Author Topic: Statistics and Analytics - insight into our performance (Read 194106 times)

PhaseOfPlay · « **Reply #960 on:** January 2, 2018, 09:44:13 pm »

Quote from: Shaved Crossbar on January 2, 2018, 09:38:20 pm

I’m not looking for a holistic outlook on penetration - as you said we can just reduce that to shots and goals. The other stats you mentioned are more general attackers stats, with the exception of through balls and successful dribbles (both of which are measured a bit weirdly).

But they are also indicators of what Packing is - passes, dribbles or shots that beat defenders. The innovation of "Packing" is applying a score to each penetration action by counting the number of defenders the action takes out of the game. The beauty of it, is that it actually gives a good indication of how well an attacker does against parked buses as they will, by definition, have more defenders to beat than an aggressive high pressure team.

Quote

Only a long ball team that produced a lot of successful long balls into the right areas - which isn’t a bad strategy at all to be honest, it’s just hard to do.

You might know more than me about it, but does the original Packing stat take into account whether the action is completed?

BrandoLFC · « **Reply #961 on:** January 2, 2018, 09:50:17 pm »

You could always DM him on twitter and ask him, probably that what he was developing is actually worth something and therefore not to be in the open. I don't think I've ever seen him delete anything else. The link here has a chart to something similar on Neymar replacements for Barca:

https://fivethirtyeight.com/features/the-neymar-deal-was-certainly-crazy-but-was-it-smart/

BrandoLFC · « **Reply #962 on:** January 2, 2018, 09:52:49 pm »

Packing – A metric which tries to determine the value of a pass or dribble by evaluating how many defenders it “takes out of the game”. Basically, how many defenders are in between the ball and the goal when a player receives the ball, and then how many still are when the next player receives it. Packing is very intuitive, and it’s something most viewers can notice while watching. The most well known company producing this data can be found here if you would like to read more about it. (also just google it, there are plenty of helpful articles)

Keita has absolutely incredible packing numbers. Even if most of his passes may look normal, he seems to always find a way to create a slight imbalance between offense and defense. He always picks the most impactful pass to give the attack whatever advantage can be gained based on current defender positioning. If you want to see an example, watch some of his full game highlights and take a look at how many times he is able to breakthrough the midfield “barrier” and push the ball into the space behind midfielders and in front of defenders. If you have watched Liverpool play under Klopp, you know how frustrating it is to watch a team have possession but be unable to exploit those gaps effectively to put pressure on a set defense.

Also as a side note, Sakho has always graded very high in this metric as well. Usually one of the only defenders near the top. It is why everyone on this sub was able to recognize that although his passing is clumsy, it is effective.

This is also from the Impect website that is referenced above:

Soccer is all about scoring goals and the less opponents a player is facing in front of the goal, the more likely he is to score. Consequently, the attacking team needs to get past the defending team and the aim is thus to outplay the opponents. To achieve this, it’s essential that after a move less defending players will be between the ball and the goal. The outplayed opponents are then ‘packed’ and currently not able to prevent the ongoing attack. This is why the idea of taking players out of the game is called Packing.

PhaseOfPlay · « **Reply #963 on:** January 2, 2018, 09:58:08 pm »

So, basically - Penetration

You can tell this stat was developed by ex-players

BrandoLFC · « **Reply #964 on:** January 2, 2018, 10:07:07 pm »

Ex-player or not it's exciting to see these type of things develop and a bit of a shame there isn't more for public consumption. Don't we have something similar to SportsVu cameras installed at Anfield?

Shaved Crossbar · « **Reply #965 on:** January 2, 2018, 10:31:28 pm »

Quote from: PhaseOfPlay on January 2, 2018, 09:44:13 pm

But they are also indicators of what Packing is - passes, dribbles or shots that beat defenders. The innovation of "Packing" is applying a score to each penetration action by counting the number of defenders the action takes out of the game. The beauty of it, is that it actually gives a good indication of how well an attacker does against parked buses as they will, by definition, have more defenders to beat than an aggressive high pressure team.

You might know more than me about it, but does the original Packing stat take into account whether the action is completed?

Yeah, the penultimate sentence of your post is what I’m looking for.

Caley’s model did take only completed actions into account (but not, im guessing, good passes at the end of a dribble, as Adama Traore scored very highly).

