So why are top players always afraight to play on low tables?????

Thats why ELO doesn't work here.

because it makes no sence to do so... lol

Chance plays a much greater role in kdice than it does in chess. In kdice with a bad start even the most skilled player can be vanquished by the inexperienced. In chess the game always starts the same way. The only advantage is going first a white. This is hardly comparable to a bad set of lots in kdice.

Personally I just use the ranking so I can play with people of similar skill. I had my rank up to 1703 today. I logged out for an hour and when I returned my rank was 1656 or so. I wanted to play with some of the 1700+ players to see how I would fare, but now I'll just have to wait.

Kdice isn't chess. Chess does not have any luck or randomness.

Kdice also isn't a coin flip. Calling a coin flip doesn't have any strategy or skill.

It falls somewhere in between.

You can't measure skill with calling a coin flip and you can easily measure skill in chess. If you try to measure skill in a coin flip everyone would eventually come out even. In chess you can measure skill levels after just one game.

In other words, with a coin flip, over the long run luck is not a factor and the true nature of the game is revealed as 50/50 chance. In the short term you see the luck: good and bad luck streaks.


In kdice short term you see the luck of getting bad position and bad rolls. But over the long run things do not equal out to 50/50. The value that they eventually equal out to is your ELO rating.

The game is new, the sample size is small and due to the luck involved we don't see the value in ELO yet as you would in chess after a few games.

Make sense?

Yes Ryan, That makes a great deal of sense.

I think the best idea posted so far was to remove people from the rankings after a set period of inactivity (ie. 14 days). I'm opposed to a drop in ranking (even a gravitational one) just because it seems wrong to me. Kasperov woudn't want to be forced to play 4 games of chess a month so that he could still be compared to some schmuck from 1936. Removing inactive players form the list makes it possible to see your ranking against ACTIVE players which is really what I'm interested in. AND players with high scores (who may have been without internet access for 4 months) can start playing again and get their name back on the ranking list.

Well, well, I have following Gedankenexperiment.

Say we have 2 Top players, equally good in playing kdice.

Now one plays on every table he just sees fit. He thinks like Ryan said, well the luck factor on the lowp players evens out on time, so the rating reflect my skill in the long run.

Now the second player is like the usual guy here. He thinks, Im not going to play on low entry tables, because if I loose I loose far to much. I will only play at 1900+ tables.

Now who of them both gets a better ELO rating?

Does the ELO rating now reflect that the second plays better kdice than the first? Or is he just better at picking the right opponents in regards to the scoring system to challenge?

the one who plays on the high tables will get better elo (me, getting 2nd on the 0-point tables, only earn 2 points)...

but the other one will got more fun :-)

nevertheless, a nonranked table would be good 4 everyone.

Currently I don't think the scores will even out. The score adjustments are too high to have people stay on a relatively stable score. If the current #1 would loses a game, even to other top players, that would take roughly 1/50'th of his total score, instantly dropping him to 3rd. An unlucky spree of say 3 games (which is not unlikely) would probably drop him out of the top 25, regardless of his true skill.
In my opinion the true number one player should stay at number one within reasonable boundaries of unlucky sprees.
Similarly, someone should not be able to get a lucky spree, take #1 with a relatively high elo rating and stop playing, because that does not resemble his true skill.
Therefore I think the K factor should be way lower on higher rank tables (perhaps a factor 4 or 8 lower than it is currently), to give the scores some inertia in order to factor out some luck. Of course this means that you will gain less too, but in a game with this much luck involved I think that is necessairy to give somewhat accurate rankings.

Alright, I'm going to put this out there.

To anyone who says it's unfair for the top players, I'm going to say your wrong.
My post is in reply to the, "well this is a game based on luck so they could get really unlucky and lose their points," statement.

I'm going to repeat your argument, slightly different.
THIS IS A GAME OF LUCK.

You know how they got to be #1? Yeah, it was luck. #2? Same dang thing.
Everyone got to the top of the ladder based on luck. All top 25.
There is no way to accurately measure rank or skill in a game like this.

Right now there are a group of people not playing, and the point drop makes the game more active. Anyone taking the rank of a game of chance this serious needs to learn that it's just a game here and have fun.

Yes, it can be serious for competition's sake, but people are taking the rank a little too serious imo.

[/End logical rant]

flagingsucks: Do you know which players plays on the 1900? Always the same. Why? They have skill. The top 25 players gained the points when there was another scoring system where you got more points so they were just lucky.

IMO the rating system can only show the big differences of the skill. It can just separate good players (always over 1900), normal players (1700+) and beginners (under 1700).