ELO most certainly DOES ignore the rain and temperature. ELO looks at rank. Rank looks at garbabe to determine skill. Thus, ELO is flawed by Rank's flaw. Rank's flaw is that it measures games won, not skill of players.Terralthra wrote:QUOTE (Terralthra @ Nov 23 2006, 02:06 PM) http://en.wikipedia.org/wiki/Bayes'_theorem
Elo doesn't 'ignore' the rain and temperature. It takes all players on a team into account when calculating win percentage, as well as how long each player was on team. If, to use your analogy, you took a whole bunch of samples that include the amount the grass grows, the sunshine, the rain, the temperature, soil density, arability, pests, etc., you could in fact build up a statistical model for how the grass will grow under any given combination of contributory factors.
Rank Discussion
A panel of expert players? Who picks this panel? Where do they find the time to evaluate each and every player in allegiance? Not to mention continual re-evaluations of each player (once a month would be the bare minimum). What happens about differences of opinion?Wasp wrote:QUOTE (Wasp @ Nov 24 2006, 02:43 PM) I'm glad you asked /smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" />
First off, in order for autobalance to work, the ranking system must be an accurate accounting of one's skill level compared to the levels of others (relative rank). This must be done by a panel of expert players who can truely evaluate the skill level of each and every individual.
If you try to mathmatically arrive at a skill level using statistics, you'll only get a rank that represents his statistics, not skill.
A statistical system based on mathematics is the only way this can practicably be worked.
Spidey: Can't think of a reason I'd need to know anything
You've got it backwards....this is a mathematical system based on statistics unrelated to individual skill.Raveen wrote:QUOTE (Raveen @ Nov 24 2006, 08:02 AM) A panel of expert players? Who picks this panel? Where do they find the time to evaluate each and every player in allegiance? Not to mention continual re-evaluations of each player (once a month would be the bare minimum). What happens about differences of opinion?
A statistical system based on mathematics is the only way this can practicably be worked.
This is the only way you'll ever determine rank........
The evaluation committee is voted in innitually. Could be lets say 10 of the "top" players total with 5 of these members necessary to create a score of another player. Those top players would have to first rank each other by voting a score. The average score becomes the players rank. Then each and every allegiance member would have to be individually scored by each of the council members. The average score becomes that players rank. Once your rank reaches a set high level, you are eligible to become a council member and score other players. Thus you create a panel by encouraging players to achieve a higher rank. You also thwart the possibility of a player altering his own rank for the purpose of negating autobalance. New players would have to automatically be up for re-evaluation more frequently than a vet would as the noobs skill would probably be changing at a rate faster than a vet's. If a noob is detected as a vet, his rank will be corrected by the committee shortly.
If this was in place along with autobalance, then it would work.
Even if this were only done initially, it would be light years ahead of the current system in determining rank.
well, i like jimmy, so he can have a rank 25, obviously i get a 30, deng annoys me, so he can have a 7, giving yanlin a 23 will make some nice rants, i accept paypal, $1 per rank, pook.. lets try this random number generator for fun /smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" /> ...
You don't get to score yourself. Average scores are used, so your extreem score would have less impact. A score of each players rank could only be increased by a limited score, preventing a council member from pushing a vet 1 to a vet 5 overnight or visa versa. The more participating (voting) council members, the more accurate the scores become.madpeople wrote:QUOTE (madpeople @ Nov 24 2006, 09:24 AM) well, i like jimmy, so he can have a rank 25, obviously i get a 30, deng annoys me, so he can have a 7, giving yanlin a 23 will make some nice rants, i accept paypal, $1 per rank, pook.. lets try this random number generator for fun /smile.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile.gif" /> ...
Do you think your statistics of historically won games really represents your ability? Will it ever? Does your school grades really represent how successful you will be in society? If you're now sucessfull in life, could you retake all those tests and get straight A's? Success should be measured by a human being, just as rank should be.
Last edited by Wasp on Fri Nov 24, 2006 4:37 pm, edited 1 time in total.
-
MadAccountant
- Posts: 2610
- Joined: Mon Jul 07, 2003 7:00 am
- Location: Ontario, Canada
a) who's the committee to determine the top 10 players?
being a top player doesn't necessarily mean good judgement of other people, just that they are good at a game.
Rank each player 1-10, perhaps 1-20 but the greater the scale the harder to assign.
But then when are ranks revisited?
Perhaps this can be done to assign the initial ELO values and then use ELO to see what happens. You may just want to use a scale of 1-5 representing 1100-1500 (or something) just to quickly get through it.
Having a perpetual rankinig system though is a headache people will quickly tire of.
being a top player doesn't necessarily mean good judgement of other people, just that they are good at a game.Rank each player 1-10, perhaps 1-20 but the greater the scale the harder to assign.
But then when are ranks revisited?
Perhaps this can be done to assign the initial ELO values and then use ELO to see what happens. You may just want to use a scale of 1-5 representing 1100-1500 (or something) just to quickly get through it.
Having a perpetual rankinig system though is a headache people will quickly tire of.
-
Terralthra
- Posts: 1748
- Joined: Fri Nov 18, 2005 8:00 am
- Location: San Francisco, CA, USA
...if you honestly are saying with a straight face that winning has nothing to do with skill, then I think this thread should just be locked. If you're not even willing to grant the basic assumption of any statistical method ( a more skilled player will win more often ), then I don't think there's any point in continuing this discussion.Wasp wrote:QUOTE (Wasp @ Nov 25 2006, 12:55 AM) Rank's flaw is that it measures games won, not skill of players.
Last edited by Terralthra on Fri Nov 24, 2006 6:09 pm, edited 1 time in total.
No, I'm not willing to grant the basic assumption "a skilled player will win more often". This would only be true in a 1 vs 1 game. To apply that logic to this game is ridiculous. And if you want this thread locked, ask someone to lock it. I doubt anyone really wants to supress opinions of this system.Terralthra wrote:QUOTE (Terralthra @ Nov 24 2006, 11:08 AM) ...if you honestly are saying with a straight face that winning has nothing to do with skill, then I think this thread should just be locked. If you're not even willing to grant the basic assumption of any statistical method ( a more skilled player will win more often ), then I don't think there's any point in continuing this discussion.