Shaved Crossbar · « **Reply #966 on:** January 2, 2018, 10:35:41 pm »

Quote from: BrandoLFC on January 2, 2018, 09:50:17 pm

You could always DM him on twitter and ask him, probably that what he was developing is actually worth something and therefore not to be in the open. I don't think I've ever seen him delete anything else. The link here has a chart to something similar on Neymar replacements for Barca:

https://fivethirtyeight.com/features/the-neymar-deal-was-certainly-crazy-but-was-it-smart/

Yeah good shout, worth a try.

Geormajesty · « **Reply #967 on:** January 2, 2018, 10:40:29 pm »

Quote from: BrandoLFC on January 2, 2018, 09:50:17 pm

You could always DM him on twitter and ask him, probably that what he was developing is actually worth something and therefore not to be in the open. I don't think I've ever seen him delete anything else. The link here has a chart to something similar on Neymar replacements for Barca:

https://fivethirtyeight.com/features/the-neymar-deal-was-certainly-crazy-but-was-it-smart/

I think he deleted a lot of his tweets and started again, rather than specifically deleting those - his overall tweet number is only 1500 now.

By the way, the next five names on the list from the Neymar article were - Felipe Anderson, Dimitri Payet, Lucas Moura, Sadio Mane and Florian Thauvin.

Geormajesty · « **Reply #968 on:** January 2, 2018, 10:42:35 pm »

PoP - while we're talking about Caley, you might be interested in his xG model as it features player finishing skill in the algorithm - https://cartilagefreecaptain.sbnation.com/2015/10/19/9295905/premier-league-projections-and-new-expected-goals

PhaseOfPlay · « **Reply #969 on:** January 2, 2018, 10:44:50 pm »

Quote from: Geormajesty on January 2, 2018, 10:42:35 pm

PoP - while we're talking about Caley, you might be interested in his xG model as it features player finishing skill in the algorithm - https://cartilagefreecaptain.sbnation.com/2015/10/19/9295905/premier-league-projections-and-new-expected-goals

Yeah I like Caley. Even though he's a Spurs fan

BrandoLFC · « **Reply #970 on:** January 2, 2018, 11:21:55 pm »

Quote from: Geormajesty on January 2, 2018, 10:40:29 pm

I think he deleted a lot of his tweets and started again, rather than specifically deleting those - his overall tweet number is only 1500 now.

By the way, the next five names on the list from the Neymar article were - Felipe Anderson, Dimitri Payet, Lucas Moura, Sadio Mane and Florian Thauvin.

No, he moved a lot of them from his normal account to his graphics account. He only posts his stats vision stuff on his graphics account now.

BabuYagu · « **Reply #971 on:** January 3, 2018, 10:53:16 am »

Quote from: PhaseOfPlay on January 2, 2018, 09:44:13 pm

But they are also indicators of what Packing is - passes, dribbles or shots that beat defenders. The innovation of "Packing" is applying a score to each penetration action by counting the number of defenders the action takes out of the game. The beauty of it, is that it actually gives a good indication of how well an attacker does against parked buses as they will, by definition, have more defenders to beat than an aggressive high pressure team.

You might know more than me about it, but does the original Packing stat take into account whether the action is completed?

Packing score is also attributed to movement. If Moreno makes a pass down the left to Coutinho, then overlaps to receive the return ball, running past 3 defenders in the process, then crossing - statistically that will show up as a completed pass and cross - which may or may not even result in a shot. In terms of packing though it would have a score of 3 for both him and Coutinho.

JCB · « **Reply #972 on:** January 15, 2018, 07:09:22 pm »

So I decided to update the EPL xG Trend data I ran a few months ago:

Here's our closest competitors' charts

Alternatively go direct to the source and see any EPL team for yourself:

https://public.tableau.com/profile/jcbviz#!/vizhome/xGTrendAnalysis/xGTrendAnalysisDashboard?publish=yes
https://public.tableau.com/profile/jcbviz#!/vizhome/xGTrendComparision/xGComparision?publish=yes

groove · « **Reply #973 on:** January 15, 2018, 07:19:07 pm »

1. Wasn't sure whether to put Coutinho in red or not but fuck it - those are Liverpool's goals!
2. Not one Real Madrid player in the top 25. Probably the first time that's happened in the last decade. Ronaldo definitely fading.
3. Lukaku and Morata both missing out. £70m a piece well spent so far.
4. Players who you think score more than they actually do: Hazard & Sanchez.
5. Kane is exceptional - let's just end that here. I also think he's one of the closest forwards to Firmino in the league when it comes to his link up play and work rate outside the box.
6. Fair play Raheem, but thanks for not turning up yesterday.
7. Why was the Premier League so hard for Falcao?
8. Aspas sneaking in there.
9. I think Carragher said yesterday that Mane wasn't having a good a season goal-wise as last season but that's not true when you look at it like this. He's actually having a slightly better one (158 vs 172).
10. Firmino ahead of world class "proper" number 9s like Aguero, Suarez and Lewandowski. What a player.

