| xG or Expected Goals... 13:59 - Sep 28 with 5661 views | itfcjoe |
....what are people's thoughts on this measure which is becoming popular?
We rank poorly for it this season, but I think it is such a crude measure and takes so little into account.
For more on it - look here - https://experimental361.com/
I personally think it is very flawed:
https://experimental361.com/2017/09/28/championship-timelines-26-27-sep-2017/
Apparently the scoreline flattered us on Tuesday and we should only have scored '1.6 goals' from our chances.
Was e-mailing about this earlier - here is some of my reasoning for not particularly liking it - would be keen to hear others thoughts
With regards to graphics on experimental361, I look at them and find
them interesting but they don't show much - I personally feel like xG
or Expected Goals is a fairly flawed statistic for a number of reasons
- mainly:
- Goals change games, if you are 1-0 up then you are likely to create
less good chances than if you are 1-0 down. As soon as a goal goes in
the dynamic of the game changes. Compared to brentford their position
looks like we should swap with them, but when one team has been
predominantly winning games and the other losing they have been
playing from such different positions.
Against Leeds, our expected goals was 1.4 whereas vs Sunderland is
1.61. We could have had 10 on Tuesday night - it uses shot locations
rather than anything else. Celina's goal on Tuesday should have been
the best part of a goal as was simple after the move. What was Wards?
It was probably set as a shot from a wide angle against a GK so a low
%age chance - when in reality is was nailed on once he was past the
keeper.......but should the xG have picked him up in front of the
keeper and measured from there? So the type of goals you score deem
how succesful you are on there.
xG doesn't take into account things like simple pull backs being
messed up, or poor execution when in a good position etc
Also - sample size is key. 46 games is too small a sample even let
alone 9 or 10.
I think statisiticians are looking for a golden bullet on football
stats like happened in baseball. But we have seen xG held up to a high
standard, like 'Chances Created' and 'Pass completion % in Final 3rd'
have been previously for individual players - but ultimately football
is too much of a fluid team game to pull something out - and every
game is different as styles change. |  |
| |  |
| xG or Expected Goals... on 14:02 - Sep 28 with 5643 views | J2BLUE |
Total b0llocks. People desperate to find some stat which agrees with their preformed opinions. There are so many stats now they are pretty much meaningless. You can prove anything if you look hard enough. I'll continue to look at the league table, the agreed method of assessing teams at the beginning of each season. |  |
| |
| xG or Expected Goals... on 14:04 - Sep 28 with 5627 views | SpruceMoose |
I think it is something only dweebs would get excited about.
An xP, or Expected Punts, stat however would catch my interest. |  |
| Pronouns: He/Him/His.
"Imagine being a heterosexual white male in Britain at this moment. How bad is that. Everything you say is racist, everything you say is homophobic. The Woke community have really f****d this country." | | Poll: | Selectamod |
| |
| xG or Expected Goals... on 14:05 - Sep 28 with 5620 views | tcblue |
Fully expected this to be a wind up.
Expected goals? Still, maybe they'll give out a trophy for them at the end of the season, something for those participants up the road to put in their hall of fame |  | | |
| xG or Expected Goals... on 14:05 - Sep 28 with 5621 views | Guthrum |
Too blunt an instrument for anything but the most general conclusions. Doesn't take into account statistical outliers like our extreme goalscoring efficiency.
Fairly concerning stats... by Guthrum 18 Sep 2017 23:14Those charts are still a bit of a blunt instrument, even the "expected goals" one. The author is looking at statistics of the number of goals scored by shots from a particular position. All well and good, more precise than overall shots "on target". However, it fails to take into account the power of the shot, positioning of the goalkeeper and whether he can visually track the flight of the ball. How secure was the 'keeper's footing? Was the shot aerial or along the ground? Even weather conditions and divots in the turf can make a difference.
Eventually, you have to start allowing for near-random factors (like manufacturing flaws in the ball) and the whole equation breaks down.
He does go some way towards this with the set of caveats at the end. But, while statistics can be a guide to what is likely to happen, they are no guarantor of what will actually come to pass in the future.
