and for context the 2009 Berro v 2008 Berro superleague.co.uk/statistics_pla ... eason=2010

and for context the 2009 Berro v 2008 Berro superleague.co.uk/statistics_pla ... eason=2010

Yes. Again, I'll refer back to my specific to the game example of Fitzgibbon/Tickle/tackling. Because we don't know exactly what the statistics are measuring and don't have a full data set, we have no idea what they are showing.

We'll agree to disagree then. Due to variables with most data sets in my line of work i'd have no metrics to work with at all if i started at the point of discounting data on that basis. As long as you know when looking at the metrics that certain variables have not been accounted for you can interpret the data accordingly.

For me i look at the data of player x making 500 tackles vs player y making 300. In my mind i look for the context. What kind of game time has each player had? What position in the defensive line does each player adopt? What are their styles? Whats the value add for play y's 300 tackles vs player x's 500. I look for all of the additionals to interpret the data. It doesnt mean i discount the figure, i simply look to rationalise it. I guess that comes from working where i do with the ambiguous data that i do.

Can you stop with this now? You say that the links you're posting are interesting, but why? You're just posting a list of numbers and pointing at them as though they have some sort of value. What is it, specifically, in game-related terms that you think they are showing?

When this was brought up earlier ([url=http://viewtopic.php?p=15115330#p15115330here[/url being the last example before you diverted the discussion back to your semantic driven comfort zone), you avoided any discussion and debate of what they actually meant and how they related to comparison between Berrigan and Houghton. Now you've brought another couple of players into the mix and still done nothing but point at the list of numbers as though that's meant to mean something. Well what does it mean?



Slightly better, but you still haven't actually gone into any deeper than asking a couple of vague questions without attempting to answer them. What [idoes[/i it mean that player x has made 200 more tackles than player y? Can we start from there? Let's give player x the name 'Houghton' and player y the name 'Berrigan' to make it slightly more specific to the discussion at hand.

The point i was making was that i for one have Swain on a pedestal and Berro as average. Looking purely at the statistics of the 2 players with no context they aren't as dis-similar as some may have suspected. (me included) Thats why i found them interesting.



I asked a couple of fairly vague questions to start a debate about what it means. You tell me what you think it means? For me i think their defensive workrate is fairly similar but i do agree that Houghton rarely wins the collision and Berro seems to hit that bit harder. Houghton clearly makes more yards per carry but then that would fit with him coming on late in the halves and running at tired defences.

The telling stat for me is looking at Berro 2008 and 2009 vs 2010. Berro surprised me with the number of assists he had for those earlier years. Is it simply that he played 7 more or is it that we played more off the 9 in those years that we do this year? Assists as i understand it are captured based on giving the final pass or final chip through. Has Berro simply been unlucky that he has created the space for other players to act as providers?

There's probably no point continuing this, but without context they mean nothing.



Right, it's starting to get a little deeper now. I've already explained the major reason why Houghton (and Swain, for that matter) makes more tackles so I'll leave that for now. Berrigan isn't targeted in anything like the same way because he regularly completely dominates the carrier and works them onto their back.

Yards per carry - what's a carry? Is Berrigan getting a couple of yards out of dummy half before giving the pass a carry? Or is it just a carry when the player is tackle whilst in possession? Without knowing, it's impossible to extrapolate on this.

Assists, again, it depends entirely on what they are. Assuming that it is the final pass/kick, you wouldn't expect that many from a hooker anyway, especially with a fairly weak pack like we had last year. He played much more at half in 2008, so that will also be a reason for him having more then. And I expect the main part of it will be luck. I certainly wouldn't be making a judgement on a hooker's performance based on their number of try-passes.

