Top 100 - Champion Data

[sfan]

http://www.heraldsun.news.com.au/foo...E19742,00.html

2004.......................2003
478 Stephen Doyle *
407 Paul Bevan *
490 Jarrad McVeigh *
365 Heath James *
362 Amon Buchanan *
310 Nick Davis 168
252 Nic Fosdike 154
177 Jason Ball 93
397 Adam Schneider 240
167 Adam Goodes 20

[Wil]

I guess now is the time to ask, what are the player rankings based on? Is there a link or can anyone put up the algorithm? I presume this is the same as the channel 10 "top 10" players during their broadcasts.

Please stop overusing the word "Champion"!

[sfan]

The player rankings are calculated by computer. Each players statistics are given a positive or negative value.

http://www.championdata.com.au/faq.html#rankings has the following info.

The Player Rankings formula is based on the Official AFL Statistics, and is calculated by computer.

The rankings formula is weighted in favour of effective use of the ball (loaded in favour of good kicking) and various types of possession gainers (loaded in favour of winning disputed ball).

Evidence based on extensive research into winning and losing factors in AFL games (1,110 games in total from 1999 to 2003) by Champion Data and the Swinburne Univerisity School of Mathematics is unequivocal ? effective kicking and winning disputed ball in critical situations is what counts most.

Accordingly, the AFL Players Rankings are geared to reward these winning factors. The formula is scientifically derived and objective.

Rankings samples

An effective long kick has to travel more than 40 metres to a 50/50 contested or better for the team. The computer adds four rankings points for each effective long kick.

If the long kick goes to a teammate for an uncontested possession, research shows this is very valuable for the team and the computer adds five rankings points to the players tally.
Effective short kicks are less than 40 metres that result in uncontested possession to a teammate. The computer also adds four rankings points for these.

Clanger kicks are when the kick goes down the throat of the opposition and the computer deducts eight
rankings points from the players tally because this hurts the team badly.

Ineffective kicks occur most often where players kick short to a contest and the computer ignores these by
giving a zero rankings value.

A goal receives eight ranking points in addition to the four ranking points for the effective kick, and the
points awarded for the possession type resulting in the kick and goal.

Another example of weighing in favour of game breaking statistic is contested marking. A contested mark results in four rankings points to the player.

A contested mark from an opposition kick is very important and the computer adds eight rankings points for these.

Publishing formula policy

The rankings value for each statistical category has been endorsed by the AFL and the Swinburne University School of Mathematics.

There are a total of 57 individual statistics categories where the computer attributes either a positive or negative value.
Several of these categories, such as ?marks from opposition kicks? ranking points, are derived from the computer linking a series of composite statistics.

Given the unique intellectual property contained in the rankings formula and the degree of underlying computer logic applied, Champion Data does not publish the full rankings formula.

[Barry Schneider]

http://www.championdata.com.au/files/rankings.pdf

This gives an overview but it does say that the rankings are their IP and therefore they are not willing to divulge the full criteria that the rankings are based upon.

Beat me to it.

[sfan]

p.s see previous thread on bigfooty

http://www.bigfooty.com/forum/showthread.php?t=129117&page=1&pp=40