The EPS squad stat thread - Rd. 6

[Tony Manx]

You can't win a penalty.

I don't disagree with your fundamental point that it would be nice to know how many penalties a player forces through jackaling, but the fact is that you don't win penalties.

I thought that was what most coaches believe is the purpose of scrummaging....to win penalties!!

 

[ratsapprentice]

It's an erroneous phrase.

It's not being pedantic, it's relevant to the question of why a "penalties won" stat isn't given - because such a data point doesn't exist.

You don't gain the penalty, you gain possession of the ball as a result of the penalty imposed on the opposition.

 

[Ragey Erasmus]

It's an erroneous phrase.

It's not being pedantic, it's relevant to the question of why a "penalties won" stat isn't given - because such a data point doesn't exist.

You don't gain the penalty, you gain possession of the ball as a result of the penalty imposed on the opposition.

This makes sense when you think about it. Ok in a 1 on 1 scenario you can "win" a penalty but what about a scenario where say 2 players compete for the ball and the guy holds on? Statistically that would show as winning 2 penalties. Likewise with scrums, you can't realistically give just 1 individual or all 8 the penalty win but you can say who the penalty is against.

 

[j'nuh]

From World Rugby: Penalty kicks and free kicks are awarded to the non-offending team for infringements by their opponents.

Penalties can be acquired as a result of a contest (i.e. won) by causing an infringement by the opposition. The infringement is caused by jackaling.

I think you're mistaking the infringement (the sanction given to the offending team) and the penalty (a kick awarded to the non-offending team). You can't take ownership of an infringement, but you can of a penalty.

And even if not, I questioned why a penalty won via a jackal isn't classed as a turnover. Whether you call it a penalty won, a baboozalie or otherwise, forcing a penalty via a jackal is winning possession, i.e. a turnover.

 

[TRF_Olyy]

I can see both sides of the argument, but I'd like to see "penalties won" as a seperate stat.

There's always the argument of "What if there's more than one player jackling?", but tackle stats are still awarded and there's often more than one player involved in a tackle.

 

[ratsapprentice]

WR being lazy doesn't make me wrong.

Incidents that take place at the ruck aren't the exception to the prevailing use of the term penalty in all other facets of play.

 

[j'nuh]

That definition was taken directly from the laws. If WR are being sloppy with the laws, then we have a much bigger concern!

 

[TRF_Olyy]

To be fair, it is a tricky situation. There's penalties for holding on, but what about penalties where you've forced a player to dive off his feet to secure it? Or come in from the side because you're there first and he doesn't have time to get through the gate?

 

[ratsapprentice]

That definition was taken directly from the laws.

Yes - and I'm saying it's wrong.

You don't award or grant a murder victim the death penalty - the penalty is "owned" by the party which receives punishment.

 

[j'nuh]

Most people use penalty as shorthand to refer to the penalty kick, rather than an abstract form of punishment. "He scored a penalty", "he gave away a penalty", "a penalty is worth 3 points" etc. You are awarded, not punished, with a penalty. It's the term used by WR and the accepted parlance. That's good enough for me.

 

[ratsapprentice]

And that nonsensical use of the term is why you are confused as to why "penalties won" don't count as turnovers.

 

[j'nuh]

Penalties won do count as turnovers though (even if you don't refer to it as "penalties won"). You didn't have possession. You caused the opponent to infringe. Now you have possession. That is a turnover of possession. IMO, it's simply an oversight by the various places which construct stats not to include winning penalties as turnovers. They probably don't put as much thought into it as we do. (Especially given the spurious accuracy, maintenance and presentation of the stats.)

I do wish we had access to the stats put together by the clubs.

 

[ratsapprentice]

I'm sure Opta, a company whose only service is providing sports stats, don't put as much thought into sports stats as we do. :lol:

Penalties "won" do not count as turnovers, clearly.

 

[Vieux Talonneur]

All this proves is that any coach worth his salt should trust the judgement of their own eyes above anything else. Stats can occasionally be useful background validation but little more and quite often downright misleading.

As soon as coaches start talking stats, metres run, heart rates and PowerPoint I despair.

 

[j'nuh]

I'm sure Opta, a company whose only service is providing sports stats, don't put as much thought into sports stats as we do. :lol:

Penalties "won" do not count as turnovers, clearly.

Okay, maybe I was being overly harsh. Opta appear to be fairly comprehensive, but they are completely stakeholder driven. If their customers don't want the data, they won't produce it.

ESPN attempt to produce layperson-friendly statistics, and they do it terribly e.g.:
They do not even present the number of minutes a player is on the pitch in a game and you have to work it out from their timeline, something you would suspect is one of the most basic pieces of data needed for any kind of accurate analysis.
They presented data in a column called "Turnovers" (not in the current design, but the previous one), and expected users to realise that this actually refers to the number of turnovers given away rather than won.
They failed to capture data on something as basic as number of rucks hit, a vital piece of information to determine the work rate of a player.
The defensive data provided is tackles and missed tackles.

I guess we will have to agree to disagree on the issue of whether penalties won count as turnovers.

 

[ratsapprentice]

All this proves is that any coach worth his salt should trust the judgement of their own eyes above anything else. Stats can occasionally be useful background validation but little more and quite often downright misleading.

I disagree - you could actually probably pick certain positions of a world XV almost entirely using the stats I posted - back three positions specifically, and the players you pick based on that would also be the players you would pick having only watched the players.

The idea that statistics are useless, is complete ******** as is the idea that you can base all of your ideas on stats.
A holistic, objective understanding of a player's capabilities is greatly (not slightly) enhanced by the use of statistics.
They allow you to quantify much of what a player actually does.

 

[ncurd]

Raw stats are generally useless being able to collate them and interpret what they mean takes real skill.

I say that as someone who loves stats and reading into them but probably isn't that good at it.

The phrase lies damn lies and statistics always annoys me, stats don't lie the person interpreting them wrongly is.

 

[ratsapprentice]

Raw stats are generally useless being able to collate them and interpret what they mean takes real skill.

Certain stats do - others don't.

 

[j'nuh]

Raw stats are generally useless being able to collate them and interpret what they mean takes real skill.

I say that as someone who loves stats and reading into them but probably isn't that good at it.

The phrase lies damn lies and statistics always annoys me, stats don't lie the person interpreting them wrongly is.

Statistics can certainly lie if the methodology behind the collection of the data is flawed. And there are inherent flaws in collecting rugby data. With the small sample sizes (players only playing around 30 games a season) and an absolute myriad of lurking variables, serious statistical analysis isn't just difficult, it's impossible. Comparisons are mostly useless. Moreover, most of the actually useful data, such as heart rates over 80 minutes, and impact in the tackle, is only collected by clubs and isn't readily available.

Off the top of my head, I would say that there are two main uses to the data produced by e.g. ESPN:
1. Some players do post obviously impressive data. Beating an average of 6 defenders a game over 20 games is clearly impressive in spite of the data limitations. Topping your team's tackle counts most games is also impressive. Basically, if your results are fairly superlative, then there's no arguments that it isn't impressive.
2. Supporting your own eyes. This is the important one IMO. e.g. you notice that a player is constantly tackling players behind the gain line but fail to get an impression of how many tackles the player makes, the stats tell you they made 15... complementing each other, they are greater than the sum of their parts.

 

[Tony Manx]

Statistics, damn statistics...