Away Match Day Thread 18/9/21 L1: Cheltenham Town v OUFC

[holdsteady]

Didn’t Arsenal Wenger (at the end of time at Arsenal when it had all gone a bit to s**t) suggest that the table should be decided on expected goals (Arsenal we’re doing better that way obviously)?

It’s a desperate stat that doesn’t really matter, we know we have a lot of crosses going no where near our players and shots from distance heading towards the corner flag, just like we know we have more of the ball than the opposition because we pass it around the back a lot.

I reckon we will play with urgency and pace and be more direct and get a result this Saturday, just worry that we will then revert to the tippy tappy stuff as soon as we have a few more points on the board.

 

[Big Ronaldo]

XG is one of those quantifiers that should be consigned to the
N-Zone, touchdown, gridiron dumpster.

I tire of theoretical what-if jargon that has little or no use.
You don’t need stats to know that we ain’t scoring.
You don’t need an effin stat to tell you that we are struggling with defending and attacking high balls and set pieces.

What those areas need are working on on the training ground.

The only stat we need is a 0 in the goals conceded and a whole lot of numbers in the goals scored.

I believe that we have the personnel in the building to do all of this and I also believe that we have the manager to implement it.
If it ain’t right…
Practice, practice f*****g practice.
Get those players up the Ka$$am and f*****g practice.
Practice until they do things right.
They are professional players!
What else do they do?

 

[werthersoriginal]

I am brave - I'll give it a go. Although it may be the blind leading the partially sighted!

xG = eXpected Goals

Pinched off t'Internet:

"Expected Goals uses various characteristics of the shots being taken together with historical data of such types of shots to predict the likelihood of a specific shot being scored. Since xG is simply an averaged probability of a shot being scored, a team or player may outperform or underperform their xG value. This means that they could be scoring chances that the average player would miss or that they could be missing chances that are often scored."

As I understand it, that is the basic version. Others add in more stats to include number of crosses, corners etc - which may be more (or less) accurate!

You can see the 'FootyStats' L1 xG here: https://footystats.org/england/efl-league-one/xg

Thanks! It's an odd mix of science and vagueness - 'The reason for using more data points to calculate xG is because many attacks can be extremely dangerous without a shot attempt'.

 

[Manorlounger]

They are professional players!
What else do they do?

Apparently some go fishing.... & play golf.

 

[Oufcrealist]

Converted chances get goals.

The argument put across was related to XG this wonderful new science of showing how good everything is. (despite losing)

I agree with you on this. XG is another example of statistical baloney.

 

[AbsoluteZero]

Last season we were at our absolute best when our full backs went on a goalscoring and assists spree and their performances completely transformed the team. In Seddon and Williams we appear to have natural left and right wing backs who can cause lots of problems going forward. So play them higher up the pitch as proper wing backs and play 3 at the back, 3 from Long, Moore, McNally and Thornily. Gorrin gets taken out of the team. The game will naturally be played further up the pitch and hopefully support the lone striker better. It could go spectacularly wrong if the players can't work it out and don't help out defensively when needed and there could be times when Gorrin or Kane does need to sit in front of the defence when we are playing away against a strong team. But we need to try something different as the stubborn 4-3-3 definitely is not working well enough to actually win enough matches and are now stuck in a habit of being on the wrong end of a close game despite us dominating all of the stats except goals scored.

 

[Mad Dog]

I am brave - I'll give it a go. Although it may be the blind leading the partially sighted!

xG = eXpected Goals

Pinched off t'Internet:

"Expected Goals uses various characteristics of the shots being taken together with historical data of such types of shots to predict the likelihood of a specific shot being scored. Since xG is simply an averaged probability of a shot being scored, a team or player may outperform or underperform their xG value. This means that they could be scoring chances that the average player would miss or that they could be missing chances that are often scored."

As I understand it, that is the basic version. Others add in more stats to include number of crosses, corners etc - which may be more (or less) accurate!

You can see the 'FootyStats' L1 xG here: https://footystats.org/england/efl-league-one/xg

So basically a complete load of B*****s then!

 

[Oufcrealist]

So basically a complete load of B*****s then!

Yes as I stated above but I used a much less offensive word

 

[Shosho]

As I use xG and various other metrics at a variety of clubs I feel compelled to chip in!

An xG value is only assigned when a shot is taken. Different models take different situations into account. For example, the xG model I created is a 'naive' one in that it assigns a value based on the distance and angle to goal along with if its a headed/foot chance, or set piece/open play. More sophisticated models are trained on goal keeper position/defenders between attacker and ball etc.