Medellin · « **Reply #974 on:** January 18, 2018, 07:50:22 pm »

Fill yer boots.

https://en.as.com/en/2018/01/17/football/1516210881_006124.html

WisconsinRed · « **Reply #975 on:** January 18, 2018, 08:02:48 pm »

We are absolutely brilliant no matter how you evaluate us

PoetryInMotion · « **Reply #976 on:** January 18, 2018, 08:07:01 pm »

Quote from: groove on January 15, 2018, 07:19:07 pm

1. Wasn't sure whether to put Coutinho in red or not but fuck it - those are Liverpool's goals!
2. Not one Real Madrid player in the top 25. Probably the first time that's happened in the last decade. Ronaldo definitely fading.
3. Lukaku and Morata both missing out. £70m a piece well spent so far.
4. Players who you think score more than they actually do: Hazard & Sanchez.
5. Kane is exceptional - let's just end that here. I also think he's one of the closest forwards to Firmino in the league when it comes to his link up play and work rate outside the box.
6. Fair play Raheem, but thanks for not turning up yesterday.
7. Why was the Premier League so hard for Falcao?
8. Aspas sneaking in there.
9. I think Carragher said yesterday that Mane wasn't having a good a season goal-wise as last season but that's not true when you look at it like this. He's actually having a slightly better one (158 vs 172).
10. Firmino ahead of world class "proper" number 9s like Aguero, Suarez and Lewandowski. What a player.

1. Fair enough.
2. Real Madrid have been struggling this season and quite often upfront as well, so not a surprise.
3. You bet!
4. For some reason, they're considered Goal Scoring Wide players but don't often hit big numbers. Hazard maybe a bit deeper than Sanchez, but the point holds.
5. Agreed.
6. He hardly turns up against us, does he? He's turned it up a notch this season though.
7. He was coming at the back of long term injuries and it has taken him quite a while to get back to the game. It also mustn't have been a pleasure playing two boring defensive managers back to back.
8. He has definitely done well before and after his stint with us. Luis Alberto doing well for Lazio and Suso for Neymar. Somehow these Spaniards never fit in with us like the others before.
9. Mane is scoring at a good rate. His overall game had a dip in between after the suspension and injury but now he's getting back to form.
10. Think Aguero has been rotated with Jesus earlier this season and Suarez had a bit of a lull before he has picked up of late. Lewa hasn't been to his standards. Cavani's the one who's still scoring at a big rate. But Firmino is still phenomenal, considering his overall game.

drmick · « **Reply #977 on:** January 19, 2018, 12:09:00 am »

Quote from: Medellin on January 18, 2018, 07:50:22 pm

Fill yer boots.

https://en.as.com/en/2018/01/17/football/1516210881_006124.html

Thats a bit worrying for LFC.

chemeddy · « **Reply #978 on:** January 19, 2018, 03:23:42 am »

The entire report from UEFA is quite interesting actually.

http://www.uefa.com/MultimediaFiles/Download/OfficialDocument/uefaorg/Clublicensing/02/53/00/22/2530022_DOWNLOAD.pdf

Financial figures are taken from FY2016. I'm not sure if LFC's figures are from Jan - Dec 2016 or otherwise. In any case, the figures are "outdated" to a certain extent.

I do hope our wage bill as a percentage of our revenues will drop in the FY2017 though. 69% does sound very high and unhealthy.

JCB · « **Reply #979 on:** January 19, 2018, 08:41:05 am »

Quote from: chemeddy on January 19, 2018, 03:23:42 am

The entire report from UEFA is quite interesting actually.

http://www.uefa.com/MultimediaFiles/Download/OfficialDocument/uefaorg/Clublicensing/02/53/00/22/2530022_DOWNLOAD.pdf

Financial figures are taken from FY2016. I'm not sure if LFC's figures are from Jan - Dec 2016 or otherwise. In any case, the figures are "outdated" to a certain extent.

