Match Thread: vs Portsmouth (a)

[Animal is a Ram]

13 hours ago, ram59 said:

If you look a bit closely at the photo, you'll see the lino is on that side of the pitch, very close to their player. Unfortunately, he seems to be looking across the back line, looking for offside and not looking at the player with the ball.

Well spotted! Clearly need my eyes testing as much as the ref  👓

[Foreveram]

9 minutes ago, ap04 said:

In other words that we have scored more and conceded fewer than would normally be due for our play (the opposite of last year). You call that skill, I call it sampling variability.

[Steve Buckley’s Dog]

9 minutes ago, ap04 said:

In other words that we have scored more and conceded fewer than would normally be due for our play (the opposite of last year). You call that skill, I call it sampling variability.

I remember when Phil Gee and Trevor Christie scored against Rotherham. My celebrations were tempered by the sampling variability of that season up to that point so it took the edge off somewhat. Same after Robbie Van Der Laan scored against Palace; it really ruined all the beers afterwards. 

Maybe, and this is just me, you need to lighten up and enjoy the ride of a promotion race otherwise perhaps football is not for you. 

[Caerphilly Ram]

43 minutes ago, ap04 said:

In other words that we have scored more and conceded fewer than would normally be due for our play (the opposite of last year). You call that skill, I call it sampling variability.

I call it what it is, we have scored more and conceded fewer than last year. It’s factual. So as I originally stated, we’ve done better this season against the teams around us. It’s led to more wins and more points. Which has us second in the league close to securing automatic promotion. At this stage I’m not even sure you know the point you’re painstakingly trying to prove.
 

“Sampling variability”…. I’m starting to wonder if Football isn’t for you friend 🤔

Yes, statistics have their place in analysing performance and trends no doubt, but without the “variability” that comes from someone smashing in a screamer from 20 yards, or pulling off a wonder save against all probability then what’s the point? It’s a sport, an entertainment, if all you care about is statistics try monitoring stocks and shares, or the footfall of your local town centre, or the weather or something.

[ap04]

18 hours ago, kevinhectoring said:

This seems completely wrong to me. Either you measure quality of chances (which is what xG does) or you look at how many goals the teams have actually scored. But if in the xG calculation you give 1.0 for a goal scored, then you’re mixing the two up

I am not mixing anything up, just saying that this is one instance where xG is not a reliable performance indicator, as in a goal being the maximum in performance scales (perfect attacking/worst defending).

Also not sure why you think it's either goals or chances (for evaluating not xG), ideally you want an analogue 0 (no threat) to 1 (goal), including execution and all non-shot events. In fact some later models use post-shot ball tracking or even goals as a 'boost'.

[Archied]

Trying to remember when / if football was a sport where you always got what you think you deserved , never mind by what metric you choose to decide what is deserved and what isn’t ,

Say you win a match one nil with a wonder goal and your keeper makes 3 fantastic saves , why don’t you deserve to win ?

you play a game where you are happy to let the opposition have the ball more but in non threatening areas , you defend well and let them do bugger all with the ball then attack with purpose when you have the ball and score more goals ,,,, you don’t deserve to win?

frankly for me there is also an enjoyment in seeing a team set up so well that they work as a unit and hardly need to tackle with the opposition keep turning back with nowhere to go ,, let’s be honest we admire defenders with that type of positional and game reading skill, it’s a gift every bit as much as a natural finisher ,

football is a game with many differing skills and tactics 🤷🏻‍♂️

[Crewton]

"Deserved to win" is one of the biggest cliches in football. Statistics do  at least help us bring some logic to the analysis of performance v results, but the League table is the only thing that really matters.

[kevinhectoring]

1 hour ago, ap04 said:

I am not mixing anything up, just saying that this is one instance where xG is not a reliable performance indicator, as in a goal being the maximum in performance scales (perfect attacking/worst defending).

You say xG should award 1.0 for a goal? You miss the point that goals involve varying degrees of good fortune.  
 

If you award 1.0 to Ward’s second goal, on a xG assessment, you overvalue the chance massively. It was neither perfect attacking nor the worst defending. It was a goal that was scored through a generous dollop of good fortune. (Fact is, he was too far out to sensibly shoot and the defender was pretty well-positioned, as was the keeper ) 

There are two quite separate things : the quality of the scoring opportunity and the fact a goal was scored. Apples and oranges. I think you’re mixing them up 

 

Edited April 4 by kevinhectoring

[FlyBritishMidland]

1 hour ago, ap04 said:

I am not mixing anything up, just saying that this is one instance where xG is not a reliable performance indicator, as in a goal being the maximum in performance scales (perfect attacking/worst defending).

Also not sure why you think it's either goals or chances (for evaluating not xG), ideally you want an analogue 0 (no threat) to 1 (goal), including execution and all non-shot events. In fact some later models use post-shot ball tracking or even goals as a 'boost'.

What you’re saying is you’ll choose to use whatever metrics you want to create and make up to suit your narrative.

I’m sorry but I’m really struggling to understand how any fan can go to such lengths to try and justify why their team shouldn’t win/draw/get a positive outcome from a match.

[Reggie Greenwood]

28 minutes ago, FlyBritishMidland said:

What you’re saying is you’ll choose to use whatever metrics you want to create and make up to suit your narrative.

I’m sorry but I’m really struggling to understand how any fan can go to such lengths to try and justify why their team shouldn’t win/draw/get a positive outcome from a match.

