Theo Walcott


#1067

a) I really don’t get why you must make it into A vs. B with stats and watching. They are not incompatible and both are needed in intelligent discussion.

a) refers to this ^…that is, who said anything about ignoring the latter?

b) Sophistry is comparing Wins and Saves to xG, when there is literally zero comparison to be had, and as a4tt suggested, is really very uncommon in someone as intelligent as yourself, I can only imagine that you are having a brain fart or drank too much yesterday like me, because I think you will see how bad that comparison is if you think about it for a second. Like I said, xG would be much closer to something like Runs Added, which is the basis of WAR. Even if it isn’t a perfect analogy to Runs Added, it certainly is nothing like Wins or Saves, which find pretty good corollaries in the suggestions I made just now in the post above.


#1068

No, no, no, no… WAR would not be remotely equivalent to goals… read my last post, it is very clear why they are totally appropriate comparisons (WINS and SAVES)… fact is, state of art in football stats is stone ages.

WRT xG, I suggest you read the entire model carefully… it is rife with problems, irrespective of the fact that it isolates one aspect of contribution to a total performance or even remotely considers the concept of WAR.


#1069

WINS and SAVES are absolutely correlated to quality - just not very well… that was my whole point. Plenty of serious baseball people even argued this not 15-20 years ago. It really isn’t that complicated.


#1070

Your point really doesn’t make sense though. Goals are literally what determine a football match, just like runs determine a baseball match, or points determine a basketball match. Of course, football is by far the lowest-scoring of all of those sports, so goals are naturally a far more valuable commodity. Goals and runs are a literal translation, not an analogy, though to make it even out the context, it would be something like 1 Run= xG x 4 . Not to mention, xG is a model-based statistic, like xFIP for pitchers in ERA, ie, it is eliminating/accounting for confounding variables so to have a more predictive/accurate value of the amount of goals that would result in a normalised statistical environment.

If goals are like runs, and WAR comes from runs added, then WAR for football would be something like a more sophisticated model of xG Added. No, xG is not like WAR, but I didn’t say that either, I was making that link I think to show you just how wrong and bad the comparison of xG (or goals) to Wins and Saves is, but I think you should really be able to see that yourself. Like a4tt said, goals are literally the meat and potatoes of football, they decide it, to repeat myself, as do runs in baseball, with the exception that goals are far more valuable in football, so yeh, if you can’t see how bad that comparison is yourself I don’t know what to say. As you are intelligent you will, though, eventually.


#1071

Of course goals determine a football match, but scoring goals is more complicated than just saying “x scored” and “y assisted” whereas in baseball it IS that simple. Seriously, I can’t believe I am having this conversation with actual football people.

Not going to argue details of xG here now, but if you read it carefully you will see the same issues that are obvious to me, notably position. WAR accounts for all aspects of contribution/detriment, including defense and have been studied in a sport that lends itself PERFECTLY to this sort of modeling with 100+ years of detailed stats.

Ummm… goals are important so case closed? I don’t think you even realize the argument being made then.


#1072

No doubt, baseball lends itself far better to statistical analysis, but that doesn’t mean statistics cannot be used in football, and it doesn’t mean some of the statistical models we have in football are quite as bad as you are making them out. Hence why I said that WAR would be something like a more sophisticated model of xG added…anyways, I was trying to point out the origins of WAR in runs so to point out the basic importance of goals, and from there to show why there is obviously not any analogy whatsoever between xG and Wins and Saves, and that that is sophistry in the absolute extreme, but I probably should’ve just left it with this

Or I probably could’ve just pointed out that Goals are literally Runs (scoring a lot of runs is not a perfect indication of a player being good, but if a player scores a lot of runs he’s usually pretty good), which is a good indication of how comparing someone talking about Goals in football (xG at that), and then you going on to comparing it to something nothing like that (Wins and Saves) is extreme sophistry, when there is an actual direct translation at hand (as he who scored the Run can have a minor or major or even totally incidental–if you get on via an error or a fielder’s choice–part in that run scored, much like in football).

Anyways, to compare it to Wins and Saves is extremely insulting to my intelligence and yours, to continue with it. I think you should see that by now, and I won’t continue to try to make you if you don’t.


#1073

Not really though you can poke holes in those stats as well if you don’t dig deep enough. A pitcher can feast off anemic batting lineups in a weak division where 1/3 of the games in a 162 game season are played. Same goes for a hitter who can feast on the #3, 4, and 5th pitchers in an opposing teams rotation. Then when they get to the post season and face legit aces, their production dries up.


#1074

WAR isn’t the same as xG… I am comparing xG to Wins and Saves because to me they are equally poor at predicting “quality” in their context, despite their staunch defenders. I outlined why they absolutely are comparable… they measure so little of the actual contribution of a single player in their totality.

