Expected Goals - what a bunch of garbage

[abo9]

Thats not much of an answer I guess google didnt help a lot. R2 does not mean a lot and 33% in the real world is not predictive.

This example is made up of random numbers, but bear with me.

Say you wanted to predict the weight of a banana. There's a couple factors you could use: It's length, width, it's color (yellow or green for example), days of growth, etc. Let's say for simplicity that we use length only, it's gonna be a pretty good predictor right? Well you could use a very simple model like:

3*length + e = Weight

Where "e" is an error term, because you omitted other parameters like "days of growth", "color", or "width". And since you don't know ALL the parameters you're missing, we usually refer to this "error" term as "luck" or "randomness" in the model.

In our banana example, if the R^2 was 90%, you would expect the "length" factor to explain 90% of the variance between the weight of each banana you look at, and 10% of the difference in weight would be explained by other factors. It's one of numerous "scoring" methods to assess the quality of a model, but should not be trusted blindly nor used as an be-all end-all. It's a good indicator though.

I guess 33% would still be fairly low in terms of "absolutes", but if that's the benchmark to beat, we just have to work to find the next best metric/model. If you want to use previous year's standings to predict this year's standings, it's a fair model to use, usually called a "naive prediction mode". I believe it would actually be the benchmark used to build more complex models trying to predict standings. That said, you could find that your "naive" model gives fairly strong results even compared to models incorporating goals scored and goals allowed (I have not looked into that, just an example), and decide that it's not worth the effort to build more complex stuff.

 

[Fourier]

so your username is Fourier and people just assume you don't know the math?

Idk if your username is related or not, and it should obviously not be a basis for authority, but I find it quite ironic.

It could have been Euler. Since I am from Edmonton.

In fact, I have made my living studying Fourier algebras.

 

[abo9]

It could have been Euler. Since I am from Edmonton.

In fact, I have made my living studying Fourier algebras.

What's the relation between Euler and Edmonton?

 

[Fourier]

What's the relation between Euler and Edmonton?

It's a play on Oiler which is how the name is essentially pronounced by English speakers.

 

[supsens]

Look I told you what it is. If you want a dictionary definition, it's not going to be any more illuminating. 33% of future goal differential can be explained by past xG differential. It's more predictive than other metrics that people were using before, like standings points or goal differential, which is why it's useful.

There's no such thing as "good" or "bad r^2. The interpretation is what you make of it. Real-world data is messy. Getting 33% of the way there, in a game where ~40% is pure luck, is honestly surprisingly effective.

If you want to disagree with how useful it is, go ahead. But hockey has moved past dinosaur thinking and most teams and private analytics companies are conducting analyses in reference to some kind of xG model.

If 40% was pure luck a guy like Mcdavid would bounce between 60 and 100 points from one year to the next and most players would have a massive fluctuations in goals and point totals
It is far from 40% luck.
It would be almost impossible for Mcdavid and Drai to finish 1 and 2 in scoring 2 years in a row

 

[ResilientBeast]

If 40% was pure luck a guy like Mcdavid would bounce between 60 and 100 points from one year to the next and most players would have a massive fluctuations in goals and point totals
It is far from 40% luck.
It would be almost impossible for Mcdavid and Drai to finish 1 and 2 in scoring 2 years in a row

I feel like you're grossly lacking a background in statistics to make good faith criticisms

 

[ResilientBeast]

This video from Vox does a pretty good job explaining "luck" in professional sports

 

[supsens]

I feel like you're grossly lacking a background in statistics to make good faith criticisms

What would you say the margin of error would be? What would you say the % of available data is off
Now when the majority of the goalies all are within a 4% save percentage of each other how much havoc would that cause?
A nice looking graph when the results are based on a 4% variance will almost always plot nice and have a good r value to confirm your doing it right will it not?

These one size fits all model ignoring who the shooter is and who the goalie is based on the assumption everyone has the same skill level is doomed to fail.

 

[ResilientBeast]

What would you say the margin of error would be? What would you say the % of available data is off
Now when the majority of the goalies all are within a 4% save percentage of each other how much havoc would that cause?
A nice looking graph when the results are based on a 4% variance will almost always plot nice and have a good r value to confirm your doing it right will it not?

These one size fits all model ignoring who the shooter is and who the goalie is based on the assumption everyone has the same skill level is doomed to fail.

