Author Topic: Statistical Comparison of LFC players to Rest of EPL - starting w/LB  (Read 33228 times)

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #160 on: July 23, 2017, 12:32:42 pm »
Trent and Matip were the two I saw. Trent wasn't really to Karius but was lax and there was some sloppy passing all over the pitch tbf. I put it down to the legs just about gone by this game due to the double sessions including one before the game I believe?

Yeah it wasnt a criticism of whoever, itīs pre-season (although Moreno isnīt afforded the same caveat for what people wrongly perceive as mistakes).

I love that by the way, I know many wonīt but Iīve heard about fatigue training before and itīs uses. It doesnīt matter how good you are when you are fresh, how good are you when you are fucked? Because in this team you are going to press and run your ass off so you need to still be able to pass, tackle, shoot and dribble when your legs are starting to wobble as that is what will happen at times.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline penga

  • What you get if you cross Pingu with Jenga.....
  • Legacy Fan
  • ******
  • Posts: 2,662
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #161 on: July 23, 2017, 12:35:10 pm »
Funny you mention that, if you look at Flanagan v Moreno, the two tackle monsters in the group. Moreno makes a huge number of tackles (almost 2x the average) but only getīs dribbled past very slightly above average (1,09 vs 0,96). This indicates that his tackles donīt always result in him being beaten and his pace allows him to remain in the game defensively speaking.

Flanno on the other hand is about 2.5x above average in tackles and 2x above the average in dribbles past also which indicates his pace might not be good enough to get him back in the game.

So I guess what it is showing is Moreno is diving in too much but because of his pace, his numbers are not as bad as they otherwise could be maybe?
What exactly constitutes dribbled past though?

For example Milner often dives in for a tackle and misses or simply gets run past without diving in, but then turns and gets into a great position to subsequently block a cross or at least recover his position - now does the initial action still count as dribbled past? If so the end result is still neutral for the attacking team vs Milner which is perhaps why he is perceived as solid even though his dribbled past stat is high.

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #162 on: July 23, 2017, 12:41:31 pm »
That's fair but at the end of the day we are comparing fullbacks and putting forward opinions/assertions mainly based on stats and limited viewings here because I'm assuming none of us actually watched all these fullbacks for the majority of the games they played except for our own. These opinions/assertions can be deemed educated guesses that help us form a reasonable perception of a player- always with the qualification that we haven't watched them enough to be 100% sure which is what we are all doing here isn't it?

Maybe I'm questioning why there are basically no negative assertions or reservations about the stats analysed by you on Robertson in this thread yet? You probably don't need me to bring them up, I just did because you missed some out. You could have included amount of aerial duels won in the Squawka graphic just like amount tackles won in another graph in your initial analysis but didn't - why not? The stat goes against your limited viewing yet you ran with the opinion that he has no real weaknesses. Also when you provide a positive assertion based on stats by Robertson how come you haven't really tried to argue against it and muse how those stats could be misleading for example? It seems a positive is a positive and a negative is argued into a positive in your opinion of Robertson so far.

A more balanced debate might have looked like this:

- Robertson has good 1v1 numbers when we take into account tackles per 90, tackle success, fouls per 90 and dribbled past per 90 compared to other fullbacks. (Basically what you have been saying in this thread), however he gave away 2 penalties (which you said in another thread) and could possibly be more exposed playing in our system as we allow opposition wingers more space.
- Robertson has decent % won in aerial duels but the amount is extremely low - 2nd last compared to other fullbacks suggesting he is passive but I don't know if it is really a problem (as opposed to suggesting he is solid based on your limited viewing which goes against the stats)
1. We havenīt got there yet
2. We havenīt got there yet

Seems like you just want to skip to the parts you want instead of following the breakdown of the entire position of full back systematically as I was trying to do in here. When I posted the squawka visual, I said the end aim was to post something like that but with more depth and proper analysis. Key word: End.