Town may appear to be poor defensively and weak in attack, but results so far do not bear that out. And how large a sample do you need before "luck" turns into "skill"?
|  |
| |
| xG or Expected Goals... on 14:07 - Sep 28 with 5614 views | portmanking |
xG is something that has been borne out of professional sports betting and trading where statisticians and traders have sought to use this model as a means of pricing up games.
I agree that it's far too rudimentary and doesn't take into consideration the plethora of in-game elements that can influence any particular attack. | | | |
| xG or Expected Goals... on 14:11 - Sep 28 with 5604 views | JakeITFC |
Think it's difficult to take the box score in isolation, but the measure is pretty much tried and tested (albeit the models differ from statistician to statistician) and you'd expect teams to regress back to the mean eventually.
It's obviously a worry that we are allowing lots of chances and not creating many. That being said, the worm has turned somewhat - as often happens, momentum breeds confidence and we are now a better side for our start. |  | | |
| xG or Expected Goals... on 14:14 - Sep 28 with 5592 views | Reuser_is_God |
Load of bollox |  |
| |
| xG or Expected Goals... on 14:20 - Sep 28 with 5574 views | SpruceMoose |
What I take from that graph is that Norwich are bland. A meh attack and a meh defense. At least it is suggesting we are outrageously bad at two things, which in turn leads to excitement right!? | |
| Pronouns: He/Him/His.
"Imagine being a heterosexual white male in Britain at this moment. How bad is that. Everything you say is racist, everything you say is homophobic. The Woke community have really f****d this country." | | Poll: | Selectamod |
| |
Login to get fewer ads
| xG or Expected Goals... on 14:20 - Sep 28 with 5573 views | LankHenners |
As a stat in itself for people to look at it's flawed for the reasons you give, and likewise I don't think it really says anything as a piece of analysis to be used as a prediction. What about own goals and penalties, which can be conceded through a range of methods from a bad challenge to a refereeing mistake?
The sample size would have to be huge before you could even begin to see if it tells you anything worthwhile, and by that point a team (or teams) could/would have changed by having better players/manager (or the reverse) and thus have better/worse defences/attacks anyway.
It was you in another thread saying that football can be a simple game often made more difficult wasn't it? Similar story here - statswise people trying to get too 'geeky' without producing anything of real use or value. |  |
| |
| xG or Expected Goals... on 14:21 - Sep 28 with 5570 views | itfcjoe |
| xG or Expected Goals... on 14:11 - Sep 28 by JakeITFC |
Think it's difficult to take the box score in isolation, but the measure is pretty much tried and tested (albeit the models differ from statistician to statistician) and you'd expect teams to regress back to the mean eventually.
It's obviously a worry that we are allowing lots of chances and not creating many. That being said, the worm has turned somewhat - as often happens, momentum breeds confidence and we are now a better side for our start. |
I think the only way it works for me, is for someone to sit down and analyse chances individually - almost in a Pro Football Focus way for 'soccer'. There would be some value in that although difficult to get the same grading system of chances across the board and would be very time consuming
Shot locations just don't mean that much to me as too many other factors around them - and then looking at the end of game to see xG offers nothing really - because Goals change Games | |
| |
| xG or Expected Goals... on 14:27 - Sep 28 with 5540 views | SpruceMoose |
Why aren't Brentford and Millwall higher in the table if they are in the top five best attacks and defenses? | |
| Pronouns: He/Him/His.
"Imagine being a heterosexual white male in Britain at this moment. How bad is that. Everything you say is racist, everything you say is homophobic. The Woke community have really f****d this country." | | Poll: | Selectamod |
| |
| xG or Expected Goals... on 14:29 - Sep 28 with 5527 views | Illinoisblue |
| xG or Expected Goals... on 14:02 - Sep 28 by J2BLUE |
Total b0llocks. People desperate to find some stat which agrees with their preformed opinions. There are so many stats now they are pretty much meaningless. You can prove anything if you look hard enough. I'll continue to look at the league table, the agreed method of assessing teams at the beginning of each season. |
wht J2 said. |  |
| |
| xG or Expected Goals... on 14:43 - Sep 28 with 5483 views | JakeITFC |
| xG or Expected Goals... on 14:02 - Sep 28 by J2BLUE |
Total b0llocks. People desperate to find some stat which agrees with their preformed opinions. There are so many stats now they are pretty much meaningless. You can prove anything if you look hard enough. I'll continue to look at the league table, the agreed method of assessing teams at the beginning of each season. |
Do you think that odds compilers know what they are doing?