I didn't do the studies so I can't answer you, but I'm sure if google and read how they were developed you'd find the information you're seeking

Glossary

Expected Goals – (xG) – is a stat examining shot quality in attempt to guess the goal expectancy of players and teams. Expected Goals is like WAR in baseball in that one isn’t going to try to explain the calculations involved. Instead, one would just explain what is used to determine expected goals. In trying to assign goal expectancy to shots, these factors are used: Shot type (Wrist/slap shot, deflection, etc.), Shot distance, Shot angle, Rebounds, Shots of the rush, & Strength (such as 5v5 or power play).

This is from 2015 but does a pretty decent job

Expected Goals are a better predictor of future scoring than Corsi, Goals

 

[Midnight Judges]

You guys can correct me if I'm wrong but xGoals does not assess the quality of individual player's shots and factor that into predictions, right?

Somebody mentioned to me a few years ago that they were considering that. Not sure if it has changed since then.

 

[supsens]

I didn't do the studies so I can't answer you, but I'm sure if google and read how they were developed you'd find the information you're seeking

Glossary




This is from 2015 but does a pretty decent job

Expected Goals are a better predictor of future scoring than Corsi, Goals

Here is a working example of real world vrs a model dealing with the average
You shoot at 15% I shoot at 10% no matter what
The expected goals is now 12.5%
No matter what I will never reach the expected goals and no matter what you will always be above the expected goals.
Do you honestly think the predictions will be right for either of us just because of “average”

 

[ResilientBeast]

Here is a working example of real world vrs a model dealing with the average
You shoot at 15% I shoot at 10% no matter what
The expected goals is now 12.5%
No matter what I will never reach the expected goals and no matter what you will always be above the expected goals.
Do you honestly think the predictions will be right for either of us just because of “average”

Ok so I don't think you understand what expected goals means and how it's developed

 

[ResilientBeast]

Your shooting % example is not reflective of expected goals.

Expect goals is a formula derived from a regression of a variety of different factors that result in a goal

Here is the list sorted by importance (from Money Puck model): Expected Goals (xG) Models Explained

1. Shot Distance From Net
2. Time Since Last Game Event
3. Shot Type (Slap, Wrist, Backhand, etc)
4. Speed From Previous Event
5. Shot Angle
6. East-West Location on Ice of Last Event Before the Shot
7. If Rebound, difference in shot angle divided by time since last shot
8. Last Event That Happened Before the Shot (Faceoff, Hit, etc)
9. Other team’s # of skaters on ice
10. East-West Location on Ice of Shot
11. Man Advantage Situation
12. Time since current Powerplay started
13. Distance From Previous Event
14. North-South Location on Ice of Shot
15. Shooting on Empty Net

Based upon the values derived from NHL play-by-play data a formula takes these inputs and spits out an "expected goal" which is effectively the probability that said shot results in a goal.

So you "real world" versus average example doesn't make any sense. And again just because the probability of something occurring is low does not mean it will not happen.

 

[supsens]

Ok so I don't think you understand what expected goals means and how it's developed

Yes I do, I did not feel like typing out 6% of my shots from zone a 7% from zone b blah blah in the end it does not change what I said. I’m starting to think you don’t understand how it works
On top of that do you have any idea if some teams are always above and some teams are always below because since it’s an average. 50% of the time you will have 1 team over and 1 team under.... ish
The “advanced” stat is a basic generalized trend that is actually further away from digging into the why and how of individual teams and players

 

[supsens]

Your shooting % example is not reflective of expected goals.

Expect goals is a formula derived from a regression of a variety of different factors that result in a goal

Here is the list sorted by importance (from Money Puck model): Expected Goals (xG) Models Explained

1. Shot Distance From Net
2. Time Since Last Game Event
3. Shot Type (Slap, Wrist, Backhand, etc)
4. Speed From Previous Event
5. Shot Angle
6. East-West Location on Ice of Last Event Before the Shot
7. If Rebound, difference in shot angle divided by time since last shot
8. Last Event That Happened Before the Shot (Faceoff, Hit, etc)
9. Other team’s # of skaters on ice
10. East-West Location on Ice of Shot
11. Man Advantage Situation
12. Time since current Powerplay started
13. Distance From Previous Event
14. North-South Location on Ice of Shot
15. Shooting on Empty Net

Based upon the values derived from NHL play-by-play data a formula takes these inputs and spits out an "expected goal" which is effectively the probability that said shot results in a goal.

So you "real world" versus average example doesn't make any sense. And again just because the probability of something occurring is low does not mean it will not happen.

