A season in stats XG and XA debate

[wunderwell 75]

https://footystats.org/scotland/premiership/xg

9th best Xg in the league and the worst Xa in the league.

But 4th and aiming for third........

Does this mean that the constant use of these stats shows how flawed they are?

[wunderwell 75]

Who Has The Worst xGA?

Fans of Motherwell FC will demand better defending in the future, with 1.86 xGA (Goals Expected Against). The worst of all teams in the Premiership this season.

[ropy 755]

I thought the xGA was a yellow card for berating the ref

[grumpy 147]

They are predicting we should score 1.06 goals per game, after 20 matches that should be 21, we have scored 25. They are predicting 1.86 goals against, after 20 games that should be 37 goals, we have conceded 27. That suggests to me that the statisticians they have employed aren't very good.

Or to put it another way, there are lies, damned lies and statistics, this site seems like proof of that.

[numpty 450]

It's almost as if the main appeal of watching sport is its unpredictability.

[Stuwell2 66]

Getting scudded 6-1 in October probably skews the stats a bit. 

[weeyin 4,130]

I took a few statistics classes at Uni, but I struggle to understand some of the meaning of some of these. Partly because there isn't much public detail available about exactly how they are derived - you need to pay for that - but mostly because they don't make much sense at footballing level.

I can see how it makes a lot more sense in a game like baseball where a lot of the ideas were developed. A guy standing pitching a ball and another guy standing and hitting it with a bat is a lot easier to analyse than 11 players playing another 11 players who are allowed to run around and defend and tackle etc.

Not to say they are all useless, but I think a lot of them are questionable, at least without a lot of context that isn't shared to the non-paying public.

[fizoxy 600]

People expect us to be shite all the time, so I'm not surprised by the stats. I think that our points and league position are a bit better than our performances have deserved, but that is exactly why sport in general is so popular, because of the level of unpredictability.

The analytics can be useful for looking at past performance, and possibly catching red flags with potential injuries, etc, but there are so many factors that come into goals and assists, that predicting future results is not reliable.

Stats are good in baseball because they have 162 games per team, there isn't a whole.lot of movement, and it is easier to isolate individual performances. The sheer number of games can help even out issues like shite umpires, good fielding plays, game situations, etc. But in football it is easier for a couple of weeks of bad form, luck, or bad refereeing to define a season.

[Winning by Name 114]

On 12/31/2021 at 1:44 PM, Stuwell2 said:

Getting scudded 6-1 in October probably skews the stats a bit. 

Totally!

It takes our xga home from 1.86 to 1.34 if you omit that game. Just shows how much crap statistics can be.

[steelboy 1,265]

Is there any way to compare season 20-21 under Robinson with 20-21 under Alexander using those stats?

Probably best to forget 20-21 under Keith. 

 

[weeyin 4,130]

I haven't looked for a while, but last time I did a search most of the good stuff that could be used for comparison is only available to paid subscriptions - which aren't cheap.

There might some basic stuff available from the betting sites.

 

[Busta Nut 876]

I will always stand by this being a lot of shite. 
I love a stat too. These aren't really stats. 

[CoF 745]

On 12/31/2021 at 2:24 PM, weeyin said:

I took a few statistics classes at Uni, but I struggle to understand some of the meaning of some of these. Partly because there isn't much public detail available about exactly how they are derived - you need to pay for that - but mostly because they don't make much sense at footballing level.

I can see how it makes a lot more sense in a game like baseball where a lot of the ideas were developed. A guy standing pitching a ball and another guy standing and hitting it with a bat is a lot easier to analyse than 11 players playing another 11 players who are allowed to run around and defend and tackle etc.

Not to say they are all useless, but I think a lot of them are questionable, at least without a lot of context that isn't shared to the non-paying public.

Ive read a few articles on it to try and get a better understanding, but it still all means nothing to me....

 

On 12/30/2021 at 8:54 PM, wunderwell said:

Who Has The Worst xGA?

Fans of Motherwell FC will demand better defending in the future, with 1.86 xGA (Goals Expected Against). The worst of all teams in the Premiership this season.

....and heres the reason why. If we keep doing what we're doing and finish 4th then I'm absolutely fine with our 1.86xGA.... whatever the fuck it means. 

[wunderwell 75]

Not looked at the split - but again our position in the table was better than suggested - to me it simply seems we allow more shots at goal, but remembering this game from this season I think skews everything

https://footystats.org/scotland/aberdeen-fc-vs-motherwell-fc-h2h-stats#1275390

Aberdeen's shots that day were not particularly troublesome.

