Monthly Rankings that Take Games Played Into Account

Have an idea to make Quadradius better?

Monthly Rankings that Take Games Played Into Account

Postby OncoByte » Sat Feb 28, 2009 10:05 pm

In brief – I propose a simple mechanism for discounting the rank of players who have not played many games compared to everyone else. I outlined a model for this in TheThrill's thread where he bitched about the ranking system.

The essential argument is that people who play few games have an advantage in the current ranking system - and a strong disadvantage to play more games once they take the top spot. Essentially, they may get lucky and win just enough games to get themselves ranked at the top, but were they to play more games, their true, lower win % would come out and they would cede the top spot.

An ideal ranking system would do the following:
1) Allow for rank lists that span several different time periods (daily, weekly, monthly, yearly, all time ...)
2) Take into account the quality of the player you defeat when adjusting your rank
3) Recognize that the more games you play, the more accurate your rank metric will be (less error) by incorporating this into the rank metric!

ELO can do #2 really well, but fails to do #3. It takes many games to arrive at an accurate ELO score for a player, therefore it is good for long term rank lists, but not short term ones where few games are played.
My proposal does not do #2 at all, but it could later be combined with ELO to tackle all three.
In brief, it assigns an “error” or penalty to your win % (our current rank metric) that is based on the number of games you have played and the number of games the x-most prolific players have played.

The properties of the error function are such that the penalty to your rank score is greater during your first few games and shrinks as you play more AND this penalty is higher if most players have played more games than you and becomes negligible if you have played more games than most.

I tweaked the error function to be a little less drastic than the one I listed in the prior thread. The only variables involved are the number of wins (wins), number of games played (games), and mean number of games played by the top x-number of players (meangames). Here is the revised formula for the Adjusted Rank Score (ARS):
Image
Increase the value of constant I set at 4 and the penalty becomes less harsh. I think somewhere between 3-5 is ideal.

Sticking with the value of 4, here's a graph that demonstrates its effects:
Image
This graph shows you how the Adjusted Rank Score works. In this example, a player plays his first game when the average number of games played by everyone else is 100. This new player wins every other game after losing his first and plays a total of 100 games. His win % converges on 50% (blue line on graph). The green line shows his Adjusted Rank Score – the metric that would be used to rank him on the leaderboard. After his first win, his ARS is 26%, just over half of his 50% wins. This is low because this player has played very few games compared to the mean. As he plays more games, his ARS grows closer and closer to his win %. By the time he plays 100 games, the fraction of his win % that is discounted by the error function is a mere 1.8%. The red line shows the number of percentage points discounted from the win % by the error function. The purple line shows what fraction of the win % this represents. You can see from the graph that when a player has completed half as many games as everyone else, his win % is discounted 12%. A win % of 80%, for example, would be adjusted to 70.4%.

In actuality, most players complete their first few games at the start of each month. Since the mean number of games at this point is low, everyone’s ARS will be close to their win %. If they keep up with everyone else, this will be true for the rest of the month. Those players that lag behind will see a penalty applied to their rank score that grows as they fall further behind. Playing twice as many games as the mean only buys you another 2% increase in your rank score (from 50% to 51%, for example).

Here’s what would happen to the leaderboard (as it stood at 3:50 pm on this last day of February) if we applied this rank metric using the mean number of games played by the top 13 players:
Image
TheThrill would come out on top. Rhox, Fluffballs, and tony see almost no difference between their win % and ARS since they have played many more games than the mean. I fall to fourth place. Respectable, but appropriate given that I have played few games. If I really wanted to claw my way to the top, I’d have to play a lot more. As I play more games, each prior win counts more since my penalty decreases. Instead of rank sitting, I now have incentive to play more games. If I don’t, I see my score and rank erode further as everyone else plays more.

I also have a simple way of melding short term rank lists with the more long term ELO-based ranking system, but I can talk about that later.
User avatar
OncoByte
Orbgasms
 
Posts: 1136
Joined: Sun Mar 18, 2007 9:40 pm
Location: Location, Location ...