Every attempt to drag me into a debate about data we havenīt even touched on yet has to be met the same way - we will get to that.

As for his tackling, his success rate is weak, not good. But is still better than Milners if that is the standard bearer he needs to beat. Also his tackles per 90 were a scratch below average (1,74 vs 2,10). Whatīs that, 10% below average. Ghoulam was less than 1/2 of the average. Also in his league, there are even more tackles committed by full backs which means if we were to make adjustments for the league, he could end up being with 1/3 or 1/4 of the average. That 10% difference which should make a negligible impact on results vs that huge difference which could have a massive impact on results is the difference between being in the acceptable range and being a outlier which is to be treated with suspicion.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #163 on: July 23, 2017, 12:47:53 pm »
What exactly constitutes dribbled past though?

For example Milner often dives in for a tackle and misses or simply gets run past without diving in, but then turns and gets into a great position to subsequently block a cross or at least recover his position - now does the initial action still count as dribbled past? If so the end result is still neutral for the attacking team vs Milner which is perhaps why he is perceived as solid even though his dribbled past stat is high.

Opta offer no explanation for it. But then if you treat data collection with suspicion then things like this become entirely pointless. You either have to trust that the data from Opta is to a high enough accuracy to be useful, or itīs not and ignore stats completely. We donīt really have the time to analyse every single tackle, dribbled past and block for every player on the list. Even if we did we would then just end up arguing the toss over what something is or isnīt.

Morenoīs slide last night = attemped blocked shot? Attempted tackle? Dribbled past? Attempted trench digging before a ground assault on Hong Kong?

Stats cut through that conjecture. Those who accept them for what they are can use them and debate over the big picture. Those who donīt trust them will still be arguing the toss over 1 individual instance of something that constitutes 0.00001% of what that player did in the game.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #164 on: July 23, 2017, 01:22:53 pm »
Block Refining
So I mentioned yesterday some problems with how the data was being looked up/imported into the table I am working on. One of the drawbacks of living in a 3rd World Country is that I would need to sell a kidney to get a new computer right now such is the price of technology here and the computing power of the ones I have isnīt sufficient to really work on a 54mb spreadsheet which is what you end up with when you have loads of opta data from loads of leagues, most of which you donīt need.

Having redone all the formulae and triple checked them, this is where we are with the table now.

Player      Crosses    Passes   Shots      Total Blocks    Scouting Comments?
Erik Pieters      0,83    1,14   0,57      2,54    ????
Ryan Bertrand      0,68    1,29   0,36      2,33    ????
Christian Fuchs      0,9    1,02   0,38      2,3    ????
Aaron Cresswell      1,28    0,72   0,21      2,21    ????
James Milner      0,97    1,08   0,14      2,19    ????
Marcos Alonso      0,57    1,34   0,17      2,08    ????
Nacho Monreal      0,54    1,17   0,31      2,02    ????
Andrew Robertson      0,82    0,69   0,49      2    ????
Danny Rose      0,59    1,17   0,18      1,94    ????
Stephen Ward      0,78    0,67   0,45      1,9    ????
Alberto Moreno      0,53    1,13   0,18      1,84    ????
Joe Average      0,63    0,84   0,29      1,75    ????
José Holebas      0,84    0,43   0,31      1,58    ????
George Friend      0,6    0,46   0,51      1,57    ????
Leighton Baines      0,57    0,8   0,19      1,56    ????
Charlie Daniels      0,53    0,53   0,38      1,44    ????
Gaël Clichy      0,68    0,68   0,08      1,44    ????
Allan-Roméo Nyom      0,48    0,68   0,27      1,43    ????
Patrick van Aanholt      0,4    0,65   0,32      1,37    ????
Nathaniel Clyne      0,3    1   0,05      1,35    ????
Kyle Walker      0,3    0,8   0,23      1,33    ????
Ben Davies      0,31    0,62   0,36      1,29    ????
Jon Flanagan      0,28    0,33   0,19      0,8    ????

