more or less than 7 players with the current spreadsheet. However if the answer to your first question is yes, then I guess we will need to change the spreadsheet to be able to accommodate these types of games. Including deciding on what the weightings should be.

[raphtown]

Tentatively, we could include forum games. 7 player ones like 1900 would have no problem being included, but we'd have to figure out a system for including the other ones without changing too much to our existing system.

[Mapmaker]

I would say that for now, since the forum game running is 7-player, include it. It's just another kind of variant, and we've so far included chaos variants and probably a few others as well. But for multiple-player, it depends on whether we can make the maths work out.

As for weightings, I think the problem is just that games take a long time. ELO was designed for leagues with games being completed more regularly, was it not? In which case it would make sense to increase the weightings.

[InterMPC]

So how much change do people think is appropriate?

Should a solo among equally rated players be worth 25? 50? 100? Any other value?