Models are trained on hundreds of thousands of shots...(think my model is trained on ~200k shots from previous years). The xG assigned is simply a % chance of a goal compared to other shots in the data from that distance/location etc. Where xG is limited is in isolated shots/matches due to variance. Opta have Oxford as 8.8xG from open play (2nd in the league) with 1.2xG from set pieces (ha!). Wimbledon have 4xG from set pieces and 9 goals...this most probably will not be sustained, however they have hired a specialist set piece coach.

Whilst xG is an indicator it confirms what we've seen with our eyes so far (and is probably worth an extensive look)...we regularly find ourselves in positions to cut back (these create high xG chances due to being close to the goal and central) however have had shots blocked. Personally, I believe these will come good (how many times does Sykes get into a position to cut back on the right edge of the 18 yard box?). If being critical, we know Taylor will always occupy the back post...its therefore reliant on our midfielders to get into positions around the penalty spot. The goal we *appear* to be trying to score this season is one Man City and Sterling have mastered. Unfortunately, when teams execute a low block we resort to relentless crossing (which as a strategy is a poor one - whilst looking like you are doing something productive).

 

[ZeroTheHero]

Thanks! It's an odd mix of science and vagueness - 'The reason for using more data points to calculate xG is because many attacks can be extremely dangerous without a shot attempt'.

I'm not sure it's that useful TBH.

The reason I say that is that wherever a shot is taken from, surely a better player is more likely to score it? So if you were a team that had three red hot strikers you could have a low xG and score a shedload and conversely you can create all the chances you like but if your lone striker can't hit a barn door with a banjo then no matter how high the xG you aren't going to win any games! And for xGA (eXpected Goals Against) the same is true if you have a defence adept at blocking the ball and a top class goalie as opposed to a rubbish defence and a Mr Butterfingers.

I think what it *might* show you is how many decent chances you create and it's probably a better measure of how you did in the game than 'possession' which is worse than useless. It does of course give managers another stat to hide behind when their team loses!

 

[Berliner]

No statistic engine will ever be invented that can measurably determine the outcome of James Henry shooting from 30 yards one week and electing to pass while clean through on the other, or whether Mark Sykes will score or collapse like a souffle in front of goal, or whether Sam Winnall will all of a sudden smash in a 30 yard one bounce volley. This is where it all falls down.

 

[Manorlounger]

I'm not sure it's that useful TBH.

The reason I say that is that wherever a shot is taken from, surely a better player is more likely to score it? So if you were a team that had three red hot strikers you could have a low xG and score a shedload and conversely you can create all the chances you like but if your lone striker can't hit a barn door with a banjo then no matter how high the xG you aren't going to win any games! And for xGA (eXpected Goals Against) the same is true if you have a defence adept at blocking the ball and a top class goalie as opposed to a rubbish defence and a Mr Butterfingers.

I think what it *might* show you is how many decent chances you create and it's probably a better measure of how you did in the game than 'possession' which is worse than useless. It does of course give managers another stat to hide behind when their team loses!

My thoughts as well. If it doesn't result in a goal then, as far as I am concerned, it's not really relevant. It may prove entertaining to have lots of shots but if we don't score, and concede, then we lose.
It bugs me when any manager claims that his side was the better when losing. He may be right but, it's still a defeat.

 

[Coffin dodger]

As I use xG and various other metrics at a variety of clubs I feel compelled to chip in!

An xG value is only assigned when a shot is taken. Different models take different situations into account. For example, the xG model I created is a 'naive' one in that it assigns a value based on the distance and angle to goal along with if its a headed/foot chance, or set piece/open play. More sophisticated models are trained on goal keeper position/defenders between attacker and ball etc.

Models are trained on hundreds of thousands of shots...(think my model is trained on ~200k shots from previous years). The xG assigned is simply a % chance of a goal compared to other shots in the data from that distance/location etc. Where xG is limited is in isolated shots/matches due to variance. Opta have Oxford as 8.8xG from open play (2nd in the league) with 1.2xG from set pieces (ha!). Wimbledon have 4xG from set pieces and 9 goals...this most probably will not be sustained, however they have hired a specialist set piece coach.

Whilst xG is an indicator it confirms what we've seen with our eyes so far (and is probably worth an extensive look)...we regularly find ourselves in positions to cut back (these create high xG chances due to being close to the goal and central) however have had shots blocked. Personally, I believe these will come good (how many times does Sykes get into a position to cut back on the right edge of the 18 yard box?). If being critical, we know Taylor will always occupy the back post...its therefore reliant on our midfielders to get into positions around the penalty spot. The goal we *appear* to be trying to score this season is one Man City and Sterling have mastered. Unfortunately, when teams execute a low block we resort to relentless crossing (which as a strategy is a poor one - whilst looking like you are doing something productive).