Ignore the scouting comments for now, we will get back to those later when we can analyse things. One quick thought though - are Clyne, Milner & Morenoīs similar blocked passing numbers a pressing thing? Rose would support that theory, Ben Davies doesn't. Yet logically pressing results in blocked passes. Does that mean Davies isnīt very good at pressing hence why Rose is preferred normally, Davies for certain games when his 1-v-1 strength is more important than pressing weakness?

Anyway, here is what that looks like in a visual format.



Here is how the 1-v-1 vs Blocks visual now.



I am starting to see a direct correlation now between how "good" people appear on this graph and the amount of tackles they make. That makes sense because if you are diving in for tackles you are taking charge of the situation and not waiting for the winger to try to beat you, cross, pass, shoot. Obviously tackling isnīt always the right move and even the best will only have a ~85% accuracy at it (The highest here is Fuchs on 82%).

At one extreme you have Flanagan (5,64 per game) who dives in far more than anybody else here. At the other end you have Bertrand and Cresswell who rarely dive in (1,33 & 0,89).

Cresswell is most interesting to me here. I was a season ticket holder at Ipswich and since then have watched as many of their games I can. Cresswell always stood out. Infact in many ways he reminds me of Robinson. But he never stood out defensively really. He is showing up here as a player who doesnīt foul much, doesnīt get beat much, blocks an absolutely insane amount of crosses and has decent numbers in blocking passes and shots too. I would suggest I have probably seen more of him than anybody else here and yet I just cannot correlate that with my viewing off him. He was never weak defensively but he certainly didnīt stand out either.

When I look at how West Ham concede goals it suggests they are actually weak at defending attacks down the wings, weak in defending against skillful players, weak at defending through balls and very strong at dealing with crosses. This made me think back to how Hodgson set us up. He would focus on shape above anything else, almost to the extent that it was all that matter. He tucked his full backs in and invited teams to cross against us. They wouldnīt commit to tackles to stay in shape therefore we struggled to regain possession.

I genuinely didnīt watch much of West Ham last season and when I did I was pretty much only watching Lanzini or the opposition, but does that match what they were doing tactically as that would explain lots of outliers Cresswell is showing. West Ham were the lowest tackling side in the league by some distance. They also lead the league in interceptions by some distance which are both signs of a very compact side who favour staying in shape over winning the ball.

This makes me think at some point we need to come back and make adjustments towards average based on the teams people play for to allow for possession, tactics, etc.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline JCB

  • Sponsors of Digger Barnes.....
  • Legacy Fan
  • ******
  • Posts: 2,021
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #165 on: July 23, 2017, 01:35:32 pm »
What exactly constitutes dribbled past though?

For example Milner often dives in for a tackle and misses or simply gets run past without diving in, but then turns and gets into a great position to subsequently block a cross or at least recover his position - now does the initial action still count as dribbled past? If so the end result is still neutral for the attacking team vs Milner which is perhaps why he is perceived as solid even though his dribbled past stat is high.

Here's a link I came across the other day that might interest you. Similiar issue, but in relation to interceptions instead. Highlights the grey area that is sports data collection.

https://everyteamneedsaron.com/2016/08/01/our-lexicon-for-defending-in-football-isnt-good-enough/

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #166 on: July 23, 2017, 01:40:18 pm »
Duels

Something I havenīt included yet in our workings is duels, this is because I am having some trouble deciding what exactly they are.

According to Opta, a duel is any 50/50 encounter on the pitch. That made me want to include them in 1-v-1 calculations but decided to research further.

