GWT: Champions League - Semifinals
- Thread starter bluesfan94
- Start date
Vasilevskiy
The cat will be back
cgf
FireBednarsSuccessor
Scandale du Jour
JordanStaal#1Fan
gary69
Registered User
I have Wirtz and Bellingham, the favourite being the one whose team wins their European competition. If they both win, then it's Bellingham because CL is higher level competiton.
Until Euro, which wll have big influence based on who does better. If Mbappe has a great Euro finals and those two have not, then maybe he can edge ahead.
gary69
Registered User
Just shows that Xg formula is flawed and should be adjusted to reflect reality more accurately.
Wee Baby Seamus
Yo, Goober, where's the meat?
close pack means the international tournaments make the difference.
RM win UCL and Brazil win Copa? it'll be Vini.
RM win UCL and England win Euro? it'll be Bellingham if he has a dominant tournament, Kane if he doesn't
Germany have a good Euro and Leverkusen complete the undefeated treble? it'll likely not be Wirtz because he doesn't have the pedigree yet, but it should be.
France win Euro? it's Mbappe
Spain win euro or make the final? it's Rodri.
unfortunately none of you are prepared for the actual outcome
which is Cole Palmer winning the Euro2024 golden boot, PL POTY, PL YPOTY, and subsequently the Ballon d'Or
RM win UCL and Brazil win Copa? it'll be Vini.
RM win UCL and England win Euro? it'll be Bellingham if he has a dominant tournament, Kane if he doesn't
Germany have a good Euro and Leverkusen complete the undefeated treble? it'll likely not be Wirtz because he doesn't have the pedigree yet, but it should be.
France win Euro? it's Mbappe
Spain win euro or make the final? it's Rodri.
unfortunately none of you are prepared for the actual outcome
which is Cole Palmer winning the Euro2024 golden boot, PL POTY, PL YPOTY, and subsequently the Ballon d'Or
Wee Baby Seamus
Yo, Goober, where's the meat?
they'll have the award ceremony in Cannes so they can call it the Palmer d'Or
bluesfan94
Registered User
What if Arsenal wins the EPL and England wins the Euros on the back of Rice?
Not really. I just shows those 2 teams get very lucky with the lack of finish from opponents, instead of defending better than others.
Wee Baby Seamus
Yo, Goober, where's the meat?
still Kane sorry, these folks are not giving Declan Rice a ballon d'or
bluesfan94
Registered User
I just can't imagine a group that lionizes trophies giving the BDO to a player who brought the first trophyless season to his team in a decade.
Wee Baby Seamus
Yo, Goober, where's the meat?
i can imagine a group that lionizes trophies giving the BDO to a guy who captained his team to a Euro championship and led the continent in scoring during the club season
bluesfan94
Registered User
Which is why I think the way England wins is important. If England wins because Foden/Saka/Bellingham/Rice/Palmer score a bunch of goals and Kane is relatively irrelevant, it's harder to see for me.
gary69
Registered User
But that's not what Xg should be about?
If the reality is that goals get scored in certain situations because teams get constantly "lucky", then the formula should be adjusted to reflect the reality more accurately how likely goals are scored in those type of situations, alas to include what you called "luck".
bluesfan94
Registered User
gary69
Registered User
Probably not, but are those two teams really the only anomalies over the years, maybe they are.
There is data for several years from numerous leagues to make informed decisions on possible changes of the formula.
bluesfan94
Registered User
Of course not. The goal isn’t for xG to match G for each team. The goal is to have it match on a broad level. So, while those teams have given up a lot fewer goals compared to their xGA, there are definitely teams that have given up more G than xGA.
Yes, that data is used to update each event’s xG.
I cannot begin to describe how much I loathe xG as a stat or people referring to it as a measuring stick for... Anything, really.
xG takes ZERO context into consideration.
Prime Conte/Allegri Juventus scored 1 goal, then maybe scored another. Then just closed the game out. Probably had 0 xG from min 30 to the end. The opponent probably creates a chance or two and gets their xG up. It tells you nothing of the game that was just played.
xG doesn't tell you who is chasing the game, who's defending the lead, etc etc.
xG takes ZERO context into consideration.
Prime Conte/Allegri Juventus scored 1 goal, then maybe scored another. Then just closed the game out. Probably had 0 xG from min 30 to the end. The opponent probably creates a chance or two and gets their xG up. It tells you nothing of the game that was just played.
xG doesn't tell you who is chasing the game, who's defending the lead, etc etc.
