Underlying Data (2 Viewers)

SeaSeeEffCee

Well-Known Member
SNADCet.png


nnQeLo1.png


cSPByJa.png


CxZZj9o.png


xUAiYLj.png


Not the best for us in terms of automatic promotion ambitions, but solid all the same. Very tight league this year, no stand out sides.
 

Evo1883

Well-Known Member
You can win the league by winning every match 1 - 0.

Your attacking stats could be poor and your expected goals would be extremely low.

It's the results that matter... If these things meant anything, we'd all be doing well on accumulators every week
 

SeaSeeEffCee

Well-Known Member
We know from last season that xG is not really worth that much, we had a higher xG last season
I disagree, if it wasn't worth much then every single bookie going wouldn't have been using this sort of data for years. The main issue with it is that is tracks long term trends which can mean that sometimes the data won't feel like it's "come true' over the course of a single season.

Two examples; United under Solskjaer last winter and Reading under Jaap Stam in 16/17. When Solskjaer first came in and United won a ridiculous amount of games in a row their underlying numbers were actually pretty awful and by the end of the season the bubble had burst and they'd fallen off a cliff. Stam's Reading on the otherhand managed to defy the numbers for a whole season, finishing third and making the play-offs with relegation standard numbers. They failed to go up, kept a similar side to the season before and recorded similar data but completely fell off a cliff and Stam ended up getting sacked with them in the lower reaches of the table.

This sort of data shouldn't be used as gospel, it's more of a guide than anything. Our data actually shows that we're performing exactly the level our points total and goals f/a would suggest, there is just a fair few teams below us under-performing which make us look a bit worse than we probably are.

You can win the league by winning every match 1 - 0.

Your attacking stats could be poor and your expected goals would be extremely low.

It's the results that matter... If these things meant anything, we'd all be doing well on accumulators every week
That's not really how it works. xG tables take into account xG both for and against, so if you were winning 1-0 every week yeah maybe your xG would be low, but you'd expect the xGA to be equally low. If you were winning games 1-0 but your xGA was much higher than your xG then you'd probably be getting incredibly lucky most weeks.
 

Ring Of Steel

Well-Known Member
What the fuck is this nonsense all about? You can have more points than other teams yet they’re top- yeah ok then

as for the comment about the bookies, if this stuff was true then everyone would be using it and putting the bookies out of business
 

SeaSeeEffCee

Well-Known Member
What the fuck is this nonsense all about? You can have more points than other teams yet they’re top- yeah ok then

as for the comment about the bookies, if this stuff was true then everyone would be using it and putting the bookies out of business
That's not how it works. It's a predictive model based over an extended period of time. Bookies use to to figure out long term trends, which are then supplemented by situational information in order to calculate odds. In terms of individuals games there are too many uncontrollable, unreliable elements to a game of football for the data to be blindly applied to singular games as xG over a single game often won't be properly reflected in the scoreline.

As for the table, the positions don't matter, it's not supposed to be there to say who should definitively be in what position, it simply shows how the table would look in every game followed it's individual xG (which won't happen most games for a myriad of reasons). The data is what is important and it's actually pretty good for us in terms of consistency. Ideally you want you numbers to be as close to what they really are as possible as that indicates that you're neither under or over achieving and are simply maintaining a consistent level. It's easy to look at teams like Accrington and think "how are they above us when we're doing better than then" and think the model doesn't work but that not how it's supposed to work. The Accrington example simply shows that despite having a positive xG to xGA ratio is most games this season they are struggling to convert that into actual points, and in that scenario would probably point to them being pretty inefficient going forward, and a bit too generous at the back. We all know that we're a better side than Accrington, that's not the point.

In retrospect I probably should have just posted the graphs and skipped the table as it is quite confusing/makes us look worse than we probably are. Would be nice if people didn't instantly dismiss stuff as rubbish just because it doesn't paint us to be the best team in the history of football though.
 