According to squawka:-

Duels
A failed tackle – A tackle that hasn’t won possession, could be a foul. Take on – Attempted dribble past an opponent. Aerial – Aerial duel 50/50 when ball is in the air. A lost aerial duel – An aerial duel that has been lost, could be a foul.
Read more at http://www.squawka.com/football-stats-definitions


This means I am on the path I want to be on now in breaking down duels which is an all encompassing metric that doesnīt tell us much other than the player does a lot of stuff and the accuracy it does that combined stuff with. I always thought duels would be a good meassure of 1-v-1 ability and yet I can see now that it includes dribbling and the amount of fouls the player wins.

I think more than anything else, duels simply shows you have aggressive a player is. This isnīt necessarily always a good thing. Itīs certainly not a bad thing either though. Its not something I think we will use though as we are already using the underlying data that feeds it in various categories.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #167 on: July 23, 2017, 02:17:07 pm »
Food for Thought - Liverpool & Spurs have by far the lowest number of interceptions in the league. Spurs have around 8 per game, Liverpool 10, 3rd lowest is Swansea with 12. West Ham lead the table with over 16 per game.

Interceptions is always used to identify players are would be good in a pressing system but should it be?

Itīs not a possession thing as Manchester United & Arsenal are both near the top of the table in possession and interceptions. Swansea are around average possession and are 3rd lowest in interceptions. Maybe it has an inverse correlation to tackles somehow? Liverpool are 2nd bottom in interceptions and 2nd top in tackles. West Ham have the most interceptions, least tackles. Nice, we might have something. Yet Spurs are bottom in interceptions and middle of the pack in tackles. Bournemouth are bottom 4 in both categories.

Spurs are considered one of the best teams defensively and yet are average at tackling and the worst in terms of intercepting the ball. Maybe recoveries? Liverpool are the best at recovering loose/second balls (huh?) and Spurs are one of the worst (huh?). One metric where all the teams who concede the fewest chances per game excel is "PutThrough/Blocked Distribution Won". (Huh?). Thatīs Optaīs way of saying a blocked forward pass.

Obviously no one metric can tell you how good or bad a side is at defending but the weakest teams in this metric are:-

Team      Blocked Passes
Hull City         269
Sunderland         269
Bournemouth         264
Stoke City         263

I think we would all agree those sides are pretty much all very shit at stopping teams from getting good scoring chances against them.

At the other end it gets more difficult. These are the best sides at blocking forward passes:-

Team      Blocked Passes
Southampton    383
Leicester City    362
Manchester United    338
Chelsea    333
Swansea City    325
Tottenham Hotspur    325
Liverpool    324

So it seems just because you are good at this, doesnīt mean you wonīt concede goals. But being bad at this will almost certainly result in your defence being opened up. There doesnīt appear to be any correlation to this and any other defensive metric. I figured blocking passes would result in lots of interceptions &/or recoveries but it appears not.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline Thats So Dimitar

  • Kopite
  • *****
  • Posts: 608
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #168 on: July 23, 2017, 03:44:23 pm »
Just read all 5 pages of the thread and have found it extremely interesting. Hopefully as a Spurs fan I can perhaps offer some biased insight in to our own fullbacks since they seem to be the ones most compared against so far.

One thing I think that is missing which I feel would be near impossible to have stats for is situations where players are making tackles. Rose I feel attempts a lot of tackles near halfway and further up the pitch. If he fails here that is bad news which is why I think he may go in with a mentality of succeed at tackling or end their play with a foul. For dribbled past I think your thoughts earlier sound about right, both Walker and Rose seem happy to let wingers carry the ball down the lone, confident they can keep up with/catch them. I assume based on what I saw you say earlier this results in high dribbled past numbers?

That last post does seem bizarre, if a team hits the ball long and the defender gets it is that an interception? If not that might go some way to explain it, teams go long to try get in behind and winning the ball off that would not count as a tackle or interception.
We beat Real Madrid

Offline penga

  • What you get if you cross Pingu with Jenga.....
  • Legacy Fan
  • ******
  • Posts: 2,662
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #169 on: July 23, 2017, 04:09:54 pm »
Here's a link I came across the other day that might interest you. Similiar issue, but in relation to interceptions instead. Highlights the grey area that is sports data collection.

https://everyteamneedsaron.com/2016/08/01/our-lexicon-for-defending-in-football-isnt-good-enough/
That author is correct! Nice article.