Not a fan a WUM 

[norwichram]

I’m very proud of our club, team and fans for the part they all played in a really good advert for football. Great professionalism and respect for the opponent and both sides and fans were committed. 

DTID 
 

 

[ap04]

30 minutes ago, FlyBritishMidland said:

What you’re saying is you’ll choose to use whatever metrics you want to create and make up to suit your narrative.

No, what I'm saying is that I can use the metric I want for Derby but have no time or interest to watch 11 other games a week so I have to go by the next best (models).

 

30 minutes ago, FlyBritishMidland said:

I’m sorry but I’m really struggling to understand how any fan can go to such lengths to try and justify why their team shouldn’t win/draw/get a positive outcome from a match.

It works both ways though, remember at Northampton everyone saying how we were a disgrace and deserved nothing from the game, guess what that was nonsense too.

[Premier ram]

2 hours ago, ap04 said:

I am not mixing anything up, just saying that this is one instance where xG is not a reliable performance indicator, as in a goal being the maximum in performance scales (perfect attacking/worst defending).

Also not sure why you think it's either goals or chances (for evaluating not xG), ideally you want an analogue 0 (no threat) to 1 (goal), including execution and all non-shot events. In fact some later models use post-shot ball tracking or even goals as a 'boost'.

 

                        

[Wolfie20]

Not again!!!

 

 

[Srg]

31 minutes ago, ap04 said:

No, what I'm saying is that I can use the metric I want for Derby but have no time or interest to watch 11 other games a week so I have to go by the next best (models).

You're not though.

You use xG to prove when we are crap, and when it doesn't, you say it isn't good and invoke 'sampling variance' instead.

You're boring.

[Ghost of Clough]

3 hours ago, Caerphilly Ram said:

I call it what it is, we have scored more and conceded fewer than last year. It’s factual. So as I originally stated, we’ve done better this season against the teams around us. It’s led to more wins and more points. Which has us second in the league close to securing automatic promotion. At this stage I’m not even sure you know the point you’re painstakingly trying to prove.
 

“Sampling variability”…. I’m starting to wonder if Football isn’t for you friend 🤔

Yes, statistics have their place in analysing performance and trends no doubt, but without the “variability” that comes from someone smashing in a screamer from 20 yards, or pulling off a wonder save against all probability then what’s the point? It’s a sport, an entertainment, if all you care about is statistics try monitoring stocks and shares, or the footfall of your local town centre, or the weather or something.

Anyone thinking a goal should count as an xG of 1.0, "becuase it was a goal" doesn't havea firm grasp of statistics. 

Imagine flipping a coin. It has an xH (expected heads) of 0.5 and xT (expected tails) of 0.5. This particular poster thinks after flipping the coin and it landing on heads, the probability should be counted as landing on heads 100% of the time.

The rest of what the poster says isn't much better than that

[Foreveram]

6 minutes ago, Ghost of Clough said:

Anyone thinking a goal should count as an xG of 1.0, "becuase it was a goal" doesn't havea firm grasp of statistics. 

Imagine flipping a coin. It has an xH (expected heads) of 0.5 and xT (expected tails) of 0.5. This particular poster thinks after flipping the coin and it landing on heads, the probability should be counted as landing on heads 100% of the time.

The rest of what the poster says isn't much better than that

Image in the future trying to explain this and the kids saying, what’s a coin 🙄

[G STAR RAM]

6 minutes ago, Foreveram said:

Image in the future trying to explain this and the kids saying, what’s a coin 🙄

Surely it will be an xC by then?

[Chester40]

14 minutes ago, Ghost of Clough said:

Anyone thinking a goal should count as an xG of 1.0, "becuase it was a goal" doesn't havea firm grasp of statistics. 

Imagine flipping a coin. It has an xH (expected heads) of 0.5 and xT (expected tails) of 0.5. This particular poster thinks after flipping the coin and it landing on heads, the probability should be counted as landing on heads 100% of the time.

The rest of what the poster says isn't much better than that

Well explained. 

I'm guessing that stats don't really give any context..

So if you score 2 early goals -  from a goal keeping error from a corner and a deflected shot from distance,  and you spend the rest of the game defending behind the ball,  launching it out, running down the clock, letting the keeper make 5 or 6 routine-ish saves.

No fan would think you had been lucky ...but the stats would suggest you were?

[DRBee]

32 minutes ago, ap04 said:

No, what I'm saying is that I can use the metric I want for Derby but have no time or interest to watch 11 other games a week so I have to go by the next best (models).

 

It works both ways though, remember at Northampton everyone saying how we were a disgrace and deserved nothing from the game, guess what that was nonsense too.

No you are just wasting everyone's time  and misapplying what can be deduced from the Xg stat of single instances.

An earlier poster said of Tuesday night  " it was a proper football match " , something most of us would agree with ( though probably not you). No stats you are likely to come up with can offer a descriptor of a 'proper good football match' ,  and in any case why do we need such stats given we are supporters not coaches and the stats don't add anything useful.

Perhaps we need a separate "stats" section of the forum where folk can swap statistics and talk the same language without polluting the match threads. 

When you write  "In other words that we have scored more and conceded fewer than would normally be due for our play (the opposite of last year). You call that skill, I call it sampling variability"  I don't think this means what you think at all. A goal scored from a low-rated goal scoring chance is a real event. Analyses show that some teams show great discrepancy in  their  goals scored compared to the rating of their goal scoring chances and it can happen over lengthy periods of time. This is not down to "sampling variability" . It is the system not conforming to the model, not due to chance, but due to imperfections in the model being used.