Motherwell FC - A Thread For All Seasons

[Busta Nut]

Maybe I am just getting auld.

[ropy]

13 minutes ago, Busta Nut said:

Maybe I am just getting auld.

I wasn’t expecting that 

[Busta Nut]

3 minutes ago, ropy said:

I wasn’t expecting that 

It's early. Auld or not. f**k them.

[RandomGuy.]

3 hours ago, 18May1991 said:

Post of the year.

Its a load of contrived, made up shite by people that have too much time on their hands. 

It's generally just misunderstood by old men tbh. The name "expected goals" doesn't really help though.

Its really meant to be looked at long term to see trends and help predict things.

[Casagolda]

On 06/11/2021 at 18:14, Casagolda said:

worth noting that for all the abuse he gets, that's Goss now got 2 assists(both from open play) in his 5 starts this season.

That’s now 3 assists in 6 starts for Goss. Well 4 technically as I think his miscue for the 1st also went down as an assist..

What a difference from set pieces though when him and Slattery are over them, we looked like scoring from just about every delivery yesterday. Goss’s ball in for the 2nd in particular was absolutely top drawer.  

I was vehemently against his signing but I do remember Alexander saying he had ‘untapped potential’ and that he thought he could get more out him. Which to be fair he has done with the likes of Devante Cole, Tony Watt etc. He’s also getting something most thought wasn’t possible out of Lamie and Mugabi so fair play to Alexander in that regard. For all the stick his tactics take at times, he does seem to be a good man manager who’s able to get the most of out individuals. 

The midfield yesterday is by far the best it’s been all season. Energetic, aggressive, snide when required but also had the composure to get the ball down and pick a pass. 

Being on the front foot and pressing from the off helped as well. Shields was an absolute pest throughout and delighted to see him get his 1st goal. Also thought Woolery had his best game in a Motherwell strip as well.

Essentially it was the best performance of the season by far as we absolutely battered Hearts from start to finish. 

[ropy]

Presumably going forward our midfield is Goss, Slattery and A.N. Other.  Who is going to take that place?

[Kapowzer]

1 minute ago, ropy said:

Presumably going forward our midfield is Goss, Slattery and A.N. Other.  Who is going to take that place?

A January signing hopefully

[Swello]

1 hour ago, ropy said:

Presumably going forward our midfield is Goss, Slattery and A.N. Other.  Who is going to take that place?

I assume the Irish boy that we've signed.

[ZX1886]

The lack of short corners yesterday was pleasing.

[Jim McLean's Ghost]

4 hours ago, RandomGuy. said:

It's generally just misunderstood by old men tbh. The name "expected goals" doesn't really help though.

Its really meant to be looked at long term to see trends and help predict things.

This seems like bullshit to me. If every event is independent how can past performance predict the future.

It would be like going to a roulette table, seeing the last 10 numbers were black and deciding that red is more likely in the future. Which any introductary stats class will tell you is complete bollocks.

[RandomGuy.]

17 minutes ago, Jim McLean's Ghost said:

This seems like bullshit to me. If every event is independent how can past performance predict the future.

It would be like going to a roulette table, seeing the last 10 numbers were black and deciding that red is more likely in the future. Which any introductary stats class will tell you is complete bollocks.

It's been proven to give you a good insight into how things will end when used correctly in conjunction with other things.

It can help you see beyond things that can be misleading. Like we had 1 less shot on target than Celtic yesterday, but clearly had far less chance of scoring, over a long period it can show you if that's happening regularly. If it is then it'll likely lead to a poor league performance as few teams are clinical enough to overcome that regularly.

Yer Celtic fan there looks like a fanny, but over the past few seasons any team with as poor underlying figures as Motherwell have generally been caught out and slumped. We regularly had very good figures while starting slow as f**k but had surges after January that brought us back into line. There will be teams that defeat this system as that's football, but it is a system clubs have used for years to gauge their performances.

[Jim McLean's Ghost]

2 minutes ago, RandomGuy. said:

It's been proven to give you a good insight into how things will end when used correctly in conjunction with other things.

It can help you see beyond things that can be misleading. Like we had 1 less shot on target than Celtic yesterday, but clearly had far less chance of scoring, over a long period it can show you if that's happening regularly. If it is then it'll likely lead to a poor league performance as few teams are clinical enough to overcome that regularly.

Yer Celtic fan there looks like a fanny, but over the past few seasons any team with as poor underlying figures as Motherwell have generally been caught out and slumped. We regularly had very good figures while starting slow as f**k but had surges after January that brought us back into line. There will be teams that defeat this system as that's football, but it is a system clubs have used for years to gauge their performances.

Not arguing that expected goals isn't an accuraate indicator of the quality of chances created. That is the only thing it is.

 

[RandomGuy.]

2 minutes ago, Jim McLean's Ghost said:

Not arguing that expected goals isn't an accuraate indicator of the quality of chances created. That is the only thing it is.

 

Last season it correctly predicted the top 5, with only Celtic/Rangers in the wrong order. 

