Xg from WBA game

[Laughing]

Ok, but is the second adjusted for the first?

If the first chance is rated 0.5 xG (for sake of argument), is the xG for the second reduced proportionately? Surely it doesn't assume there was potential for us to score 2 goals from that passage of play?

Maybe not in the example we saw tonight, but if 2 sitters are missed, one after the other in the same passage of play, the xG on the chances combined cannot total more than 1?

It's an anomoly of xg. It has no memory.

There's lots of times where xg is silly. Overall it's the best measure of goal attempts and their quality we have.

To be clear, according to xg neither were sitters. For azaz chance 1 in 10 or goes in for Mcgree 1 in 3 it goes in. Neither sitters.

 

[B-Man]

It was the Maja chance late on. 0.83 xG

 

[BiggEggo]

I know xg is a contentious issue but how on earth can they determine that WBA had a higher xg than us in that game.



Absolute nonsense

I think those numbers became skewed after our goal when they threw everything at us and we had to dig in.

XG is smoke and mirror stuff, as is possession %.

 

[parnabyscrosses]

I am seemingly the only one here who has no idea what the f**k ‘xG’ is or stands for!

 

[locomansimon]

Xg totally meaningless !

 

[sweet_left_foot]

I've calculated my xF's given...

 

[Gunslinger]

losing the xG battle is another black mark for Carrick and the team

3Pts and a clean sheet barely compensates for it

 

[B_G]

Regarding xG, how does it handle something like McGree's miss?

The miss came immediately after a decent chance and a good save from Azaz. I assume xG is smart enough not to count that as 2 separate Goal scoring opportunities.

McGree's chance doesn't happen if Azaz scores?

Azaz attempt was 0.10 xG (1 in 10 chance)
McGree attempt was 0.33xG (1 in 3 chance)

The more I look at xG. The more flimsy the data is.

 

[Ipad1977]

I think xG is our version of Sunderland's attendance, it's an indicator that we're doing something right (in our case our overall performances, in their case their ticketing approach and admin) and that we're clearly the best in the league in said department. We'll take pride in that whilst our actual league position doesn't deliver the goods (and how much have you heard them mention attendances in the last month or so?)

However once we start getting decent results together, it gets flung in the bin in favour of the far more important metric of the league table.

 

[anth25]

Azaz attempt was 0.10 xG (1 in 10 chance)
McGree attempt was 0.33xG (1 in 3 chance)

The more I look at xG. The more flimsy the data is.

This is the point I was trying to make. There is so much made of Xg in the media and on-line yet anyone watching that game tonight would never believe that WBA had a higher XG.

It doesn't bother me at all as the result is the only important metric. I think people just need to stop placing so much importance on it when it is such a flawed system

 

[Humpty]

I know xg is a contentious issue but how on earth can they determine that WBA had a higher xg than us in that game.



Absolute nonsense

xG has its place but there's far too much importance placed on it when determining and understanding the balance of the game.

 

[changingman]

It was the Maja chance late on. 0.83 xG

I'm surprised that wasn't 2.0xG to be honest. A truly awful miss.

 

[Humpty]

I thought Maja was just lucky to get a toe to it to be honest. I don't think he really ever had much chance of putting it in the net.

 

[BoroPhil]

It's a strange one that Maja miss, I can see why it could be very high xG as he basically misses an open goal from 6 yards out but it looks like he is struggling to make any sort of contact.

We did incredibly well to restrict them to basically just that chance and other.

 

[B_G]

It's a strange one that Maja miss, I can see why it could be very high xG as he basically misses an open goal from 6 yards out but it looks like he is struggling to make any sort of contact.

We did incredibly well to restrict them to basically just that chance and other.

I don't think xG recognises whether a player is balanced or at full stretch etc. I also don't think it takes ito account the height of the ball, just the X, Y position.
It just recognises the place the contact was made, with which part of the body the contact is made and players positions on the pitch.