SeaSeeEffCee

Well-Known Member
bought to you by the guy who instantly dismiss stuff as rubbish just because we conceed a goal ?
Thank you for your worthwhile contribution to this thread. It’s goes nicely with your other completely pointless one earlier too. See we could try and have an actual discussion, e.g. the intended function of an Internet forum, or you know you can continue to add absolutely nothing by making offhand remarks about something completely unrelated to try and score internet points, as per.
 

wingy

Well-Known Member
Pretty sure these were posted during last season , demonstrating where we were efficient or not .
 

wingy

Well-Known Member
Agree with CCFC here OP.
I know you have taken some stick here for being negative ,but don't think that is his intention
 

Mild-Mannered Janitor

Kindest Bloke on CCFC / Maker of CCFC Dreams
I think it shows why we are nervous, those last minute swings in games this year demonstrate to us how close every game is now going to be.
Let’s hopefully start with 3 points tonight to settle those nerves and adjust the stats and the bookies position
 

Sky_Blue_Ste

Active Member
Those that out perform the data, it eventually catches up with them & vice versa - obviously there are always going to be exceptions but on the whole it does make a difference.

We are out performing the chances we are making & it may catch up with us.
 

Mucca Mad Boys

Well-Known Member
Those that out perform the data, it eventually catches up with them & vice versa - obviously there are always going to be exceptions but on the whole it does make a difference.

We are out performing the chances we are making & it may catch up with us.

The stats are probably doesn’t represent our situation because we draw so many games. The teams around us have won 3-4+ more games than us, but have also lost a minimum of 3 games more than us also. Also, we failed to win an away game before Wycombe, but have gone on to only drop 2pts out of 12 since then.

With that in mind, you’d expect us to be quite neutral.

These stats have a use, but are absolutely not gospel.
 

Johnnythespider

Well-Known Member
Those that out perform the data, it eventually catches up with them & vice versa - obviously there are always going to be exceptions but on the whole it does make a difference.

We are out performing the chances we are making & it may catch up with us.
We are missing a lot of the chances we are making, we dominate possession in most games we play, we could be top but we're not, yet !!
 

Flying Fokker

Well-Known Member
I like all the numbers. I then look at the table now and see we are in with a bloody good chance of promotion. Seasons are made of dreams.
 

ccfcway

Well-Known Member
That's not how it works but I can see you're going to make absolute no attempt to understand.

Even tougher after tonight. It’s not a problem, the offer still stands. Not saying it’s wrong, just saying I will put my money where my mouth is if it comes in . Surely your ok with that ?
 

SeaSeeEffCee

Well-Known Member
Even tougher after tonight. It’s not a problem, the offer still stands. Not saying it’s wrong, just saying I will put my money where my mouth is if it comes in . Surely your ok with that ?
The issue is that you're completely missing the point. No-one is saying that Accrington are going to finish above us. The table isn't supposed to be taken as some gospel 'this is where everyone should be 100%'. But of course everyone sees us in 8th on it and freaks out and shouts about how the data is wrong and can't be trusted without actually understanding it.
 

Johhny Blue

Well-Known Member
The issue is that you're completely missing the point. No-one is saying that Accrington are going to finish above us. The table isn't supposed to be taken as some gospel 'this is where everyone should be 100%'. But of course everyone sees us in 8th on it and freaks out and shouts about how the data is wrong and can't be trusted without actually understanding it.
The little circles are pretty
 

Fergusons_Beard

Well-Known Member
You all need to wise up on xg data.

Brentford owners been using it for years cleaning up and selling data to rich gamblers.

Take last weekends Tranmere Portsmouth game. 0-2 to Pompey. Yet Tranmere had an xg of 1.5 And Pompey had an xg of 0.9.

That means that the chances Pompey created were low and probably lucky shots.

Tranmere on the other hand had good chances yet failed to score.

If that game is played 10,000 times With the same xg Tranmere would win more than Pompey.

Looking at their xg over their 9 game streak they’ve been pretty lucky-scoring way over their ‘expected’ outcome.

Don’t be ignorant- if interested I’d recommend The Expected Goals philosophy by James Tippett.






Sent from my iPhone using Tapatalk Pro
 

Users who are viewing this thread

Top