But irrespective of all that, once you have the data by watching and synthesizing all the various forms of data and reach a conclusion about a player, why go back to the stat-isolated argument when we all agree it the by far the weakest in football?

Runs/RBI are a pretty poor measure as well, but at least better studied and slightly more correlated with WAR (or total quality) than WINS and SAVES, but like I said, plenty of baseball people still feel these are legit, good measures. I don’t think you understand the argument if you don’t see why it is valid comparison or you truly overrate football stats massively.

I will give you some credit, you are at least trying to learn the baseball stats - I have been studying sports stats for 40+ years, so I appreciate that effort.


#1075

Mate, give me a break with your intelligence comment. If you don’t see how I have explained the comparison very clearly, then you are choosing to not want to understand it… you talk about correlation - the only correlation that really ultimately matters is whether or not something is correlated with helping the team win (ie., quality)…

WINS and SAVES are correlated, but not well… they DO represent an amalgamation of personal quality and team quality, much like goals and assists… they also factor in “clutch” elements (BS of course).

RUNS and RBI if you choose as better analogs, fine… FFS stop being a pedantic fck… they suffer from the same fing issue, maybe to a lesser degree.


#1076

Oh it certainly isn’t. Bad arguments based on watching with no objective evidence, and prone to all kinds of psychological/judgment biases, are the worst in football. These, in fact, are what abound on football message boards, and often in this thread. They are what we could call the Lengooner argument. Stats-based arguments are far better in comparison, far better, and again, you are underrating some of the quality of stats out there in football at the moment, without disagreeing that there is a long, long, way to go.

As for the rest, you literally aren’t even reading me, and your arguments continue to absolute miss my points in their totality. Yes, it is insulting to the intelligence.

For example, when I literally just said, in the very post you are responding to:

One would almost think I had said WAR is the same as xG. :roll_eyes:


#1077

Yes, and football suffers from these same holes 100x… sample size (games/season) is much smaller, they only play same opponents twice, under DIFFERING conditions (home/away), etc… like I said, football is abominable at this sort of thing.


#1078

You are comparing them without understanding them or clarifying… they aren’t remotely the same.

Equivalent to the Pires/Theo comparisons. How the f*ck does that advance the convo?


#1079

Theo’s a little bit shit.


#1080

Pot, kettle… I don’t think you are reading me at all either… or at least purposefully trying not to understand.


#1081

But also a little bit decent. On his day.


#1082

If stats adequately measure 10% of a player’s total performance, then maybe not so, eh?

Also ignoring the wisdom of the masses watching, plus experts… of course some fan watching a single game is prone to shockers when judging players.


#1083

Perhaps I haven’t understood WAR correctly, I can accept that, even if I think the Runs= Goals, Runs Added= Goals Added, a formula/metric using Goals Added as the base input before controlling for all other important factors, is a rather logical and step-by-step comparison, certainly more logical than the one you’ve made, even if I’ve fully misunderstood how WAR works, in which case someone engaging in genuine argument would understand so quite easily and still understand the force behind the point the person they were arguing with was making, realising that the other person is coming from a place of relative ignorance about one part of the analogy. :slight_smile:

But I certainly understand Wins and Saves, and I certainly understand that a comparison of goals to Wins and Saves to goals/xG is specious in the extreme, and really one of the worst, most misleading, most nonsensical comparisons I have read on here in a while, which is saying something.

That’s nice that you certainly know more about baseball modeling than me, but it doesn’t change how bad the analogy/comparison was.

…experts…yeah, like the ones that talk for Sky Sports. Or do you mean the ones that write columns for places like statsbomb, who don’t take such a rash and dismissive stance to the idea of, shock and awe!!, applying objective data and logical statistical models to analysis of player value, while always recognising the limits (and trying to make them less) of stats in football? Those ones I don’t ignore completely. :slight_smile:


#1084

We will agree to disagree, but I don’t think you fully understand (or if you do, insanely overrate) xG.

WINS and SAVES are correlated with quality, as I said… they are just problematic for a variety of reasons in terms of judging total quality… just like xG - it really is that simple. If you don’t agree with that incredibly simple statement, then you are ignoring fact in advancement of an agenda.

WRT expert, etc… yeah, take it to the extreme and make a joke of it… that is valid and fair argument… that makes sense… or maybe just the fact that Theo in this case doesn’t start for the 5th place team, like many, many realize and have said for a while.

And yeah, you don’t understand WAR or apparently the history of evolution of sports stats in the most sports stats sport in the history of the world.


#1085

Yeah, like myself. :wink:

Maybe you should read the argument above before you came in, I feel like you have a totally mis-informed idea of my stance.


#1086

You placed yourself in the same sentence as A4T at least more than once - maybe you can read his stance on Theo and rethink?