Wee Baby Seamus
Yo, Goober, where's the meat?
it's a piece of the puzzle, not be all and end all. sophisticated analysts absolutely incorporate game state when discussing these things.
- Jan 23, 2017
- 19,584
- 27,118
It's more useful the bigger sample size you have, but there are some obvious flaws in judging a game purely by xG
- It treats all the players the same: Kylian Mbappe and Emre Can generate the same xG from the same chance, when one is obviously more likely to go in.
- It treats all goalkeepers the same.
- It heavily relies on relies on where the shot was taken, and ignores things like defender position between the ball and the goal, what kind of a stance the attacker is shooting from etc.
- xG only generates from shots and not possession. So let's say a team gets a 3 on 1 attack, but fumbles the ball and never manages to take a shot: it's zero xG, which obviously isn't on par with reality.
Sometimes the xG is in line with what happened in the game, sometimes it's not. But it tends to even out in the long run.
- It treats all the players the same: Kylian Mbappe and Emre Can generate the same xG from the same chance, when one is obviously more likely to go in.
- It treats all goalkeepers the same.
- It heavily relies on relies on where the shot was taken, and ignores things like defender position between the ball and the goal, what kind of a stance the attacker is shooting from etc.
- xG only generates from shots and not possession. So let's say a team gets a 3 on 1 attack, but fumbles the ball and never manages to take a shot: it's zero xG, which obviously isn't on par with reality.
Sometimes the xG is in line with what happened in the game, sometimes it's not. But it tends to even out in the long run.
hatterson
Registered User
xG is only a small piece of the puzzle. It's not everything but it's also not nothing.
It's also interesting to look at things like xGOT or PSxG which is basically a Post Shot xG. According to one model PSG had 2.96 xG but only 0.69 xGOT another has them at 3.2xG but only .76 PSxg which completely matches the eye test that PSG bottled a ton of very good chances.
Some xG models also include information on where the defenders are, where the keeper is, etc.
Looking at Real-Bayern the xG from Opta for the game was 3-0.4 for Real however the PSxG was 3.04 - 1.58 since Kane had a couple shots that were very low xG but ended up being very well taken shots so had .35 and .31 PSxG and their goal was a .06 xG chance that turned into a .59 PSxG chance due to the quality of the shot.
Granted I think their PSxG model needs a bit of work since it claims that Joselu's second goal was .65 PSxG but I'm not really sure how you can have a shot directed to where it was with the pace it was actually be stopped by a keeper who was where Neuer was (and in fact they say his first goal was a higher PSxG than his second despite it going directly through Neuer's legs and if he had dropped his legs different he actually makes a save on it. But those type of insanely high danger chances are usually where the models struggle the most because they have so little data. Players don't often end up with open net tap ins, so there's not a lot of events to base it on.
It's also interesting to look at things like xGOT or PSxG which is basically a Post Shot xG. According to one model PSG had 2.96 xG but only 0.69 xGOT another has them at 3.2xG but only .76 PSxg which completely matches the eye test that PSG bottled a ton of very good chances.
Some xG models also include information on where the defenders are, where the keeper is, etc.
Looking at Real-Bayern the xG from Opta for the game was 3-0.4 for Real however the PSxG was 3.04 - 1.58 since Kane had a couple shots that were very low xG but ended up being very well taken shots so had .35 and .31 PSxG and their goal was a .06 xG chance that turned into a .59 PSxG chance due to the quality of the shot.
Granted I think their PSxG model needs a bit of work since it claims that Joselu's second goal was .65 PSxG but I'm not really sure how you can have a shot directed to where it was with the pace it was actually be stopped by a keeper who was where Neuer was (and in fact they say his first goal was a higher PSxG than his second despite it going directly through Neuer's legs and if he had dropped his legs different he actually makes a save on it. But those type of insanely high danger chances are usually where the models struggle the most because they have so little data. Players don't often end up with open net tap ins, so there's not a lot of events to base it on.
hatterson
Registered User
I referenced this above, but some xG models (like Opta) do include defender and keeper position data as well as things like phase of play (counter vs established possession vs set piece, etc)
Granted your point about defender position in between is valid since statistically a block is not credited as a shot and thus is automatically 0 xG in a shot based xG model.
Ad
Upcoming events
-
-
-
-
-
BET ON ONLY ONE HORSE WATCH LIVE - FINGER LAKES RACE 1 - Farmington, NYWagers: 1Staked: $20.00Event closes- Updated: