Forum
unknown3761736
|
Ryan wrote
at 10:11 PM, Friday December 1, 2006 EST
The rating in kdice is based on the ELO rating system which is also used for chess.
Basically you get a higher rating if you beat others with a high rating. This also means you lose more when you lose to people with low ratings. So if you have a high rating you can gain more by playing other with high ratings. There are minimum 200 rating tables which will help you do this. ELO on wikipedia: http://en.wikipedia.org/wiki/ELO_rating_system |
|
Ryan wrote
at 12:45 AM, Wednesday December 6, 2006 EST I like the idea of starting with 1000. If I do this I will bump up everyone's current score 1000 points. Maybe I will do this the same tiem I change the scoring system.
|
|
joby.d wrote
at 1:37 AM, Wednesday December 6, 2006 EST Hello all, after reading this topic I especially liked the 1st suggestion and Pegasus' suggestions. Here is another one:
EP1 (based on rating at start of game) = 1st (+2), 2nd-3rd (+1), 4th (0), 5th-6th (-1), 7th (-2) Round up to the nearest whole number in case of tie. AP1 (based on final placement in game) = 1st (+2.25), 2nd-3rd (+1.25), 4th (0), 5th-6th (-1), 7th (-2) * Draws (Assuming a draw button for 2-3 player games): 3-way draw (+1.5), 2-way draw (+1.75) R1 = R1+(32*(AP1-EP1)),0); I used non-whole numbers because I don't fully understand how the current system works and wanted to ensure people expected to get 2nd-3rd still gain points when they get 2nd-3rd. All the best! |
|
0blivion wrote
at 3:19 AM, Wednesday December 6, 2006 EST I understand why the ELO system is being used. However kdice and chess are very different games... with kdice depending much more on luck, and even in the mid- to end-game, when skill becomes important, luck plays a fairly large role. With chess, ONLY the players' skill is a factor and so the ELO rating works, but with kdice, i think the relative importance of luck as a factor makes the ELO system too biased against highly rated players.
Just my 2c. thnx for ur time XD |
|
Pegasus wrote
at 5:59 AM, Wednesday December 6, 2006 EST Oblivion that suggests ratings are too volatile, and could be improved by reducing the K factor from 32 to something lower. It would take longer to earn high ratings tho.
|
|
Nuklear wrote
at 11:17 AM, Wednesday December 6, 2006 EST 2 Ideas:
- How about adding bonus points for finishing off an opponent. You could also subtract points for a player who gets 2nd or 3rd and does not finish off any opponents. -I feel its time to add a 300+ table and maybe even a 400+ |
|
Guntfred wrote
at 11:32 AM, Wednesday December 6, 2006 EST I am not an expert in ELo Rankings, but this version seems to be broken as installed.
If a player starts at 0 and can not lose points, but can "give" points to other players, the average points per player will go up over time. One reason to start the chess ELO ranking not at 0 ist this . Every point I win HAS to be reduced from another player, or the system will not work. |
|
Ryan wrote
at 12:13 PM, Wednesday December 6, 2006 EST Guntfred:
I'm thinking about raising the starting rating. But I'm curious what is supposed to happen when a player, who started at 1000, hits 0... do they go negative, are they still aloud to play? |
|
empath wrote
at 3:31 PM, Wednesday December 6, 2006 EST That's a very good point. A player could create 7 accounts, start a game, join with all 7 accounts and concede with everybody but himself
|
|
Measure wrote
at 4:33 PM, Wednesday December 6, 2006 EST Ryan, I don't think that anyone could ever get down to zero if you started everyone at 1000. a player who got down to even 500 points would have a hard time losing any points to players that were at 900 or 1000.
Ok, maybe someone could, but it would probably take months or years of efforts to do so. You just don't lose enough points per game when your rating gets low enough, and there won't be a lot of people with low enough ratings to enable a continued slide down to zero. |
|
Precambrian wrote
at 5:56 PM, Wednesday December 6, 2006 EST I think the chess approach of starting everyone out at 1600 makes more sense than starting at 0.
|