Hasek in his prime would not let you reach expected goals a prediction that he will allow the average amount of goals is wrong
A large goalie who is positional sound may not let you score in close again you can throw those expected goals out the window
A slow goalie that can’t move left to right will have a higher % of goals against than expected from further out.
Using a basic average generalization is not “predictive”

 

[ResilientBeast]

Yes I do, I did not feel like typing out 6% of my shots from zone a 7% from zone b blah blah in the end it does not change what I said. I’m starting to think you don’t understand how it works
On top of that do you have any idea if some teams are always above and some teams are always below because since it’s an average. 50% of the time you will have 1 team over and 1 team under.... ish
The “advanced” stat is a basic generalized trend that is actually further away from digging into the why and how of individual teams and players

You're getting too granular lol

All the factors I listed above impact whether a specific shot is a goal. The summation of those various factors were subjected to a regression to assign weighting regarding their impact. There is no "averaging" here.

Read the article I linked in the last post it specifically details how it's calculated.

 

[ResilientBeast]

Hasek in his prime would not let you reach expected goals a prediction that he will allow the average amount of goals is wrong
A large goalie who is positional sound may not let you score in close again you can throw those expected goals out the window
A slow goalie that can’t move left to right will have a higher % of goals against than expected from further out.
Using a basic average generalization is not “predictive”

They're saying based upon shot quality the probability of a shot being a goal. Facing an exceptional goalie does not change this result.

Despite complaining about how shitty you think the metric is, you seem to ascribe a lot of belief that people think it's super meaningful and perfectly describes hockey. You're evidently just here to yell about how bad you think the stat is, can't you find a better usage of your time?

 

[supsens]

You're getting too granular lol

All the factors I listed above impact whether a specific shot is a goal. The summation of those various factors were subjected to a regression to assign weighting regarding their impact. There is no "averaging" here.

Read the article I linked in the last post it specifically details how it's calculated.

Averaging that is all it is. Lol how do you think they end up with these numbers

 

[ResilientBeast]

Averaging that is all it is. Lol how do you think they end up with these numbers

Regression? That's not "averaging"

I'm providing you resources that help explain the metric, you don't bother reading them and then just keep parroting the same bs

 

[ResilientBeast]

Here is a working example of real world vrs a model dealing with the average
You shoot at 15% I shoot at 10% no matter what
The expected goals is now 12.5%
No matter what I will never reach the expected goals and no matter what you will always be above the expected goals.
Do you honestly think the predictions will be right for either of us just because of “average”

God the more I think on this "example" it doesn't even really relate to the topic at hand.

I take 100 shots at 15%
You take 100 shots at 10%

Between the two of us the shooting average is 12.5%. So?

That doesn't tell us anything about expected goals since we're missing so much information. It tells us that I score 15 goals in 100 shots and you score 10 goals in 100 shots. Big deal

 

[bossram]

If 40% was pure luck a guy like Mcdavid would bounce between 60 and 100 points from one year to the next and most players would have a massive fluctuations in goals and point totals
It is far from 40% luck.
It would be almost impossible for Mcdavid and Drai to finish 1 and 2 in scoring 2 years in a row

I feel like you're grossly lacking a background in statistics to make good faith criticisms

Yah, it's best to not bother with this guy anymore. You can't reason with a guy that has no clue what he's talking about.

I'm literally referencing team level correlations between xG and goal differential and then this guy comes to me with "lel y didn't u predct McD getting 60 points then GOTEEEM"

This guy thinks a regression is averaging, so there isn't much to work with here.

 

[supsens]

Yah, it's best to not bother with this guy anymore. You can't reason with a guy that has no clue what he's talking about.

I'm literally referencing team level correlations between xG and goal differential and then this guy comes to me with "lel y didn't u predct McD getting 60 points then GOTEEEM"

This guy thinks a regression is averaging, so there isn't much to work with here.

Sure and hockey is 40% luck because your math does not work.

 

[ResilientBeast]

Sure and hockey is 40% luck because your math does not work.

Lol it's not his math

Did you read or watch anything we posted to help show you what we meant?

 

[supsens]

Lol it's not his math

Did you read or watch anything we posted to help show you what we meant?

I don’t care who’s model it is when it’s wrong it’s wrong
writing it off as luck is ridiculous

 

[ResilientBeast]

I don’t care who’s model it is when it’s wrong it’s wrong
writing it off as luck is ridiculous

Ok so you're not here for a conversation at all you're here to miss the point entirely and keep talking about math you don't grasp.