Forum
A Message To Ryan
|
'dasfury' wrote
at 6:23 AM, Thursday April 16, 2009 EDT
I have an idea for a new scoring system that in my opinion could change kdice from a great game to a fantastic game. I was thinking that instead of the leaderboard being based on the amount of points accumulated each month it should be based on the ppg of players. This would solve a lot of problems with the current system. Players would now strive for the highest position possible and the problem of v-flagging would be significantly reduced. It would dramatically increase the emphasis of the game on skill rather than how many games you can play a month. And it would be a pretty easy system to impliment too as you would still have all the same tables, scoring system would not need to be changed just the way people gain points. Or better still having two seperate leaderboards one for ppg and one for accumulated points and seperate trophies for each at the end of the month. That way everyone who cant play 600 games a month has something to aim for and I think both systems would work great together. The only problem that I can see with the ppg system is that we would need to decide a minimum number of games for the month to be played to be eligible for the leaderboard. I know you are busy with the kdice trainer at the moment but I'm hoping you could post your reaction to it. Any other views on this system would be welcome.
|
« First
‹ Previous
Replies 11 - 16 of 16
|
the full monte wrote
at 11:50 AM, Thursday April 16, 2009 EDT one of the reasons that ELO allowed for players to sit after just a handful of successful games was because each win as 1st gave you 1/1=1 points, whereas each win while 5th gave you 1/5=0.2 points, etc, etc. This is a HUGE gap in point-gaining. Perhaps if the difference wasn't so dramatic, it would balance out the numberOfGamesPlayed vs skillRequiredToWinGame seesaw. So, instead of points added being = 1/eloRank, make it something like sqrt(1/eloRank) or 1/(eloRank+2) or something like that.
Also, I would love to hear skrum's advice. And the wrench in the system is how to incorporate tourney finishings into elo-based algorithms? |
|
Homer Simmpson wrote
at 4:16 PM, Thursday April 16, 2009 EDT well you could also do something like have your score based on say PPG*(games played)^1/3 this way yesterday when i was at 600 ppg with only 6 games my score would be 1091 whereas les who has 368 ppg and 80 games played would be something like 1586. and then on top of that put on a minimun number of games like say 30 or so.
this way PPG would be the most strongly emphasized while high ppg over several games would be secondary. Shad in feb would have gotten 1587 which is just barely above les is right now and shad had 203 ppg (which isnt bad) and 478 games (thats a lot) also with a 30 games min. you whould see ppg multipliers ranging from 3.1 (30 games) to 10 (1000 games) and if you think thats too large a difference then try it with games^1/4 2.34 to 5.62 or games^1/5 1.97 to 3.98 top 6 scores from January 09 using each of these 3 multypliers g^1/3 1st. MadHat_sam 2561 1776 1426 2nd. Shevar 1384 679 679 3rd. Smumustang 1109 660 660 4th. Potato27 1041 472 472 5th. Shadolin 834 375 375 6th. Orlafede 804 389 389 also this would be sam's ranking over the last 6 months using the ppg*games^(1/5) formula Mar. march deserves to be expanded Feb. 1st 1405 fede 2nd with 1012 Jan. 1st 1425 Shevar 2nd 680 Dec. not enough games dasfury 1st with 767 and monte 2nd with 706 Nov. 4th 251 jpc4p 1st with 704 Oct. 3rd 980 dasfury wins 2057 FriendlyLassy 2nd with 1037 march 1st. 'dasfury' 1929 2nd. Das Jr. 1576 3rd. orlafede 1156 4th. MadHat_Sam 1010 5th. Shadolin 697 6th. Parian 690 7th. Joselito Michuad 596 anyway i like the idea of a more ppg with games played multiplier based scoreboard for month to month ranking. and then maybe a permanent ppg leaderboard |
|
|
Homer Simmpson wrote
at 4:37 PM, Thursday April 16, 2009 EDT another idea would be to do say (ppp^(3/2)*games)/1000
march would look like this 1st Das Jr. 2145 2nd 'dasfury' 1547 3rd Shadolin 1383 4th Joselito M. 1333 5th MadHat_sam 1323 6th Parian 995 7th orlafede 800 and of course you could play with what exponent you want to raise ppg to. you could also combine this with a games played multiplier to look like say something like PPG^(3/2)*Games played^(1/2) or PPG^(3/2)*ln(Games Played) which i kinda like that last one it gives some incentive to play more games but as you play more games it gives less incentive to play more games then it did previously. you would probably want to play around with what base you use for you logarythimic equation though |
|
Leek Step wrote
at 5:31 PM, Thursday April 16, 2009 EDT bring back elo would be fun! just for the kicks!
But really the fundamental difference between kdice now, and kdice when it was good, is a larger rotating pool of competitors. Makes the games more unpredictable when you don't have to sit with the same people all the time. Easiest way to do this is the BONUS ROUND! Win a game, advance to BONUS ROUND to play 6 other table winners. That would make it impossible to rely on the same helpers from game to game, in the high pts games. |
|
detenmile wrote
at 12:36 AM, Friday April 17, 2009 EDT i had another idea but i forgot it. sorry
|
|
BLUNTMAYNE420 wrote
at 1:53 AM, Monday July 25, 2011 EDT bumb
|
