Lineups, Formations, Tactics, Stats, etc.

That's pretty stunning - 20 goals against vs. ~10 xGA?? Unbelievable. Looks like probably a mixture of bad luck (opponent overperform) and bad goalkeeping. But regardless I'd much rather have GA >> xGA then the other way around because if there is a reversion to the mean, things get better.

 

I think a lot of the goals we’ve given up are on the defenders more than the ‘keeper. Or, is this supposed to factor out the defense?

Go Quakesfans!!

 

It kinda factors out the defense. xGA is about a player getting into position to take a shot. That shot’s xG depends on how make-able that shot is. Take the 1st CO goal. Paul Marie makes a horrible back pass and Yarbrough fouls in the box. Two big defensive screwups, one by a defender, one by GK, and it results in a PK - a super high XG shot.

But here’s the thing. Our collective xG is mid-pack, which means we are allowing shot types at a rate no worse than the average team. Yet our GA is 2X our xGA! So either our opponents have been nails with shots against us or our GKing has been poor of some combination.

 

Is there a precise statistical evaluation of “how make-able the shot is” to derive xG for a given shot?

 

No. They measure everything by assumed average, not factual numbers involving xG and xGA. Ive learned theough schooling recently no one in the profession takes Xg and XGA serious as a metric.

Will Kuntz said at a web class,he doesnt recruit off xG or xGA, those arent factual metrics, that it was created so soccer commentators had more to talk about. Its not a factual metric used by anyone in MLS/Soccer currently.

Theres no base to it.

 

I'm shocked!

 

Thanks!

By the eye test, which is probably biased and inaccurate, it looks to me like our defenders are crap and give up a ton of shots with next to no pressure. I blame most of the goals we’ve allowed not on ‘Brough, but on the defenders in front of him.

That’s my observation, which might not be right. What do the rest of you think? Is our GK terrible? Or is it that our defenders are crap?

Go Quakesfans!!

 

I think Yarbrough has been pretty solid, Daniel was struggling early, possibly trying to play through his injury, our defenders have made a hash of things...we've given up a lot of softies. Not that I value that stat, but it just screams poor defending to me. Luchi can't even get the team to play Dom jr ball now, which is the only way he has survived in coaching, without a somewhat sound D, he doesn't have a job...because he isn't doing squat on the other side of the ball. I'm really surprised he hasn't gone more conservative ( double pivot 4-2-3-1 or empty bucket 4-4-2) to stop the bleeding in the back..

 

Keeper isnt the problem. Our problem is Defense. Rodrigues isnt all that good. And when paired with 3 different style defenders were going to be scattered

 

Yes.

“Expected goals (or xG) measures the quality of a chance by calculating the likelihood that it will be scored by using information on similar shots in the past. We use nearly one million shots from Opta’s historical database to measure xG on a scale between zero and one, where zero represents a chance that is impossible to score, and one represents a chance that a player would be expected to score every single time.”

https://theanalyst.com/na/2023/08/what-is-expected-goals-xg/

 

“Similar shots” seems highly scientific. Not.

 

Interesting. Who makes the determination of these metrics? I don't see Yarbrough as the problem. I see a back four (especially centerbacks) who are have been terrible.

 

The xG value is based on 20 different characteristics. Some of the most important ones:

• Distance to the goal.
• Angle to the goal.
• Goalkeeper position, giving us information on the likelihood that they’re able to make a save.
• The clarity the shooter has of the goal mouth, based on the positions of other players.
• The amount of pressure they are under from the opposition defenders.
• Shot type, such as which foot the shooter used or whether it was a volley/header/one-on-one.
• Pattern of play (e.g., open play, fast break, direct free-kick, corner kick, throw-in etc.).
• Information on the previous action, such as the type of assist (e.g., through ball, cross etc.).

https://theanalyst.com/na/2023/08/what-is-expected-goals-xg/

 

Wondo vs Tommy T taking the shot?

Also, each of these bullet-point "characteristics" has a level of subjectivity built in.

 

Many of the characterists are not subjective: distance from goal, angle of goal, GK position, etc. And the shots are evaluated by machine model, based on training from a million or so shots. They train the model to accurately predict the likelihood a shot will go in based on what actually happend in real life (the million or so training shots).

