Post-match: US-Honduras 2/6, general Post-match thread [R]

I guess next you will go to a baseball site and argue that the era stat is useless because it does not record valid runs after an error properly.

This discussion is silly, the industry stats are what they are and all parties have agreed and use them. I have already offered that if you build a better mouse trap I will help you raise some venture and we can have a $300MM per year business. Stop with the sniping, man up and do something about it or STFU!

 

Never yell Kill!! In the presence of a pit bull nor Opta in the presence of Spartacus.

 

Um.....ERA isn't the best metric for evaluating a pitcher. And your appeal to authority bias isn't helpful either (the industry using an imperfect metric doesn't somehow make the metric better, and chances are specific teams may use proprietary formulae and programs and not OPTA).

There's nothing wrong with admitting those parts of OPTA are subjective and imperfect. They're imperfect, incomplete starting points.

 

...

You're the one persistently referring to those silly industry stats.

You stop posting OPTA stats, then I'll STFU about OPTA.
At this moment the "industry standard" is crap.
Open. Your. Mind.

 

Yes it is important, except we don't know anything about the context of the "key pass" stat. So it's meaningless.

And how are you going to take that into account?

And how are you going to take that into account?

If you need a starting point as valuable as the "key pass" stat, then just make something up controversial and start a discussion from there.

That is an interesting practice of course.

 

Zusi's only real impact at the NT level has come in the first half against a bunkering Jamaica, where he then disappeared when Jamaica came out of its shell.

I fail to see how even his workrate can be counted on for a contribution and what exactly does workrate with Zusi even mean if he can't use it to be a good defender and break up attacks and lacks the speed to run into space and threaten and cause the defense to stay honest/account for him?

 

Lets see I can ignore reporting from every media and industry outlet about matches, stop analytical discussions with colleagues about USMNT matches or I can just ignore you. Choices, choices.

 

Or you could address the completely legitimate, nuanced criticism of OPTA as the "end-all-be-all-this-is-perfect-and-i-will-brook-no argument" thing.

 

The USMNT is a team that requires its AMs and wingers to help the defense. It's standard procedure, from the times of Sampson. McBride defended like a lion when it was needed. Same for Eddie Lewis, Cobi Jones, up to the modern times with Dempsey and Donovan covering a lot of the field and helping defend in crucial games.

Kljestan and Zusi didn't do any defending once they came in. And, since they didn't do much attacking, it meant after the second change, Honduras had TWO free men: Maynor Figueroa (who was roaming the whole field from his CB position, while Sacha just watched him) and Mario Martinez (Zusi kept about five steps on the side or behind him, that is, not marking him at all: Edu was split between marking him or Boniek, who was cutting in from their right while Johnson was busy with Peralta).

 

I will address it by saying any moron who made the statement that OPTA is "end-all-be-all-this-is-perfect-and-i-will-brook-no argument" thing should be quickly executed.

Who said such a thing?

 

Well, you're refusing to entertain any debate on it's merit or accuracy, simply by doing the e-equivalent of plugging your ears and yelling "everyone uses it so it must be true la la la".

 

I have no interest in a general debate about it's merit or accuracy any more than I am interested on a general debate on the merits of democracy or hot sex with a couple of Houston Texan cheerleaders. It is what it is.

You and others are welcome to have that debate if you wish. One can denigrate any stat method by focusing on example boundary cases or operator errors. Soccer stats, baseball stats, basketball stats, football stats are all not perfect.

This started with me saying that (per OPTA) LD had produced > 23 chances last summer followed by MB and Torres with 6 and JJ with 3. If someone believes that those are numbers are off then please review those matches and correct them. Until that happens ESPN. MLS, and the USMNT will use those numbers for analysis purposes.

FYI in presentations on OPTA data and vendor/client comparisons OPTA comes out as > 93% accurate which is good enough for most purposes.