Youre right there's probably little point carrying on the conversation. If i was to give one piece of advice (and please dont think this is condasending as it's not intended that way) i think you need to put the methodology to one side and questioning the validity of the data and learn to trust the source in order to open you mind and discuss the actual players and their attributes and have a debate. The stats provider clearly do this a lot and the coaches of SL clubs use the same source i believe. Going back 10 pages to my point to Kosh, just because you dont know the methodology of the data doesnt automatically mean it's wrong. As long as there is consistency between how a tackle or an assist is captured for each and every player the data is usable imo.

So, going back to what I said about metres per carry, how can this be debated?



Only if everything each player did was in the same proportion.

Quote ="Big Dave T"On that note i've never compared Berro and Swain statistically as they are different players but thought this was interesting too: 2004 Swain v 2009 Berro superleague.co.uk/statistics_pla ... eason=2010 2005 Swain v 2009 Berro superleague.co.uk/statistics_pla ... eason=2010'" Very interesting stats...I think the reason most hold Swain in such high esteem and what sets him apart from Berro is Swainey played with passion every time he stepped onto a pitch.

Quote ="Big Dave T"On that note i've never compared Berro and Swain statistically as they are different players but thought this was interesting too: 2004 Swain v 2009 Berro superleague.co.uk/statistics_pla ... eason=2010 2005 Swain v 2009 Berro superleague.co.uk/statistics_pla ... eason=2010'" Very interesting stats...I think the reason most hold Swain in such high esteem and what sets him apart from Berro is Swainey played with passion every time he stepped onto a pitch.

Metres per carry in simplistic terms would be the metres divided by carries. Metres would be measured by seeing where the ptb ball is and how many metres are made over the advantage like before been tackles. A carry would be each time a player takes it in and is tackles. I'd hope that a professional statistical research company capturing this data would have an agreed approach set down with guidelines as to what constitutes each metric. I'd also hope this has been derived in conjuction with the clubs to ensure it measures what they want measuring in the way that they want it measuring.

Sometimes simplicity is the best approach and ignoring variables and classing them as a working assumption that they are a constant is the best way of approaching these complex situations.

So do you think one of the issues may be that Berro is too naturally gifted and looks like he's taking it easy when maybe he isnt?

No, I just don't think he puts as much effort in any-more tbh.

That's my view too but the more i debate it the more i'm open to been convinced otherwise.

reading your debate has given me a nose bleed!

and ive got a headache. Anyway, it's nearly pub o'clock.

Right, let's assume that a carry is as you've said then (a terrible definition as far as I'm concerned, not through your fault, but through the fault of whoever it is that's decided to try and measure something as variable as what a carry is). The terrible, average, coasting Berrigan, this season, has an average of just over 6m per carry. The NRL-bound Houghton has an average of just under 7.5m per carry. What does that tell us? 1.5m is the difference between success and failure? Or Berrigan makes so many more carries that he doesn't have the energy to make the extra 1.5m that would elevate him to greatness? Or something else?



Sometimes, but not in this case. Again, Fitzgibbon/Tickle/tackling. That simple hypothetical remains unchallenged yet is still the massive stumbling block that these statistics haven't even got close to overcoming.

See for me the metres per carry is an irrelevant stat for a hooker. The better use of metres per carry is in the forward armwrestle where our great game has changed.

Moa making 8 metres per carry compared to Cusack making 4 metres per carry theoretically tells you that Moa makes twice the yards Cusack does therefore giving a better platform. Obviously before you say it the stat doesnt say how many defenders Cusack pulls in or how quick the ptb is he gets.

The telling stat for me for hookers would be around pass accuracy and speed.

Sometimes overly complication statistics and needing to consider all variables is often the blockage to moving things forward imo. (as evidenced by the debate not really moving on for 10 pages)

Stats are for s

Succinct! Probably about right too. On the field you need a team that works for each other and you need players that care as well as have skill etc.

fair point

That, and the chargedown at Cardiff.

Back to berrigan.

Were not missing him are we?

Houghton is every bit as good you know.

only just!

I thought Berro was pretty annonymous yesterday!