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #170 on: July 23, 2017, 04:17:08 pm »
Just read all 5 pages of the thread and have found it extremely interesting. Hopefully as a Spurs fan I can perhaps offer some biased insight in to our own fullbacks since they seem to be the ones most compared against so far.

One thing I think that is missing which I feel would be near impossible to have stats for is situations where players are making tackles. Rose I feel attempts a lot of tackles near halfway and further up the pitch. If he fails here that is bad news which is why I think he may go in with a mentality of succeed at tackling or end their play with a foul. For dribbled past I think your thoughts earlier sound about right, both Walker and Rose seem happy to let wingers carry the ball down the lone, confident they can keep up with/catch them. I assume based on what I saw you say earlier this results in high dribbled past numbers?

That last post does seem bizarre, if a team hits the ball long and the defender gets it is that an interception? If not that might go some way to explain it, teams go long to try get in behind and winning the ball off that would not count as a tackle or interception.

No, dribbled past requires a gain of position over the other. If you are just running down the line side by side pretty much, you are not going past anybody. The point that changes is when the player gets infront of you on the same line on the pitch (cutting across for example).

As for that last one in bold, I think they are counted as recoveries which makes sense based on what I have seen when comparing Lucasī stats to events that occur in the game.  Here are Optaīs definations:-

Interception
This is where a player intentionally intercepts a pass by moving into the line of the intended ball.

Recovery
This is where a player wins back the ball when it has gone loose or where the ball has been played directly to him.

So I think Inceptions is when you press passing lanes, like how Guardiolaīs Barcelona did a lot. They would press the ball carrier while others would press lanes they felt the ball would be played into to try and intercept passes. Whereas Klopp presses the ball carrier and players use their cover shadows to prevent passes to teammates which usually results in long balls & therefore recoveries.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #171 on: July 23, 2017, 04:25:05 pm »
I guess the overall point I was making about interceptions is that, looking at the numbers for Liverpool and Spurs who most would name as the best pressing sides in the league at the moment, it seems interceptions has a low correlation to pressing. However, blocked passes is something I have never really seen used when analysing players and seems to have a much higher correlation.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline Flinstone

  • Whatever you like. Would like to slip it in without making it awkward... Chinese Information Minister
  • Legacy Fan
  • ******
  • Posts: 3,747
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #172 on: July 24, 2017, 01:25:34 am »

 One of the drawbacks of living in a 3rd World Country

Brazil is a first world country. Quite sure they were firmly aligned with the Americans when the Cold War was raging
The West is finished, China is the future

Offline RayPhilAlan

  • Legacy Fan
  • ******
  • Posts: 2,132
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #173 on: July 24, 2017, 01:31:15 am »
Just wondering if there's any way of measuring that essential fullback skill where the winger comes at you, you don't give him a sniff, and he turns and plays it back.

It's not a duel, as there's no take-on, no possibility of a change in possession. But it's a small victory for the fullback every time it happens.

Back in the early pages of the Keita thread there was talk of 'packing' - an attempt to measure how much a player's involvement advances play towards the opponent's goal - and it is pretty obviously a strength of Naby's.

I wonder if there is also a 'negative packing' measurement, where you can assess how a fullback can force play back away from his goal. Not sure if you or any private individuals would have access to this sort of data, but it might reveal some answers.


Offline Anfield89

  • Legacy Fan
  • ******
  • Posts: 2,986
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #174 on: July 24, 2017, 01:49:00 am »
He flew by at speed but he was back up and ready quickly, while the right back or other central defender were nowhere to be seen.