 

And no actual people in any relevant credible soccer role use xG & xGA. I'll stick with the employed vs the manufactured stat made up.

xG & xGA don't tell the whole story, a player could take 20 shots and 19 on goal and never go in his xG its going to give miscued results because they could be contributing nothing else to the field.

Thats like saying how many headers John Doyle had a game made him a great defender.

 

How do you know those things were accurately measured (within what margin of error?) and not merely guesstimated?

And what weight does each of the characteristics receive? For example, is the distance more important than angle? And if so, how is it weighted?

I take it the proficiency of the shooter has no bearing, since you ignored that aspect of my prior post. Yet, isn't that the most important characteristic? Messi or Ramiro? Who ya got? xG all the same if the distance and angle and GK position are identical, amirite?

 

It's just a tool. No one is claiming it's a be-all / end-all. People can use it or not. For as long as there's been soccer, seems like people have talked about "fair result" or not. This is a way to try to be a little more objective about that. We can talk about shots, but then what about the quality of the shots, etc. Doesn't hurt to try to be a little more objective about it. But again, if you don't like it, or you don't believe in it, that's fine.

 

It's not just me and @Kaitlyn.. According to her, none of the MLS data crunchers rely on it, either. Don't know if that's true or not, but if it is, does that cause you to reevaluate?

 

I don't have a strong opinion about it either way. However, I find it somewhat complex. In my view, the metrics commonly used in baseball don't translate directly to soccer, no matter how much effort is put into it. For instance, in soccer, a shot on target can be considered a quality shot, whereas a shot that sails into orbit like a Ramiro Corrales or Simon Elliott special would not be considered a good shot.

 

The characteristics are measured by the "machine". There's no human intervention. I imagine that there are video analysis tools that can determine player positions on the field fairly accurately.

As for what weight do the characteristics receive, that's why you train the model - to figure that out. You design the model to match the empirical data as best you can. The "machine" does that for it actually - it trains itself.

One of the uses of xG is evalaute the proficiency of the shooter. A player with a significantly higher G than xG is a good finisher. A player with a significantly lower G than XG is a poor finisher. Cade had one of the worst G-xG in MLS last year I think. Messi probably has a positive G-xG.

When you see a huge discrepancy like our xGA vs. GA, you could say that we've just played against a bunch of great finishers or that they were average finishers who just happened to make great finishes against us, or that our GK'ing wasn't very good.

In his prime, Messi usually out-paced xG pretty significantly, as you'd expect.

https://www.messivsronaldo.app/detailed-stats/xg-expected-goals/

 

LAFC don't use them which Kuntz confirmed. Waibel has talked about in his pressors how they evaluate talent Seattle, is based on a player by player basis to fill needs.

Waibel in Seattle has talked about using some form of Analytics but more of a Creative Metrics approach as to PPDA (Passes per defensive actions), Progressive Distance covered, Team Sequence, pass per vs SCA (Shot created actions - Run On, Run oFf, positioning, and tactical awareness.

Wilfred Nance also uses a different metric to evaluate his talent, which has made him successful in using average to slightly above average without any real game changers. He talked about finding and tinkering with the right combinations of players based off chemistry and metrics to team sequence and build up. He wants talented players but smart athletes vs Talent alone.

Luchi and Leitch unfortunately are Muppets playing with action figures.

 

How did the computer models do during Covid?

https://www.ncbi.nlm.nih.gov/pmc/ar...shows the percentage,figure was less than 20%.

 

Bolded - Well thats true, there is tools and assist tools that measure metrics like that distance covered, run of play and positioning and touches per run of offensive action.

There is tools (and tools and software used by the professionals I own), that I can load the match up into the program and i can track individual players. I can tell you Rodriguez is the most out of position player on our team, right behind him is Ebobisse. Where Espinoza, Yueill and Marie are 3 of the better players by positional metric. Tommy Thompson metrically is a better player on numbers than Chris Wondolowski ever was, but that is debunked almost immediately by on field performance.

Some CB's have better G-Xg than strikers.