MCFC Analytics

[Damocles]

City have more goals from substitutes than any team in the league. In fact, we're 30% more than our closest rivals of Everton and around 50% ahead of Manchester United.

Proportionally however, 16.4% of our goals came from substitutes putting us fourth in the league behind Moyes' Everton on a massive 20% (1 in every 5 goals comes from a sub!), Lambert's Norwich with 17.4%, and Hughes' QPR with a close 17%. United had a measley 8% of their goals from substitutes.

The lowest was Fulham, who only managed a single goal all year from a sub, or 2.3% of their total goals.

The interesting question is what exactly this tells us about the teams. Do they have good squads? Are they more determined? Do they leave it late? Is it a boon for their managers? Is it a sign that their managers were found out? Ideally, I can cross reference this later with exactly what situation the teams were in when the sub goals were scored. Of course, some players might have gone off for fitness and injury reasons instead of a tactical change. More data is needed before analysis but that Everton stat is pretty shocking.
Most of the "big teams" of Chelsea, Liverpool, Spurs, United and Arsenal are floating in the 8-10% range rather than our 16.4%.

EDIT: By the way, if anybody has any questions that they would like investigating statistically, then I'd be happy to have a go for you.

 

[Stato]

City have more goals from substitutes than any team in the league. In fact, we're 30% more than our closest rivals of Everton and around 50% ahead of Manchester United.

Proportionally however, 16.4% of our goals came from substitutes putting us fourth in the league behind Moyes' Everton on a massive 20% (1 in every 5 goals comes from a sub!), Lambert's Norwich with 17.4%, and Hughes' QPR with a close 17%. United had a measley 8% of their goals from substitutes.

The lowest was Fulham, who only managed a single goal all year from a sub, or 2.3% of their total goals.

The interesting question is what exactly this tells us about the teams. Do they have good squads? Are they more determined? Do they leave it late? Is it a boon for their managers? Is it a sign that their managers were found out? Ideally, I can cross reference this later with exactly what situation the teams were in when the sub goals were scored. Of course, some players might have gone off for fitness and injury reasons instead of a tactical change. More data is needed before analysis but that Everton stat is pretty shocking.
Most of the "big teams" of Chelsea, Liverpool, Spurs, United and Arsenal are floating in the 8-10% range rather than our 16.4%.

Think ours is down to the way we grind teams down, so that by the time the subs are on there fresh and the opposition are knackered.

 

[Didsbury Dave]

Here's some data visualisation done by a Tweeter:

<a class="postlink" href="https://efex.bimeapp.com/players/dashboard/113A000D95C24E0185188E1B6E8E173B" onclick="window.open(this.href);return false;">https://efex.bimeapp.com/players/dashbo ... 1B6E8E173B</a>

Notice how absolutely terrible we are at long balls and how high Rodwell is in the tiki-taka ranking.

Checking through my own data collation, I thought I'd give Nasri's passing stats a look over for our good mate Didsbury Dave:

Total Successful Passes: 1539
Total Unsuccessful Passes: 266
Total Succ. Passes excl Crosses and Corners: 1498
Total Unsucc. Passes excl Crosses and Corners: 148

Succ Passes in defensive third: 87
Succ Passes in middle third: 762
Succ Passes in offensive third: 649

Succ Short Passes: 1454
Succ Crosses/Corners: 41

Corners Taken: 93
Of Which Are Short Corners: 25
Succ Corners into box: 21

Succ Long Balls: 0 (all season!)
Unsucc Long Balls: 9

Pass Forward:463
Pass Backwards: 207
Pass Left: 538
Pass Right: 497
(1035 side passes)

Key Passes: 69
Assists: 9

Total touches: 2152 - 5th behind Yaya, Barry, Silva and Clichy

the stats mean little in isolation, although you could make an argument about how few of the total passes are forwards.

Nasri is clearly an efficient passer. What he doesn't do is enough incisive passes for me. And he sometimes takes too many touches. If we stick with 352 he's not going to get that many games either.

 

[Bodicoteblue]

Did nobody hear Colin Schindler on 5live this morning ridiculing the use of stats & performance analysis? "one day football will come back to me" Dinosaur!

 

[Damocles]

Did nobody hear Colin Schindler on 5live this morning ridiculing the use of stats & performance analysis? "one day football will come back to me" Dinosaur!

I don't listen to fucking morons.

As DD accurately said, statistics on their own are useless as football is a game about the creation, defence and exploitation of space. We do not currently have the data to work here as the data is event driven where the events are things like "touching the ball". When events become as granular as "moving" and we can see the entire pitch on a second by second scale, then statistics will entirely run the game. Until that day, we have to rely on them as a supplement to an educated football brain.

And Colin Schilinder doesn't even know what performance analysis is, let alone enough to have a basis of an opinion on it. I sincerely hope that he drops dead, and I mean that with all seriousness. He is everything that is bad about this species. The day I meet him, I have promised myself that I will squarely punch him in the face.

 

[Damocles]

Who was really the MoM?

<a class="postlink" href="http://public.tableausoftware.com/views/Premiershiplastyear/Whathappenedlastyear?:embed=y" onclick="window.open(this.href);return false;">http://public.tableausoftware.com/views ... r?:embed=y</a>

Nice data visualisation there. This is gaining speed, you guys should keep up with it.

 

[roaminblue]

Got the lite version the other day. I'm a real beginner with statistics. But really want to learn more.

I'd be interested in breaking the data down and finding which sets of data mean we score more goals. Any input on which metrics I should be taking into account? Or at least find corellations if not causality.


One other idea I had was when thinking about rvp. I want to try and break down the service he received at arsenal and compare the service he could have received at us and utd and see whether he would have been a worth while signing for us, and whether he is for them. Purely for entertainment purposes, as I know that teams will adapt the way they play

Is there an easier way of finding the data and taking it out of the spreadsheet?<br /><br />-- Sun Aug 19, 2012 7:08 pm --<br /><br />

Who was really the MoM?

<a class="postlink" href="http://public.tableausoftware.com/views/Premiershiplastyear/Whathappenedlastyear?:embed=y" onclick="window.open(this.href);return false;">http://public.tableausoftware.com/views ... r?:embed=y</a>

Nice data visualisation there. This is gaining speed, you guys should keep up with it.

Where are you getting this info from by the way?

 

[Damocles]

Top question roaminblue.

I'm waiting back to hear from Gavin whether I can poist schemas but it shouldn't be a problem.

What you have now is a shitload of data. You need a way to query that data. I use SQL or Structured Query Language after putting it in a DB.

Your RVP question is difficult but if you give me a couple of days, I can answer.

 

[roaminblue]

Top question roaminblue.

I'm waiting back to hear from Gavin whether I can poist schemas but it shouldn't be a problem.

What you have now is a shitload of data. You need a way to query that data. I use SQL or Structured Query Language after putting it in a DB.


Your RVP question is difficult but if you give me a couple of days, I can answer.

That's brilliant, could you do me a favour? Would it be possible to give me a quick break down of the process you undertake? Nothing in depth but just procedure, I think I could read up on it.

With regards to the first bit (strikers and what makes them score more goals) I'm extracting all of the data manually, what sorted of db would you recommend for a beginner?

Do you mind if I drop you a pm tomorrow with the metrics that I am using as evidence? I'm really only doing this as a learning process tbh, but I am keen to learn more, and if I ca. Cut down how time consuming it is

 

[Damocles]

Anybody recieved the advanced package yet or doing anything interesting with the data?