MLS - May 18 - Red Bulls (Home - CF)

in addition to what happened during game play, in the last week we saw:

-a Toronto FC player get involved in a brawl in a game he didn’t even suit up for
-multiple headbutts and kicks to the groin between opposing players in said brawl
-a coach accusing the NYCFC coach of punching a teenager two months ago
-a raccoon having one of the best animal on the field runs in MLS history
-NYCFC subsequently adopting the raccoon as its new mascot
-a major crowd control miss enabling tons of fans to be late to the game
-the home tailgate being moved because of a dead body

Yeah. We are nothing if not interesting. What a week.

:oops: you cant just add that and not give details
 
  • Like
Reactions: FootyLovin
I’m one of those terrible people who can’t help but check out forums for teams I dislike when they have a bad result . Over on MF they seem to be doing two things 1- whining about the ref 2- foaming at the mouth with hatred because we’re all a bunch of euro posers who didn’t support mls from day 1.
They can “speak for” themselves. I was a charter season ticket holder in 96. I doubt most of them were.
 
  • Like
Reactions: CCMore and gbservis
"xG says [___] but I know what I saw"
Totally agree with your takedown of people overestimating their own eye-test skills vs what science and statistics tell us is absolutely better judged in other ways. My only caveat is that I can recall at least one play off the top of my mind (and I think there were other similar ones) where NJ had a 2v2; both CBs covered the ball and were very lucky to catch the cross which otherwise would have given a tap in.

I think those plays, which don't show up in xG, can loom large in the mind.

Of course, I can also remember a play where we had a 3v1 I think and Santi scuffed the cross allowing the 1 NJ player to clear.
 
I didn't hear Twellman's take - I will go back and listen if I have time.

Xg often gets used to evaluate things it's not good at figuring out. The relative performance of two teams in a single game (or single half) is one of them. Xg needs a larger sample size to be really meaningful.

It's also a fairly narrow measure; it's focused on shot quality (and to some extent, quantity). It doesn't measure the other things that go into that.

We probably had a high xG because our two goals were close to tap ins. That makes those plays seem a lot easier than they were as each required a high quality assist and some good progressive passing before that. If either cross had failed to connect by a few inches, then the xG is zero for that play.
 
Totally agree with your takedown of people overestimating their own eye-test skills vs what science and statistics tell us is absolutely better judged in other ways. My only caveat is that I can recall at least one play off the top of my mind (and I think there were other similar ones) where NJ had a 2v2; both CBs covered the ball and were very lucky to catch the cross which otherwise would have given a tap in.

I think those plays, which don't show up in xG, can loom large in the mind.

Of course, I can also remember a play where we had a 3v1 I think and Santi scuffed the cross allowing the 1 NJ player to clear.
I didn't hear Twellman's take - I will go back and listen if I have time.

Xg often gets used to evaluate things it's not good at figuring out. The relative performance of two teams in a single game (or single half) is one of them. Xg needs a larger sample size to be really meaningful.

It's also a fairly narrow measure; it's focused on shot quality (and to some extent, quantity). It doesn't measure the other things that go into that.

We probably had a high xG because our two goals were close to tap ins. That makes those plays seem a lot easier than they were as each required a high quality assist and some good progressive passing before that. If either cross had failed to connect by a few inches, then the xG is zero for that play.
ASA has a Goals+ metric that attempts to measure various actions that make goals more or less likely, including passing, dribbling, interrupting, etc. John Muller has a recent article suggesting that the concept of "progressive actions," currently subject to varying definitions but generally meaning things that move the ball substantially closer to the opposing goal, is the best all-round metric for soccer.

I think they complement rather than supplant xG, which I agree is limited but IMO gets to the heart of things. I take the point you both make about dangerous plays that yield 0 xG. A team can create a real chance and then fail just short of a shot which means zero xG. I think it's good that we have metrics that account for all the pre-shot stuff and another metric that only measures when you do in fact generate a shot.

I think of it like this: the thing you want is to score goals. Excluding own goals, to score you have to shoot, and to shoot you need to move the ball forward, and before you can move the ball forward you need possession, and to gain possession you have to play defense. That's a bit simplified but you get the point. Goals+ and progressive actions measure everything that gets you to the shot, which is the next-to-last thing. xG measures the final thing that gets you to the thing. And if you don't take a shot, well that should be worth 0 on that measurement. For the Wolf goal, there was extended play from the back, leading to Santi's linebreaker, the Jones run, his pass to Wolf, Wolf's little dance and finally the shot. On one hand every one of those steps has value even if the next step fails because at a minimum maintaining possession and moving the ball away from your goal lessens the danger of conceding. But on the other hand, it doesn't matter if the play breaks down because (1) Santi's pass gets intercepted or (2) Malachi makes the wrong pass to Wolf or (3) Wolf is dispossessed before he can shoot. Because every one of those means the same thing: you didn't take a shot and if you don't shoot you don't score goals.