I do hope our wage bill as a percentage of our revenues will drop in the FY2017 though. 69% does sound very high and unhealthy.

In the previous year's report it was 56% of revenue. As listed in the footnote this was a 30% hike, presumably in "anticipation of the new Premier League tv increases".

We also just finished (or were in the middle of) redeveloping the main stand (p117) thus anticipating further increases from that. The next report should see that drop back to acceptable levels.

clinical · « **Reply #980 on:** January 19, 2018, 09:11:33 am »

I know it's taken from SSN

but maybe Mignolet isn't quite as bad as we think? Still cost us, but also De Gea is something else.

http://www.skysports.com/football/news/11661/11212584/david-de-gea-premier-leagues-top-performing-goalkeeper-stats-reveal

rocco · « **Reply #981 on:** January 20, 2018, 09:44:32 am »

Quote from: JCB on January 19, 2018, 08:41:05 am

In the previous year's report it was 56% of revenue. As listed in the footnote this was a 30% hike, presumably in "anticipation of the new Premier League tv increases".

We also just finished (or were in the middle of) redeveloping the main stand (p117) thus anticipating further increases from that. The next report should see that drop back to acceptable levels.

I take it the 30% hike was across all clubs as certainly not just us , as 56% is a healthy enough figure

has gone odd · « **Reply #982 on:** January 20, 2018, 11:48:37 am »

with klopp now stating karius is no1 i decided to look at his prem league stats.

to be fair, in the 14 games in the PL he has started, we have lost just the once but he has conceeded 16 goals. not sure what to make of that ...

vagabond · « **Reply #983 on:** January 20, 2018, 11:52:39 am »

Quote from: Medellin on January 18, 2018, 07:50:22 pm

Fill yer boots.

https://en.as.com/en/2018/01/17/football/1516210881_006124.html

Amazing that Spurs can get away with that wage bill. Also no excuses for not coming top 4 if you have one of the top 4 wage bills, as we do.

JCB · « **Reply #984 on:** January 28, 2018, 12:57:06 am »

So to take my mind off of today's disappointing game I put this together, something I've been meaning to do for a while.

Please note that the data is for EPL games only and is up to date.
The size of the circle denotes time played.
Positions are those listed by opta, I have not bothered amending them.

groove · « **Reply #985 on:** January 28, 2018, 01:51:12 pm »

Good stuff that, thanks! Wonder if Ayew would be a decent squad signing...

JCB · « **Reply #986 on:** January 28, 2018, 02:31:34 pm »

Quote from: groove on January 28, 2018, 01:51:12 pm

Good stuff that, thanks! Wonder if Ayew would be a decent squad signing...

Richarlison also stood out as a possible option. I was also interested to see that Mahrez posted similar numbers to D. Silva this season.

JCB · « **Reply #987 on:** February 9, 2018, 06:14:56 pm »

Bob Sacamano · « **Reply #988 on:** February 13, 2018, 04:07:41 pm »

538 gives us a 95% chance to make the Champions League which seems...high to me. They give us a 6% chance to win the Champions League this year, so I assume it's 89% to make top 4. Given that we are only 1 point ahead of Chelsea and 2 ahead of Spurs that really seems high to me...I would've guessed closer to 70%...but apparently their model sees us as the 2nd best team in the premier league.

BassTunedToRed · « **Reply #989 on:** February 13, 2018, 04:27:59 pm »

Quote from: Bob Sacamano on February 13, 2018, 04:07:41 pm

538 gives us a 95% chance to make the Champions League which seems...high to me. They give us a 6% chance to win the Champions League this year, so I assume it's 89% to make top 4. Given that we are only 1 point ahead of Chelsea and 2 ahead of Spurs that really seems high to me...I would've guessed closer to 70%...but apparently their model sees us as the 2nd best team in the premier league.

This has us at 83.7% for a top four finish (albeit it hasn't been updated with Chelsea's result from yesterday at the time I'm writing this).

http://www.sportsclubstats.com/England/Premier.html

So a figure in the 80s seems to be standard, and you can look to bookmakers too - Paddy Power have us at 1/9, which means they think we have about an 84% chance too.