Ipswich have opened 23/10 at home to Bristol City this week, despite being the league's highest scorers, ahead of Bristol City in the league and at home.
That price is almost entirely based on xG, shots on target (taken/allowed) and shot conversion rates. | | | |
| xG or Expected Goals... on 14:45 - Sep 28 with 5468 views | itfcjoe |
| xG or Expected Goals... on 14:43 - Sep 28 by JakeITFC |
Do you think that odds compilers know what they are doing?
Ipswich have opened 23/10 at home to Bristol City this week, despite being the league's highest scorers, ahead of Bristol City in the league and at home.
That price is almost entirely based on xG, shots on target (taken/allowed) and shot conversion rates. |
And it's printing money.... | |
| |
| xG or Expected Goals... on 14:47 - Sep 28 with 5460 views | JakeITFC |
| xG or Expected Goals... on 14:45 - Sep 28 by itfcjoe |
And it's printing money.... |
Well I happen to agree - as I said, we are on an upward trend. | | | |
| xG or Expected Goals... on 14:54 - Sep 28 with 5437 views | J2BLUE |
| xG or Expected Goals... on 14:43 - Sep 28 by JakeITFC |
Do you think that odds compilers know what they are doing?
Ipswich have opened 23/10 at home to Bristol City this week, despite being the league's highest scorers, ahead of Bristol City in the league and at home.
That price is almost entirely based on xG, shots on target (taken/allowed) and shot conversion rates. |
Odds compilers set a starting point designed to attract money to all outcomes. As money comes in they will adjust the price. Of course they base the starting price on stats and as we're an unfashionable club we don't attract much money.
23/10 looks massive value | |
| |
| xG or Expected Goals... on 15:28 - Sep 28 with 5394 views | SW6Tractor |
There are too many variables in what is a fluid sport to make it robust. So many aspects you cannot account for with limited data available. |  | | |
| xG or Expected Goals... on 15:29 - Sep 28 with 5391 views | chicoazul |
This kind of stuff is some of the worst sht out there. |  |
| |
| xG or Expected Goals... on 15:49 - Sep 28 with 5349 views | ITFCBlues |
| xG or Expected Goals... on 14:45 - Sep 28 by itfcjoe |
And it's printing money.... |
Tuesday was printing money - was nailed on and we were priced as if Sunderland are the better side |  |
| |
| xG or Expected Goals... on 16:28 - Sep 28 with 5303 views | Herbivore |
I've yet to see any team or player expected goals stat match the actual goals scored, this suggests it's a load of guff. |  |
| |
| xG or Expected Goals... on 16:32 - Sep 28 with 5292 views | SomethingBlue |
Might post longer thoughts another time but I think it's snake oil. | |
| |
| xG or Expected Goals... on 16:37 - Sep 28 with 5274 views | itfcjoe |
| xG or Expected Goals... on 16:32 - Sep 28 by SomethingBlue |
Might post longer thoughts another time but I think it's snake oil. |
Would love to hear them | |
| |
| xG or Expected Goals... on 16:38 - Sep 28 with 5273 views | JakeITFC |
| xG or Expected Goals... on 16:32 - Sep 28 by SomethingBlue |
Might post longer thoughts another time but I think it's snake oil. |
I don't think it's the Rosetta Stone when it comes to cracking football, but the logic itself is indisputable, surely? | | | |
| xG or Expected Goals... on 16:40 - Sep 28 with 5265 views | Mullet |
Utter numberwang innit.
Literally doesn't add up. Like most stats, utterly pointless on their own let alone combined with others without the requisite analysis and context.
Football is not a science, it's far more than that and that's why we love it and why it needs words, pictures and Morrissey quotes. |  |
| |
| |
By continuing to use the site, you agree to our use of cookies and to abide by our Terms and Conditions.