Postby driven2sin » Sun Mar 01, 2009 3:30 am

how can anyone argue with all those color graphs.. like ross perot all over again
User avatar
driven2sin
Orbgasms
 
Posts: 1766
Joined: Sun Jan 07, 2007 3:13 am
Location: casting judgement from beyond your sight

Postby TheThrill » Sun Mar 01, 2009 4:46 am

IMO its a great idea.

but even if this was the worst idea ever, hey u put me at #1 how could i complain?
TheThrill
Quadradius Master
 
Posts: 195
Joined: Wed Jan 24, 2007 4:37 pm

Postby Rydash » Sun Mar 01, 2009 11:09 am

Onco, I must disagree with you in the first line.

I don't find this to be "simple". :P

But on the other hand, I didn't actually READ this...just looked at picture....
Changed for the sake of mobile browsers everywhere.
User avatar
Rydash
QTorus Stimulator
 
Posts: 758
Joined: Thu May 17, 2007 4:43 pm

Postby blamin8or » Tue Mar 03, 2009 8:09 am

this is awesome. I assume you're proposing we base everyone's rank on the mean games played by everyone and not just the top 13, it'd be interesting to see the whole top 100 reorganized according to this. I'm interested to see how this would affect players falling off the board though. As for a simple way incorporating this into ELO, i do not believe you.
blamin8or
Inebriate Radial
 
Posts: 16
Joined: Mon Nov 17, 2008 6:12 pm

Postby dadada » Tue Mar 03, 2009 11:56 am

I don't like the idea of using the mean games played by and number of members on the list as the base for the error percent. There's no reason why someone playing 250 games in a month (creating such a skewed distribution) should hurt the raiting of someone playing an acceptable 75 games.

Instead I suggest a base that would be easier to calculate, and allow for comparing between months (where the mean games played could possibly change - though I have not collected data on that). Use 3x (or 4x) where x is the number of days so far in the month. People averaging more than 6 games a day will still have percent error bordering on miniscule, and this way people can know exactly how many games they need to play to have a specific error percent (if of course they put in the effort to calculate it).

:Insert Fancy Graphic Here:
"take sharp pointy sticks and make alot of holes"
dadada
Recruited Soldier
 
Posts: 19
Joined: Sat Mar 08, 2008 11:41 pm

Postby OncoByte » Tue Mar 03, 2009 3:06 pm

dadada wrote:I don't like the idea of using the mean games played by and number of members on the list as the base for the error percent. There's no reason why someone playing 250 games in a month (creating such a skewed distribution) should hurt the raiting of someone playing an acceptable 75 games.

Instead I suggest a base that would be easier to calculate, and allow for comparing between months (where the mean games played could possibly change - though I have not collected data on that). Use 3x (or 4x) where x is the number of days so far in the month. People averaging more than 6 games a day will still have percent error bordering on miniscule, and this way people can know exactly how many games they need to play to have a specific error percent (if of course they put in the effort to calculate it).

:Insert Fancy Graphic Here:

Good suggestion - you could definitely use a recommended number of games played instead of a mean as the basis for establishing the error.

If you do use the mean of the 100 most active players, I suspect that no single individual would have much of an impact on the mean. In your example, a player with 250 games under his belt would only raise the mean by 2.5 games. Balance that with a large tail of people playing 20 or so games and the mean won't get out of control.
User avatar
OncoByte
Orbgasms
 
Posts: 1136
Joined: Sun Mar 18, 2007 9:40 pm
Location: Location, Location ...

Postby driven2sin » Wed Mar 04, 2009 4:35 pm

just like back in ol' England, some measurements were based on the King (if not, well, change a wikipedia page to make it so)

So here, in QR, we should base this metric on our King.. The Learned Pig.


So we start off with a base number. The weight of his bowel movements on Easter Sunday. (a steamy 15 pounds)

Then you must divide this by the amount of porn files on all of his available computers as of June 6, 2006 (45,800)

Times that by all his sleepless nights during childhood (600)

Divided by the amount of times he has boned a teenager (1)

And then multiplied even further by the amount of pills he has taken in the 90's (10,000)

Then count the amount of weeks from that number as a year, to the death of Bob Hope and then subtract from that number the year of America's Independence (all that leads to 200... trust me)

Then divide that by his balls (2)

For a final number of 100.

That is the amount of games that should be used as the Official QR Ranking Determinator per monthly ranking
User avatar
driven2sin
Orbgasms
 
Posts: 1766
Joined: Sun Jan 07, 2007 3:13 am
Location: casting judgement from beyond your sight

Postby Rhox » Thu Mar 05, 2009 7:23 pm

POTY.
Oncobyte wrote:This is a luck containing game. We like that. The randomness means that you have a tremendous variety of options and almost every game you play is different than the last. The same strategy doesn't work every time. Sometimes, you will lose entirely because of bad luck. Play another "hand" of QR and get over it.


2007 - 1073-482, 69%
2008 - 477-269, 64%
2009 - 499-228, 69%
2010 - 514-233, 70%
2011 - 46-45, 51%
2012 - In progress

CAREER: 2715-1313 (68%)
User avatar
Rhox
Orbgasms
 
Posts: 2459
Joined: Mon Mar 26, 2007 12:30 am
Location: NC

Postby OncoByte » Thu Mar 19, 2009 9:16 pm

Current rank list using my adjustment for games played:

Image

Notice that tony is ranked over sanzo. I think this accurate. For sanzo to beat tony on the current leaderboard and play as many games as tony, sanzo would have to:
1) Play as many games as he has already this month (29)
2) Win 24 of these
3) Then win 9 out of an additional 11 games.

This is seems like a tall order to me.

As it stands now, sanzo can safely rank sit. He knows he has played enough games to stay on the leaderboard through the end of the month (unlike poisonivy). In fact, tony's greater number of games played is actually a disadvantage! Each win tony gets at this point is worth incrementally less than it would be for sanzo.

For example, if sanzo and tony each win 5 out of the next 5 games they play, sanzo's win% goes up by 2.5% while tony's only climbs 1.3%.

While it is true that a loss would cost sanzo more, sanzo doesn't even have to take a chance he might lose! He can let tony take all the chances as he tries to catch up. If instead, sanzo saw his rank score fall as more games were played by everyone else, he would feel forced to play or risk losing his rank. If he got to where he is now by luck, his luck might run out and he could suffer a painful loss. Of course, he might skillfully pull off a few wins, in which case, he deserves his place in the sun.

This still doesn't address the quality of opponents, but it does help take out the effect of lucky streaks.
User avatar
OncoByte
Orbgasms
 
Posts: 1136
Joined: Sun Mar 18, 2007 9:40 pm
Location: Location, Location ...

Something like this should be done

Postby Shmoo » Wed Dec 23, 2009 2:16 pm

For me it is not about accuracy (though that may be a consequence). It is about not punishing players for playing good players.

With Elo or some variant I am punished for playing underrated players.
A game should have enough luck that the worse player can win SOME of the time!
Shmoo
HotSpot HotShot
 
Posts: 27
Joined: Wed Apr 25, 2007 12:19 am

Postby OncoByte » Sun Mar 21, 2010 4:18 pm

It's been a year so I'm reposting what the leaderboard would look like if the number of games played were a factor in the rankings.

Quick recap -
1) Players who play fewer than the mean number of games get a penalty deducted from their win percentage (details in the first post).

2) The further your game total is from the mean, the higher the penalty.

3) Ranks are recalculated after the penalty has been applied to the percentage of games won.

Here's the recalculated rankings for this month:
Image

The '% Error' is the penalty applied to percentage of games won. The mean is only the mean of the top 13 players.

Hey - I'm first! No way! I never realized that was going to happen!
User avatar
OncoByte
Orbgasms
 
Posts: 1136
Joined: Sun Mar 18, 2007 9:40 pm
Location: Location, Location ...

Postby driven2sin » Mon Mar 22, 2010 6:27 pm

it's still bollocks

people can just play crappy players or their fake accounts... you need ELO or something similar that weighs games vs strength of opponent but have a cut off so if two people with a large difference in ranking, it just gives a small move to either side..

anything else.... bollocks!
User avatar
driven2sin
Orbgasms
 
Posts: 1766
Joined: Sun Jan 07, 2007 3:13 am
Location: casting judgement from beyond your sight


Return to Suggestions

Who is online

Users browsing this forum: No registered users and 1 guest

cron