AP player rankings

[Peat]

Somebody has developed an algorithim to measure player performance and put up a chart of the best 30 players in each position in the AP over the whole season based on it. Pretty interesting stuff, even without fully understanding how performance is being measured.

http://www.rugby.net/demo-responsive.html

 

[Which Tyler]

That looks fantastic - now to try and work out what the hell it all means - and that may take some time.

First thoughts though - how the hell is Roko not getting caps?

 

[Mumbles119]

Second thoughts though - how is Tom Wood ranked so high?

 

[TRF_Olyy]

Sale well represented,
Sam James the best centre in the prem

 

[Peat]

One big thing to note with this is that if a player is playing in a different position for a game, or playing an international, then he receives no points for his ratings in that position while the other guys do. So if you look at Robshaw, he's ranked as the second best flanker in the league until he misses 9 games (presumably mainly international duty). Haskell's 6th prior to that point. Piutau's ranking is banjaxed by being moved around positions.

And so on.

As for Wood - looks like he had a strong end to the season.

 

[Don't Skip Leg Day]

You've all got to remember than Tow Wood is a good rugby player, good brain, good work rate etc doesn't make many mistakes so is a good stat player.

It's only by human judgement we know he's not a top international player because of the lack of his carrying game and lack of physicality. Stats arnt everything.

 

[TRF_stormer2010]

I've always been interested in these types of rankings and have always had an issue with how these things get presented. If only I had the time and Excel skills.

 

[Which Tyler]

One big thing to note with this is that if a player is playing in a different position for a game, or playing an international, then he receives no points for his ratings in that position while the other guys do. So if you look at Robshaw, he's ranked as the second best flanker in the league until he misses 9 games (presumably mainly international duty). Haskell's 6th prior to that point. Piutau's ranking is banjaxed by being moved around positions.

And so on.

As for Wood - looks like he had a strong end to the season.

I'd noticed the international absences - which is strictly correct I guess; but for match days, could easily include stats earned there - after all, surely meters run etc are harder to earn at that higher level. Equally, It shouldn't be all that difficult to include stats earned in a different position.

My excel skills could probably cope with putting something like this together; but not my access to stat.s; time; inclination; or general computing skills to put it up somehow accessible to all.

 

[TRF_stormer2010]

I'd think take the AP only but allow for a slight modifier 'rewarding' consistency rather than one-off performances. I'd also either have a slight modifier 'rewarding' plyers who have played more matches or a modifier somehow taking opposition performances/rankings into account.

I have a few ideas about how to make something of this sort into a marketable game of sorts if anyone has the time to listen and willingness/ability to take it further? PM me.

 

[Which Tyler]

I'd still think that the internationals "deserve" some points for playing at a higher level; and that something should be done for someone like Devoto with 208 points to be 28th "best" FH AND 935 points to be 22nd "best" centre - for that case; I'd simply add them together, and shove him in his primary position - to make him 17th "best" centre.

 

[RedruthRFC]

Am I missing an explanation of how these scores are reached? I couldn't find one. Without this, they are somewhat meaningless IMO as we have no idea of whether they are something resembling a decent measure of how effective a player is devised by clever statisticians and rugby people or just a basis for a fantasy league scoring system slung together by a numbers nerd. The fact that they group all props together and the shortcomings that TRF_Peat has pointed out make me think that the former is latter is more likely to be the case.

 

[Which Tyler]

http://www.rugby.net/s/lmpAyv5MR4Q#Content:contentId=5171798064758784

 

[RedruthRFC]

I'd still think that the internationals "deserve" some points for playing at a higher level; and that something should be done for someone like Devoto with 208 points to be 28th "best" FH AND 935 points to be 22nd "best" centre - for that case; I'd simply add them together, and shove him in his primary position - to make him 17th "best" centre.

I agree that penalising versatile players, injured players and unavailable players is crazy. I'd advocate a minimum number of matches played in a position in order to be included and for ranking to be based on average score, not total. I don't see that much relevance in a fly half's performances at centre to his ability to play fly half and a player's performances in each position he's played for significant time can be assessed on its individual merits.

 

[RedruthRFC]

http://www.rugby.net/s/lmpAyv5MR4Q#Content:contentId=5171798064758784

Thanks for that. It answers most of the questions I had. I like the idea of weighting performances based on the relative strength of teams, this was a possible shortcoming I was going to point out.
@TRF_Peat, it appears that the system is capable of coping with the shortcomings you point out, but the specific graphic linked to isn't doing so.

 

[Peat]

There are a bunch of other graphics available on that site that throw more of a light on things.

 

[j'nuh]

Beautiful presentation of data, but dubious methodology.

 

[dt678]

Cheers all

I'm part of the team that developed this rating system so it's great to see the in depth discussion. Happy to answer any questions and also want to point out the bigger picture European ratings, which include the test rugby of the last year (RWC & 6N):

http://www.rugby.net/g/EU.html

Looking across both test and club rugby evens out the handicap the test players have when we look at just the AP ratings.