So not B*****s it is absolute B*****s

 

[Mad Dog]

As I use xG and various other metrics at a variety of clubs I feel compelled to chip in!

An xG value is only assigned when a shot is taken. Different models take different situations into account. For example, the xG model I created is a 'naive' one in that it assigns a value based on the distance and angle to goal along with if its a headed/foot chance, or set piece/open play. More sophisticated models are trained on goal keeper position/defenders between attacker and ball etc.

Models are trained on hundreds of thousands of shots...(think my model is trained on ~200k shots from previous years). The xG assigned is simply a % chance of a goal compared to other shots in the data from that distance/location etc. Where xG is limited is in isolated shots/matches due to variance. Opta have Oxford as 8.8xG from open play (2nd in the league) with 1.2xG from set pieces (ha!). Wimbledon have 4xG from set pieces and 9 goals...this most probably will not be sustained, however they have hired a specialist set piece coach.

Whilst xG is an indicator it confirms what we've seen with our eyes so far (and is probably worth an extensive look)...we regularly find ourselves in positions to cut back (these create high xG chances due to being close to the goal and central) however have had shots blocked. Personally, I believe these will come good (how many times does Sykes get into a position to cut back on the right edge of the 18 yard box?). If being critical, we know Taylor will always occupy the back post...its therefore reliant on our midfielders to get into positions around the penalty spot. The goal we *appear* to be trying to score this season is one Man City and Sterling have mastered. Unfortunately, when teams execute a low block we resort to relentless crossing (which as a strategy is a poor one - whilst looking like you are doing something productive).

So basically a complete load of B*****s then!

 

[werthersoriginal]

My thoughts as well. If it doesn't result in a goal then, as far as I am concerned, it's not really relevant. It may prove entertaining to have lots of shots but if we don't score, and concede, then we lose.
It bugs me when any manager claims that his side was the better when losing. He may be right but, it's still a defeat.

Sometimes the better team loses - a goal can involve luck, or come from a penalty or a scramble following a corner. This sounds like it's a way of converting into data some of the things that give us our subjective impressions such as 'we were all over them / couldn't get near them', and I can believe it's more meaningful than possession stats.

 

[Coffin dodger]

So basically a complete load of B*****s then!

There are so many other variables which can come into play that are not mentioned ie:
Weather is the wind with you or against or is it a crosswind, is it freezing cold or boiling hot.
Are you shooting towards your own fans or opposition or even a fence end.
Is it the first minute of the game or the last.
Did you get your end away last night.
And a thousand others.

 

[Mad Dog]

There are so many other variables which can come into play that are not mentioned ie:
Weather is the wind with you or against or is it a crosswind, is it freezing cold or boiling hot.
Are you shooting towards your own fans or opposition or even a fence end.
Is it the first minute of the game or the last.
Did you get your end away last night.
And a thousand others.

So basically....

 

[The mighty Aldo]

I remember us in div 2 and just about in the top division but the game has changed a lot since then - you don’t get the likes of Walsall and Bury in the championship. There is potential at the club as there is with any club in a wealthy area but at the moment we have about 7000 fans and a scruffy unfinished ground, and I’d say we are doing pretty well. If we do get new investment and a definite plan for a new ground the stakes could get higher, as you say.

Walsall and Bury were higher than Bournemouth a few years back, but look at them now.Ever since I started supporting the Yellows Bournemouth were always a fourth tier team, they sometimes went to the third tier So you do get the likes of Walsall and Bury up there.
Wouldn't surprise me to see Fleetwood or Morecambe in the Championship soon.Wycombe were there last season for Gods sake
Any team with a good management structure who can get the right manager has a chance to go through the leagues

 

[Mellowyellow]

Not that I don’t agree with a lot of what u say, but how much better do u expect us to do? In the last two seasons we’ve just had two playoff campaigns in a row, and our highest league finishes in over twenty years, competing with some massive clubs now in our league along the way, and having some fantastic cup runs as well.

I believe that its partly because yes we have finished in the play offs twice, but what kr says he has learnt isnt put unto action. For example needing a more physical presence after Blackpool. Yet here we are again the last few games, going up against physical teams and being muscled off the ball.
He wanted a better quality squad that would take us up. He's now got that, having been backed well by the board, but weekly the same mistakes away from home are made, in a 1 dimensional system that teams already know about.
Is it the players or manager that arent performing?

 

[Daviddidcotexile]

It is interesting how noble came in at the last minute for West ham. No scramble for GPS or some other techy addition. Do other clubs do what we do re all this statistical analysis or do they stand by the side of the pitch laughing whilst they gallop past our defence and score.