 

Statistics, in order to be meaningful, require that:

a) They refer to repeatable events;
b) A significant number of the variables are known.

Soccer is the sport least likely to have useful statistics because:

a) Plays are not set, like in football or baseball --that is, they come from free-standing positions that are, often, never repeated;
b) There are 11 variables per team, instead of only 5, like in basketball, or 6 like in volleyball or ice hockey.

Soccer is like field hockey or softball (cricket also has set starting position per play, so it's like football or baseball).

 

If the statistic and it's methodology are sound, it's pretty hard to denigrate it. That's the problem with this part of the OPTA stuff. It's subjective, it doesn't have a large enough sample size and there's not a set parameter as far as what defines what it's trying to quantify. The combination of flaws make it a less useful tool than it might appear superficially. There are baseball stats that, when taken with other metrics, are extremely accurate predictors of performance, and evaluations of said performances.

Like I said earlier, there are more and more clubs using proprietary metrics and software for this sort of thing. They probably break down the field by zone and look at sample sizes of shots, saves, on frame, goals, etc for thousands of matches to try and provide a decent measure. OPTA is a very good start, but it's an imperfect thing, and just because ESPN uses it and MLS/USSF cite it publicly doesn't mean it's valid, and it definitely doesn't mean it's what MLS clubs/USSF actually use internally.

Source?

 

I agree with this. Soccer will be one of the hardest sports to come up with meaningful metrics for. There are some very intelligent people and groups working on it (I seem to remember voros being involved in some consulting work, I know Billy Beane did some work for either SJ or some EPL club, and I know FSG has some people doing some stuff for LFC).

Soccer stats are in their infancy. I'd say it's similar to the early attempts at using SABR to evaluate defensive performance.

 

That is your opinion. I am very, very familiar with OPTA and can say I have not heard those concerns from vendors, partners, customers, associations, and clubs.

It is safe to say that 90% of entities that do any kind of analysis are using OPTA data, either the base data stream or additional interfaces on top of the base stream.

If you don't believe in the data then I agree don't use, it but when ESPN.com, or MLSSoccer.com, or SKY is reporting chances that dataset is what they are using.

I'll use your data, so how many chances did LD create last summer?

 

Define "chances created".

 

Opportunities knocked up?

 

"Key passes". (i.o.w. passes that lead to a shot)
For example: https://www.bigsoccer.com/community/...-match-thread-r.1982165/page-26#post-27226447

 

SJE / Liverpool / John Henry / FSG / Damien Comolli

 

Any way you want to. In fact if you define a set of objective stats, update them at the end of every match, provide them to any entity that may want to use them (for a fee if you wish of course, after all we are not communists), I promise only to use your stats on BS instead of the terribly flawed OPTA ones.

 

There is some value simply in knowing who makes the last pass before a shot. Over time there will be enough of a sample to show that a player who clearly makes more key passes is providing value. Are each and every one of those key passes equally valuable, obviously no they aren't. Just like every shot taken isn't as equally valuable, or every pass completed. The pass that played in Boniek Garcia was roughly as valuable as his slow rolling cross. I think Beane, in addition to key passes, has talked about crediting passes in a possession sequence that lead to a shot as being a tool he was looking at to derive a player's value. I would expect a player who regularly makes those valuable secondary assist passes to also regularly make assists (GKs and CBs may be positional exceptions).

 

I know you're just being snarky, but I would happily watch games and chart them, break the field down into zones and try to assign value to certain actions, given the time. Perhaps I'll start this endeavor and watch some archived USMNT games so that by the time the Costa Rica I'll have a workable thing going that everyone can tear apart via peer review.

And I think that watching how advanced soccer metrics develop and evolve over time is going to be really fascinating.

 

It may be possible to also add value by comparing the difference in run of play finishing rates on the key passes a player makes to the average finishing rates of the shot takers. The better the chance created the more likely it is to be finished. Then use that as a weight for a player's key pass per 90 rate.