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

Does it?
Impermanence I tell you.

 

@ Honduars, Minute 37:09 1T.
That's a "key pass" according to OPTA.
That's that TheHoustonHoyaFan thus considers to be a "chance" ("it is considered a chance (key pass").

 

Do you have a clip of it on youtube or something? I don't have my DVR'd version at my disposal... I am here at work...working...

Can you paint the picture of what happened?

 

Your standards are just too high for reality.

 

Just so everyone is clear what you are railing against.

You are against OPTA which is the #1 soccer stat service used by every major soccer entity (associations, clubs, media) on the planet. ESPN and MLS in the US for example uses OPTA and their definitions.

We need to be clear that every major soccer entity considers a "key pass" a chance not just me.

 

Yeah, it's pretty sad really.

 

And here a no "key pass" for Jozy. This is OPTA in a nutshell for you: pure crap.

 

Good luck convincing the world of football/soccer/futbol that those OPTA stats are crap.

 

You gotta start somewhere...

 

Soccer statistics like this are in their infancy; I'm sure they'll improve over time, but tracking things like "chances created" right now is really subjective. I'm not even sure where you would begin narrowing it down...you'd probably have to break the field down into different grids or areas, go back through thousands of game logs to determine the number of shots from each area, the amount on frame, the amount of saves, goals scored, and then somehow go from there.

 

Now this is more like it. Good post.

 

I don't really have a problem with this, because you could make this argument against any stat, including goals or assists.

If a ball deflects off an attacker and goes in, he'll be credited with a goal even if he had no awareness of it. (Ching got a goal in qualifying this way in the last cycle).

If a player delivers a 5 yard backpass 30 yards away from goal and the recipient shoots and scores from 30 yards away, the player will be credited with a key pass and an assist.

In individual instances, the stats may not tell you much, but with a large enough sample size, the noise begins to vanish. If one forward has twice as many goals as another, or one midfielder has three times as many assists or key passes as another, that's telling you something meaningful. That can't just be coincidence. (Although it could be correlated with other meaningful stats, like total passes, so you'd want to run those numbers against some other variables.)

Now if the stats aren't actually being properly compiled by the standards set by their own metrics, well that's a completely different problem.

 

Wow! a key pass? this one is very subjective to me. So any pass that leads to a shot no matter how terrible the chance is, is consider a "key pass"

 

So you are saying the passback that Tim Howard got and cleared but end up being a goal 65 yards away is consider a "key pass".

 

Correct. Any deliberate pass that leads to a shot is a key pass. That is essentially what all the parties agreed to in order to take the subjective element out of that stat.

There is another OPTA stat called Big Chance or Clear Cut Chance which has the subjective element of the operator determining if the pass created a "likely" chance to score.

 

Yes. The pass would also be an assist, and the clearance would also go down as a goal.

But over the course of let's say 10 games, you wouldn't expect Howard to be one of Everton's leading goalscorers as a result of the way the stat defines goals -- and if he is that tells you something meaningful about the lack of finishing elsewhere in the team. Similarly, you wouldn't expect the person who made the pass in that individual instance to be leading the team in assists or key passes, unless he actually was consistently putting his teammates in dangerous attacking positions across the field (or unless everyone else on the field was failing to do that as well).

 

It is considered an assist and goes into the official EPL records as an assist.

 

The wind that day was credited with an unofficial assist.

 

Yes, but that doesn't make this so called "key pass" stat any less crappier. And as the Jozy non-key-pass example shows, it's also highly subjective as OPTA simply decided when they think something is worth to be called a "pass" or a "key pass" and when not (ie. Jozy's pass to Dempsey).

I'm sorry but this is incorrect.

Well nothing in soccer is coincidental. I mean, it's not a coincidence that someone playing with Dempsey on his side will record more key passes, than with Zusi on his side.

Does it tell us anything about the quality of those key passes? Nope. Does this mean there's noise reduction? We don't know. Does this mean the "key pass" stat (which also includes set plays(!)) is meaningful? No it's not.

 

I asked OPTA if they could start tracking key posts, but they never wrote back.

 

Yes. Except when OPTA decides it's not. We don't know based on what grounds though. Blind, lazy, dumb, prejudiced, weird, etc.? For one thing they make a lot of mistakes and are highly inconsistent.

 

Adding "Deliberate" took out the "subjective element"? Hilarious.

 

Great post! Very good way of putting it. It's getting around the 'noise' and making sense of the numbers.

If it happened over and over and over again.... yea! That attacker decided it was better to drop it back to Killer Timmy (instead of pushing forward) to boot it in the net!

'Too much ********ing noise' they said!

 

It absolutely is meaningful; it just doesn't tell the entire story. Which again, is the same with every stat. Context is always important.

Just because "assist" figures include set plays as well, or goals include penalty kicks, that doesn't make them meaningless. You just have to take that into account when you're trying to draw conclusions. (Other examples: if an average attacking midfielder is playing with an excellent striker, who is consistently able to get a shot off in difficult circumstances or scores with ease, then you would want to take that into account in evaluating that midfielder's assist or key pass figures; or if a fullback is combining with a winger on his side of the field who likes to take a lot of speculative shots, than maybe you can't draw much from his "key pass" numbers.)

Even the most useful stats are just a starting point for discussion. From there you have to figure out what else is going on and what conclusions you're justified in drawing. But the way to go about that is to find systematic variables that could account for the differences over time between one player's stats and another's, not single instances where the stat failed to tell you anything of consequence.