Only top 6 club it gone wrong was Livingston, where it preferred St Mirren. It can give you a good indicator into what to expect once you've worked out kinks in the modelling and have a big enough sample size. Just now mines, which includes xG but isn't solely that figure, is saying Motherwell are due a few bad results, and Hearts/County have a slight chance of seeing their results change, but everyone else could easily continue performing as they are. On paper its saying you should be down with us and St Mirren and in a battle for the top half rather than comfortably in it, but theres always a team who overperforms these figures (United last season were murdered by these same statto types), so I'm beginning to think the tight nature of every game up here allows for that.

I won't argue with you though, as I don't want to be one of these folk who try and ram it down folks throats constantly trying to claim its some magical force that football must obey as its quite clearly not. Ultimately its just a tool to be used or ignored, and theres no right/wrong decision to be made there for fans.

[Jim McLean's Ghost]

2 minutes ago, RandomGuy. said:

Last season it correctly predicted the top 5, with only Celtic/Rangers in the wrong order. 

Only top 6 club it gone wrong was Livingston, where it preferred St Mirren. It can give you a good indicator into what to expect once you've worked out kinks in the modelling and have a big enough sample size. Just now mines, which includes xG but isn't solely that figure, is saying Motherwell are due a few bad results, and Hearts/County have a slight chance of seeing their results change, but everyone else could easily continue performing as they are. On paper its saying you should be down with us and St Mirren and in a battle for the top half rather than comfortably in it, but theres always a team who overperforms these figures (United last season were murdered by these same statto types), so I'm beginning to think the tight nature of every game up here allows for that.

I won't argue with you though, as I don't want to be one of these folk who try and ram it down folks throats constantly trying to claim its some magical force that football must obey as its quite clearly not. Ultimately its just a tool to be used or ignored, and theres no right/wrong decision to be made there for fans.

Now did it predict the top 5? i.e. use past data to inform what will happen in the future

Or did the Xg data correlate with final table. A big difference. And it is hardly a triumph to be able to say that teams that have better chances than their opponents will finish higher up the table. How did it do at the bottom of the table?

[RandomGuy.]

7 minutes ago, Jim McLean's Ghost said:

Now did it predict the top 5? i.e. use past data to inform what will happen in the future

Or did the Xg data correlate with final table. A big difference. And it is hardly a triumph to be able to say that teams that have better chances than their opponents will finish higher up the table. How did it do at the bottom of the table?

From memory it was predicting the top 5 from around February onwards. It had us in the top 6 from September time anyway.

At the bottom I think it had United bottom, and Killie and Hamilton mixed up, but the rest was in order. Also sure it had been bunching teams up correctly from the turn of the year.

I don't see how you can agree that xG can match up with the final table then also claim you can't predict a final table using xG?

[Jim McLean's Ghost]

1 minute ago, RandomGuy. said:

From memory it was predicting the top 5 from around February onwards. It had us in the top 6 from September time anyway.

At the bottom I think it had United bottom, and Killie and Hamilton mixed up, but the rest was in order. Also sure it had been bunching teams up correctly from the turn of the year.

I don't see how you can agree that xG can match up with the final table then also claim you can't predict a final table using xG?

Because I don't think you understand what a prediction is. Me telling you last week's lottery numbers isn't much of a prediction. To predict something you need to use something you have to tell you about something that hasn't happened.

So Xg got the bottom 2 teams wrong. So even when it has all the data it doesn't correlate to reality.

[rowsdower]

Is there a good primer in how xG is calculated? I'm interested in how all the variables are accounted for, like shite goalies, time of the game, whether they're under pressure from a defender, etc. Is there a next generation of stats on the way?

I'm a big baseball fan, and love the stats there, but those make sense to me since there are less variables, and each team plays 160+ times.

[RandomGuy.]

8 minutes ago, Jim McLean's Ghost said:

Because I don't think you understand what a prediction is. Me telling you last week's lottery numbers isn't much of a prediction. To predict something you need to use something you have to tell you about something that hasn't happened.

So Xg got the bottom 2 teams wrong. So even when it has all the data it doesn't correlate to reality.

It predicted 5/6 of the top half and bottom half in February. How is that not it predicting a future result using past results?

It took the xG figures from the past ~20 games to predict how the table would look after another ~18?

It'll never, ever, predict every single position correctly and I don't think anyone's ever claimed it has? 

I understand you won't change your mind on this, but I'll tell you for a fact half the clubs in the division including your own have predictive systems based around it to gauge how things are going/will go, and clubs aren't shy in ignoring things if they think it's a waste of time.

[Jim McLean's Ghost]

6 minutes ago, RandomGuy. said:

It predicted 5/6 of the top half and bottom half in February. How is that not it predicting a future result using past results?

Because anyone could look at the table in February and tell you the final standings. It is a prediction just not a very impressive one.

[JamesP_81]

On the xG thing the only time I've found it useful is as a retrospective stat after a game when it gives a bit of insight into whether one of the teams overperformed relative to the chances that they had on the day , ie if a game finished 2-1 and the xG was say 1.2-1.4 then the winning team have produced a better quality of finishing and the losers have squandered chances . Perhaps it gives a bit more of an insightful review of the game rather than just having to rely on shots on and off target , but beyond that I wouldn't give it a second glance.  As for predicting future results I'm in the 'yer da' camp on this one and am sure there's many other things that could show some kind of trend if extrapolated over a large enough period and calculated in just the right way.  

More importantly tho, Saturday's game was f*king brilliant. Tony Watt's shitehousery in particular deserves special praise.