Player Ranking System

Dan

Dan

Kim Jong Dan
Staff member
Administrator
Tipping Member
Righto I have been doing some research into player rankings and ideas on how we can rate the players in a way that tries to cut through the bias a little. However I cant find any solid player ranking system that really suits Rugby league and what we do.

As such I want to see if we can come up with a good little system here or formula for ranking players.

If we can I can program a system for us to do our own rankings, league wide if we want. We can rate them on our opinions as part of that, because to be honest that should be a factor, but I think we should be rating them on other things like hitups etc we somehow need to factor in positions as well for those rankings

Dan
 
Dan said:
Righto I have been doing some research into player rankings and ideas on how we can rate the players in a way that tries to cut through the bias a little. However I cant find any solid player ranking system that really suits Rugby league and what we do.

As such I want to see if we can come up with a good little system here or formula for ranking players.

If we can I can program a system for us to do our own rankings, league wide if we want. We can rate them on our opinions as part of that, because to be honest that should be a factor, but I think we should be rating them on other things like hitups etc we somehow need to factor in positions as well for those rankings

Dan
Click to expand...
I've thought about this for a little while cause I'm really into stats and league doesn't really have anything available to fans. Individual teams have their own but not really a unified rating. In order to capture something like that though I believe we'd need to take extensive stats which aren't currently recorded. Things like for defence: initial effective contact and ineffective contact, tackles from marker, missed tackles from marker, tackles resulting in an error, penalties conceded for offside, dominant tackles, penalties conceded for hands in the ruck, linebreaks conceded (could be a split stat), try saving tackle, prob a few others I've forgotten. And most of these rated against min played.

For attack things like: metres gained from dummy half, half breaks, quick play the balls, slow play the balls, amount of times tackled player surrenders, dummy runs, dummy runs assisting in a linebreak, offload to advantage, offload to disadvantage, kicks to open space, kicks that find a man on the full, run metres gained directly after a dominant tackle (the line is more set). And others I've prob forgotten plus all rated against min played.

I've actually thought that if nrl had a bigger audience it could be a business in itself. IMO to rate a player you need to look past just the flashy stats (the fanstasy league numbers) and also look at the total work rate and effort plays that go unnoticed.
 
Lurker has gone where I was thinking, what stats are available?

Once this is known I would suggest the ultimate way is a weighting system, this could even be put to a ST vote in order to get a group average. There is a business approach system based on KT that could work eg (scores out of 10)

Tackles made 1
Tackles missed 5
Line breaks 8
Tries 10
Try assists 8
Line break assists 6
Hit ups 2
runs 1
off loads 3

etc etc

If you allowed ST to vote on each catagory then you will get a group average of what the wider audience considers most important.
 
Didn't Dessie have some stats guru that came up with some you beaut player ranking a few years back, Moneyball style. Remember reading an article about it and it was all secret squirrel stuff.

Anyway what you are trying to do has already been done. It's called Supercoach.
 
I don't have a lot to add except:

1) I can't see how you can fully compare different positions (eg hooker v winger); it may need some kind of ranking by position and maybe after that try to bundle things together.

2) Any system would need to weight who the stat was made against. Perhaps a top-4/bottom-4 skew. Or score better for advantage gained over top-ranked players compared to the rest; and close to zero points when dominating the bottom 10% of NRL players (but a heavy penalty for letting them get the best of you).
 
Chip and Chase said:
http://www.couriermail.com.au/ipad/haslers-stats-are-right-on-the-money/story-fn6ck6f9-1226223360330

http://www.dailytelegraph.com.au/sp...t-won-manly-comp/story-e6frexnr-1226223454383
Click to expand...
yeh C & C different teams have them but they're lock and key type of thing. there is no unified system. Most of the stats I want I would have to manually record lol. In the states data/ stats is so much more transparent and the average punter can get some really advanced stats. So far no one has really tried to money ball the league because it is time consuming and there isn't really enough of a sample size. Because if you took it to another level you could assess how those stats perform against other teams playing styles. What grounds players play better on. How weather or time of day or week affects different players. Consistency of players, do they peak and trough, does a short turnaround affect some players worse than others.

If these stats were actually recorded you could conduct a multivariate analysis to examine team performance. For anyone that liked an intelligent punt it makes it a ****load easier.
 
Bonus points for anyone wearing a maroon and white, but only if they are contracted to Manly for the next season.
 
Time to play the ball would be interesting. Kitey in his last year for us was probably around 5 mins 36 seconds from when he was tackled to upright and rolling the ball under. Used to kill me. Massive moment killer. On the flip side you would want to know something like a tackle for delay (i.e. a tackle that forces a slower play the ball). There would be an absolute ton of things you could dig into but league stats are a joke and a lot of them are misleading.

If you are looking for valuable metrics you'd really have to analyse what @lurker is saying above which would require watching every game and recording everything then determing metres and things gained as a result a lot of different actions. Don't know how you do that.

I have a feeling betting agencies would have teams of analysts doing this. Seems like something that would be have been up Tommy Waterhouses alley.
 
I think it used to be a CVR (don't know what that stood for) system. Something something rating.

It really is a case of stats that we see not showing the full truth. Someone may have 150m with hitups, whilst someone else has 80m. But that person doing the 80m was doing it on our own line when we were getting smashed and he kept putting his hand up to take a hard run and only earning 5 or 6metres. Plus you need to take into account the effect that hit up had, did it get a quick play of the ball? Did it attract 3 or 4 defenders? Did it really take the sting out of the opposition defence?

And then a case in point for me is Jamie Lyon. In his hey day (Say around 2011) I didn't think Jamie had the biggest work rate, but what he did oozed class and won us games. Whether it was a last ditch 1 on 1 tackle (ala against the Broncos in 2011 PF which stopped Matt Gillett) or it was setting up the wolfman for a try his efforts were huge in the context of the win.
 
Killer03 said:
I think it used to be a CVR (don't know what that stood for) system. Something something rating.

It really is a case of stats that we see not showing the full truth. Someone may have 150m with hitups, whilst someone else has 80m. But that person doing the 80m was doing it on our own line when we were getting smashed and he kept putting his hand up to take a hard run and only earning 5 or 6metres. Plus you need to take into account the effect that hit up had, did it get a quick play of the ball? Did it attract 3 or 4 defenders? Did it really take the sting out of the opposition defence?

And then a case in point for me is Jamie Lyon. In his hey day (Say around 2011) I didn't think Jamie had the biggest work rate, but what he did oozed class and won us games. Whether it was a last ditch 1 on 1 tackle (ala against the Broncos in 2011 PF which stopped Matt Gillett) or it was setting up the wolfman for a try his efforts were huge in the context of the win.
Click to expand...
Great example is that Roosters semi where Hiku had 1,000 metres but that's because it was a defensive battle and each team was just banging out 5 tackles/kick so naturally the fullbacks were going to make a heap of yards on kick returns...
 
Ok that's way too deep.

Let's just come up with a simple idea for rating players after a game on their performance
 
Maybe you could contact Spotsdata the company that came up with the CVR (contribution value rating) see if they will throw you a bone? As mentioned above if your looking for a money ball sort of method they filter something like 50 stats into the rating including the 1%'ers that do win games.
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 1 (members: 0, guests: 1)

Members online

Total: 234 (members: 5, guests: 229)

Latest posts

Team P W L PD Pts
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0

Online statistics

Members online
5
Guests online
229
Total visitors
234
Totals may include hidden visitors.

Latest posts

Newest members

Stay Connected

Back
Top Bottom