Finally, I will note it is quite rare for many xG haters to actually promote Goals+ or progressive actions or anything else as a meaningful concept. They would mostly rather complain that their team should have scored 4 times by the half on 7 shots "if they were just more clinical" and xG be damned, or create a narrative in their head like "Team X was actually more dangerous" regardless of what any measurable activity reveals. That's not you two, but a lot of xG criticism is just people who don't want their soccer opinions held up to any objective measure.
 
Last edited:
ASA has a Goals+ metric that attempts to measure various actions that make goals more or less likely, including passing, dribbling, interrupting, etc. John Muller has a recent article suggesting that the concept of "progressive actions," currently subject to varying definitions but generally meaning things that move the ball substantially closer to the opposing goal, is the best all-round metric for soccer.

I think they complement rather than supplant xG, which I agree is limited but IMO gets to the heart of things. I take the point you both make about dangerous plays that yield 0 xG. A team can create a real chance and then fail just short of a shot which means zero xG. I think it's good that we have metrics that account for all the pre-shot stuff and another metric that only measures when you do in fact generate a shot.

I think of it like this: the thing you want is to score goals. Excluding own goals, to score you have to shoot, and to shoot you need to move the ball forward, and before you can move the ball forward you need possession, and to gain possession you have to play defense. That's a bit simplified but you get the point. Goals+ and progressive actions measure everything that gets you to the shot, which is the next-to-last thing. xG measures the final thing that gets you to the thing. And if you don't take a shot, well that should be worth 0 on that measurement. For the Wolf goal, there was extended play from the back, leading to Santi's linebreaker, the Jones run, his pass to Wolf, Wolf's little dance and finally the shot. On one hand every one of those steps has value even if the next step fails because at a minimum maintaining possession and moving the ball away from your goal lessens the danger of conceding. But on the other hand, it doesn't matter if the play breaks down because (1) Santi's pass gets intercepted or (2) Malachi makes the wrong pass to Wolf or (3) Wolf is dispossessed before he can shoot. Because every one of those means the same thing: you didn't take a shot and if you don't shoot you don't score goals.

Finally, I will note it is quite rare for many xG haters to actually promote Goals+ or progressive actions or anything else as a meaningful concept. They would mostly rather complain that their team should have scored 4 times by the half on 7 shots "if they were just more clinical" and xG be damned, or create a narrative in their head like "Team X was actually more dangerous" regardless of what any measurable activity reveals. That's not you two, but a lot of xG criticism is just people who don't want their soccer opinions held up to any objective measure.
I wonder how teams break down by average xG and how this relates to total xG. I imagine some teams are much quicker to take the first decent look and count on the rule of large numbers working out for them. Other teams will be more patient, foregoing early shots in hopes of manufacturing that higher quality opportunity. As described before, all of those attempts at higher quality that don’t end in shots reduce the total xG. But presumably they would increase the team’s average xG. Anyone have a scatterplot of teams by average xG vs total xG?
 

Am I crazy to think Cushing being selected MLS coach of the week is further proof that the MLS has no evidence that Cushing did what TOR accused him of? If there was still an ongoing investigation I can't see the MLS making Cushing coach of the week and drawing attention to him.

Herdman needs to face some serious consequences for accusing a fellow coach of such serious misconduct at a press conference when he did not see it himself or have any proof.
 
I wonder how teams break down by average xG and how this relates to total xG. I imagine some teams are much quicker to take the first decent look and count on the rule of large numbers working out for them. Other teams will be more patient, foregoing early shots in hopes of manufacturing that higher quality opportunity. As described before, all of those attempts at higher quality that don’t end in shots reduce the total xG. But presumably they would increase the team’s average xG. Anyone have a scatterplot of teams by average xG vs total xG?
I'm not aware of any analysis of that. I think the problem is you are likely to get a finding that teams with a higher proportion of high quality chances to low quality chances tend to score more and win more and congratulations, you've just proved that it pays to be better than the other team. How do you disentangle the effect of shot selection from that of just being really good?
Also I doubt there is any magic ratio. I suspect it is always situation specific and there will be no substitution for the time honored method of needing players to analyze every situation on the fly to judge whether to shoot, dribble, pass, etc. That's a guess and I'm willing to be proven wrong but I just don't see how it would make sense to forego even low probability opportunities to shoot in pursuit of a ratio. MLS teams average 12.3 shots per game. Take 'em when you can.