3 hours ago, Thrice said:

 

At least CD stand by their claims and methods rather than changing their mind half way through the season when their prediction fails.

I don't think it's a good thing.

If you cannot re-evaluate things and change your mind you are doubling down on your own tripe.

The champion data guys formula is not a great one. They haven't identified this and haven't changed it in any meaningful way. They probably don't care, they are likely making money hand over fist from the fools running the AFL.

Anyone who dismisses the Champion Data analysis should do a bit more reading to understand the basis of their ratings. There are a tonne of stats that go into the formulas they use and it really does make sense if you bother to understand it. Theres a reason clubs shell out thousands a year to get the data direct.

Anyone interested in understanding the game at a higher level than is presented in the Herald Sun should pick up the AFL Prospectus (where the Sun got their ratings from).

And yeah, based on his performance over the past two years Pedersen is above average (in the top 35% of key forwards in the comp). While someone like Viney has only performed average (between 36-65% of all midfielders). No doubt Viney will hopefully and should go up to the above average category this year. The ratings also show McDonald is one of the best key defenders in the comp and is still only 24yo. 

11 hours ago, P-man said:

The players just rated Jones as the 33rd best player in the league heading into the season. Likewise Lewis and Viney were rated inside the top 50. Not bad for "average" players.

 

I'd dig deeper to figure out how Petracca was ranked above average in his position over the past two years but I don't care enough to do so.

 

As always this is a waste of time. We'll have a top 6 midfield in 2017. That much I do know.

 

 

I assume Petracca is categorised as a half forward/mid rather than a pure mid. Hence the difference in ratings.

10 hours ago, rjay said:

I thought I might rank Champion Data...

Rating: Poor to very poor

Ratings based on previous years really don't tell the story of a young list so they've about as useful as...

They also have "relative ratings" which measures players against all other players in the same position at the same age since 2010. So for instance Tom McDonald performs something like 65% better than all other 24yo key defenders over the last 7 years. Oliver is another who is about +35% compared to other mids his age since 2010. So they do have a measure that tries to factor it in.

The system of value given to the raw data might well be reasonable but I have an issue with the determination of player positions and the potential lack of reflection for multiple or specific individual roles.