Plus don't forget Chelsea have four big six matches left, including one against Spurs where one or both of them will drop points, and Spurs also have to face Man City whereas we don't.

groove · « **Reply #990 on:** February 13, 2018, 04:39:34 pm »

1/9 = 90% but whatevs, we're in a good spot! Pretty much every xg model I've seen has us or Spurs as 2nd best team in the league. Coupled with a favourable fixture list it all looks good.

BassTunedToRed · « **Reply #991 on:** February 13, 2018, 04:42:20 pm »

Quote from: groove on February 13, 2018, 04:39:34 pm

1/9 = 90% but whatevs, we're in a good spot!

I was allowing for their edge which makes sure they're always in profit

But yep, who cares really. Up the Reds!

groove · « **Reply #992 on:** February 13, 2018, 04:46:45 pm »

Ohhhh of course.

JCB · « **Reply #993 on:** February 28, 2018, 11:45:25 am »

A nice little article about our stats with and without Coutinho prompted by our very own BassTunedToRed.

https://liverpooloffside.sbnation.com/liverpool-fc-tactics-longform/2018/2/28/17057204/the-curious-case-of-liverpool-without-philippe-coutinho-statistical-analysis-barcelona-transfer-142m

BassTunedToRed · « **Reply #994 on:** February 28, 2018, 12:20:51 pm »

Quote from: JCB on February 28, 2018, 11:45:25 am

A nice little article about our stats with and without Coutinho prompted by our very own BassTunedToRed.

https://liverpooloffside.sbnation.com/liverpool-fc-tactics-longform/2018/2/28/17057204/the-curious-case-of-liverpool-without-philippe-coutinho-statistical-analysis-barcelona-transfer-142m

Hadn't seen that - cheers!

Nessy76 · « **Reply #995 on:** February 28, 2018, 12:31:15 pm »

Quote from: JCB on February 28, 2018, 11:45:25 am

A nice little article about our stats with and without Coutinho prompted by our very own BassTunedToRed.

https://liverpooloffside.sbnation.com/liverpool-fc-tactics-longform/2018/2/28/17057204/the-curious-case-of-liverpool-without-philippe-coutinho-statistical-analysis-barcelona-transfer-142m

Interesting piece. Of course, it glosses over some other factors, like the presence of Van Dijk and Karius, both of which are arguably as much of a change as the absence of Phil.

BassTunedToRed · « **Reply #996 on:** February 28, 2018, 12:36:55 pm »

Quote from: Nessy76 on February 28, 2018, 12:31:15 pm

Interesting piece. Of course, it glosses over some other factors, like the presence of Van Dijk and Karius, both of which are arguably as much of a change as the absence of Phil.

Whilst they're obviously relevant, they're not the sole reason - 13 of the 20 matches without Coutinho haven't had van Dijk either, and the goalkeeper split is currently 10 each.

Nessy76 · « **Reply #997 on:** February 28, 2018, 01:02:55 pm »

Quote from: BassTunedToRed on February 28, 2018, 12:36:55 pm

Whilst they're obviously relevant, they're not the sole reason - 13 of the 20 matches without Coutinho haven't had van Dijk either, and the goalkeeper split is currently 10 each.

True, and it's always impossible to get a perfect comparison with anything like this. We also changed our style of play somewhat, with Klopp asking for more discipline from the full backs.

Still, in terms of the attacking output of the team, it is definitely interesting to see that Coutinho has really not been missed.

JackWard33 · « **Reply #998 on:** February 28, 2018, 02:09:27 pm »

Quote from: BassTunedToRed on February 28, 2018, 12:36:55 pm

Whilst they're obviously relevant, they're not the sole reason - 13 of the 20 matches without Coutinho haven't had van Dijk either, and the goalkeeper split is currently 10 each.

Do you see any merit in these with 1 player vs without 1 player stats / metrics
I mean its an 11 man game ... how can there ever not be a shit ton of noise in the comparisons - given that its often going to be a small sample in the first place - surely they're close to worthless?

JCB · « **Reply #999 on:** February 28, 2018, 02:26:33 pm »

What I found interesting was the frequency distribution graph. I've taken the liberty of adding a rough area graph to simulate a histogram distribution and to push the point more.

Our mode has deviated from the 0-1 goal which implies there's a higher probability of a win. Of course both Nessy's and Jack's points have merit and it's not like we will categorically state "Look, correlation...therefore cause". But it' a pattern nevertheless and warrants pointing out.

There might be a lot of noise here, but to paraphrase Nate Silver it doesn't mean there isn't any signal in the noise.