 

[grumpy 147]

1 hour ago, wunderwell said:

Not looked at the split - but again our position in the table was better than suggested - to me it simply seems we allow more shots at goal, but remembering this game from this season I think skews everything

https://footystats.org/scotland/aberdeen-fc-vs-motherwell-fc-h2h-stats#1275390

Aberdeen's shots that day were not particularly troublesome.

 

Having looked at that link, it would suggest that they expect a goal per 8(ish) shots at goal, whether it  is a two foot tap in or a forty yard effort that clears the stand. 

Looks like their statistical analysis is even more absurd than I thought at first reading.

[wunderwell 75]

2 minutes ago, grumpy said:

Having looked at that link, it would suggest that they expect a goal per 8(ish) shots at goal, whether it  is a two foot tap in or a forty yard effort that clears the stand. 

Looks like their statistical analysis is even more absurd than I thought at first reading.

I agree but the fact that Sky Sports are quoting on most of their punditry means we will hearing a lot more of it!

[grumpy 147]

Just now, wunderwell said:

I agree but the fact that Sky Sports are quoting on most of their punditry means we will hearing a lot more of it!

I won't, I wouldn't give $ky any of my hard earned cash, all they do is squander it by giving it to the EPL. 

[wunderwell 75]

Dodgy stick time!

[Spiderpig 1,093]

All those stats are just made up pish, no doubt keeping  a team of geeks in employment making it up as they go, to give other  geeks something to discuss. The only stats that matter are games played, won, drawn , lost and total points. 

[weeyin 4,130]

While, of course, total points is the ultimate stat we focus on, there is no doubt other details can be useful. Most of the huge clubs track and crunch the data to their benefit. I saw one of the Liverpool data scientists on the telly talking about some of the things they review. One thing they looked at was how likely it was for a goal to be scored from play at every position on the pitch. As well as their own stats, they used the data from every other EPL team over the past few years.

Based on that, they identified the optimal areas of the pitch they wanted to get into every time they had possession and then, in training, worked on tactics to take advantage of that.

Then they went and won the Champions League and the EPL title, so their might be something to it. However, as I mentioned earlier, none of that data is freely available. Instead, we get glimpses of individual stats that don't really mean much in isolation.

[Ya Bezzer! 1,779]

On 12/30/2021 at 9:59 PM, grumpy said:

They are predicting we should score 1.06 goals per game, after 20 matches that should be 21, we have scored 25. They are predicting 1.86 goals against, after 20 games that should be 37 goals, we have conceded 27. That suggests to me that the statisticians they have employed aren't very good.

Or to put it another way, there are lies, damned lies and statistics, this site seems like proof of that.

I predict we'll never score 1.06 goals in a game.

 

[Ya Bezzer! 1,779]

I like stats (obviously) but I prefer them to either be form based or actions that have a real world effect on a match (key passes, aerial duels, tackles etc.)

Even so I think stats only give you clues here and there to how you are performing.  I still prefer to actually watch the game.

I think it's also worth pointing out that there is a great plethora of football stats sites out there on the web but very few of them have reliable data for things like appearances and goals never mind headers won in the box from crosses from the right wing against opponents with blonde hair.

In fact there is a major website out there that puts out vast amounts of Scottish football stats.....yet they got the the appearances and goals wrong for literally half the players in the league last season...

[ropy 755]

2 hours ago, Ya Bezzer! said:

I like stats (obviously) but I prefer them to either be form based or actions that have a real world effect on a match (key passes, aerial duels, tackles etc.)

Even so I think stats only give you clues here and there to how you are performing.  I still prefer to actually watch the game.

I think it's also worth pointing out that there is a great plethora of football stats sites out there on the web but very few of them have reliable data for things like appearances and goals never mind headers won in the box from crosses from the right wing against opponents with blonde hair.

In fact there is a major website out there that puts out vast amounts of Scottish football stats.....yet they got the the appearances and goals wrong for literally half the players in the league last season...

I don’t even know how they work out the possession stats.  You often see 50/50

[weeyin 4,130]

4 minutes ago, ropy said:

I don’t even know how they work out the possession stats.  You often see 50/50

I read somewhere that the possession stats that are often used actually count the number of passes rather than minutes/seconds in possession.  The more passes you make, the more "possession" is allocated to you.

If you see 50/50 though, I'd guess that just means they haven't kept track!