 

[Laughing]

Azaz attempt was 0.10 xG (1 in 10 chance)
McGree attempt was 0.33xG (1 in 3 chance)

The more I look at xG. The more flimsy the data is.

It does have some problems BG. I argued this point last season. As a tool to predict future outcomes it is quite poor. I know this because I tried to make it work. There are better stats to use to predict future results.

However where it is good is during in-play management to predict who scores the next goal and where the game is going.

With all stats, there is a bell curve and it is wrong a lot of the time. However, over a season, it is accurate on a game by game basis at predicting the results as the game is happening. This is a fact and it's the only way I use the stat now.

From a team perspective for the coaching staff it allows them to see where tweaks may be needed. If, for example you are conceding goals at a higher rate than you would expect then the team shape is probably OK, defensively, and the issue is player mistakes rather than how the team are shaped when defending. We have seen that with boro quite a bit this season.

In the example you gave, over thousands of games in very similar circumstances, a player scores the mcgree chance 1 in 3, and the Azaz chance 1 in 10. That is also a fact. It doesn't take into account some of the circumstances of the chance, but does a better job than any other stat, and, of course, it is objective a quality football fans aint known for.

 

[B-Man]

You’re correct, it doesn’t. I think it takes into account distance from goal, angle, whether it’s a shot or a header maybe, and some also account for how the chance was created and goal keeper positioning, but not all.

As mentioned, the fact Maja didn’t make a full connection isn’t factored in, if it was it’d have been much lower!

Never sure why people get so annoyed by it. It’s better than just looking at shots or possession and deciding if you should’ve won, and worse than watching the game and forming your opinion from the full 90 minutes. Even the creators of it say that.

 

[Laughing]

B

I don't think xG recognises whether a player is balanced or at full stretch etc. I also don't think it takes ito account the height of the ball, just the X, Y position.
It just recognises the place the contact was made, with which part of the body the contact is made and players positions on the pitch.

BG xG takes into account in order of importance

The distance to goal
The angle to the goal
position of defenders
Which body part the ball was struck with
The assist type
the phase of play

You are right it doesn't include how well balanced a player is or whether he was stretching to make contact. It does include how the ball was struck, i.e. header, foot, shin, bum etc.

 

[Laughing]

Never sure why people get so annoyed by it. It’s better than just looking at shots or possession and deciding if you should’ve won, and worse than watching the game and forming your opinion from the full 90 minutes. Even the creators of it say that.

Not entirely true B-man. xG has been around as a stat for years, mid 60's I think it was first used by a statistician called Charles Reep ( I think). He calculated how many shots on average it takes to score a goal. The answer 10!

The term xG was first used in the 1990's and I can't remember by whom but it was part of a study on artificial pitches and how the effect the proffessional game.

Opta took the concept and commercialised it. Free xG ratings are available too, of course.

Not sure Opta ever said you get a better feel for the game by watching it, though, of course you do. However, you would have to say that you get 100 fans watching a game, you get 100 different opinions. We see that on every matchday thread on here!

 

[B_G]

B

BG xG takes into account in order of importance

The distance to goal
The angle to the goal
position of defenders
Which body part the ball was struck with
The assist type
the phase of play

You are right it doesn't include how well balanced a player is or whether he was stretching to make contact. It does include how the ball was struck, i.e. header, foot, shin, bum etc.

Yes, but there's no difference between whether the ball contacts an outstretched toe nail or whether player lashes it through his instep. It just registers "right foot". Obviously this is a huge difference in reality.

Again height of the ball makes a big difference. Maja's chance was a really difficult height which isn't taken into consideration.

Going back to Lath's chance against Sunderland which registered at 0.48. The ball was again at quite an awkward height for Latte and he was off balance and at full stretch (partly down to his first touch, but xG doesn't account for that either). There's no way that was a 1 in 2 chance, or a considerably easier chance than McGree's last night.

I think there's a lot of holes in the way xG is calculated.