He wasn't ready Klavan and the keeper moped up. Physically he has got the tools he just doesn't have any awareness or composure. Needs to learn and shouldn't be in the first team until he does.

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #175 on: July 24, 2017, 02:50:16 am »
Brazil is a first world country. Quite sure they were firmly aligned with the Americans when the Cold War was raging

While the was first used to determine a countries allegiance, it is now used to discuss how developed a country is. Itīs a subject of contention here. The government believe they are not and self-declared they are not to the WTO. Yet a lot of the markers like infant mortality rates, education, infrastructure, human development index, poverty of majority of citizens, high inflation, life expectancy at time of birth, etc etc all exist outside of a handful of major cities in the biggest states. The government decided that investing in football stadiums was of more value than investing in many of these things and therefore the people of Brazil will state they are a third world country until nobody in the country needs to live their life in 3rd world conditions.

This is ignoring the fact that the government here is outrageously corrupt and their actions are almost entirely out of self interest and preservation due to the ongoing corruption investigations which has resulted in many threats of the military seizing power if the government tries to interfere or shut down the investigations.

So while Brazils politicians are declaring us equals with China, Russia, Saudi Arabia etc in terms of development, those living here under the ongoing threat of military dictatorship in conditions you would find in similar parts of Africa disagree. Whilst I am alright here in Rio where they realise tourists will not come to if they donīt invest, I believe it only matters how the worst of us live, not the best and support their decision to declare us a 3rd world country until such times the conditions for the poorest of us improve.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #176 on: July 24, 2017, 02:55:58 am »
Just wondering if there's any way of measuring that essential fullback skill where the winger comes at you, you don't give him a sniff, and he turns and plays it back.

It's not a duel, as there's no take-on, no possibility of a change in possession. But it's a small victory for the fullback every time it happens.

Back in the early pages of the Keita thread there was talk of 'packing' - an attempt to measure how much a player's involvement advances play towards the opponent's goal - and it is pretty obviously a strength of Naby's.

I wonder if there is also a 'negative packing' measurement, where you can assess how a fullback can force play back away from his goal. Not sure if you or any private individuals would have access to this sort of data, but it might reveal some answers.

Would love that and itīs something I was actually discussing by PM. I said right at the start of the thread too when discussing tackling that not tackling and instead pushing people into less dangerous areas and eventually away from goal is the perfect outcome as you are then never taking any risk unlike making a tackle.

Btw - I know that someone on the anfield index manually collates all that information, or used to. At the time he did it, Henderson & Ojo both scored pretty highly on a consistent basis when they played which was why I was pretty hype about him for 2016/17 (especially when you look at this production per 90 despite being small sample size). Shame injury curtailed that and itīs possible that window of opportunity may have closed (though hopefully only temporarily).

I agree they would be useful to know and I was encouraging people to watch those with low tackling numbers to see if this was why, or if they were too passive (either by instruction of by choice) or just simply were not involved in many situations where tackling was a required action of them. During the season I hope to pay more attention to people like Cresswell & Bertrand and see why their tackling numbers are so and yet they are considered good full backs.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Pressing per 90?
« Reply #177 on: July 24, 2017, 10:05:07 pm »
So I have today been looking at Recoveries, Interceptions, Tackles & Blocked Passes all per 90. I figure these numbers are a strong indicator of how much pressing / defending a player has to do per 90.

Itīs very hard to separate this out to show whether it is active or passive. For example, Klopp's Liverpools defensive numbers will be mostly active. We hunt down the ball, force mistakes, etc. Whereas maybe Hodgson's Liverpool is more passive. The focus is primarily on maintaining shape and defensive numbers will be obtained when players or the ball enters into the defensive shape.

For example, if you are a CM for Hodgson, you focus on shape and pick up interceptions when people try to pass through midfield to the forwards through your area of influence. However, if you are a CM for Klopp, when we lose the ball our wide players press full backs and try to force loose passes into midfield which our midfield will press the receiver before the ball is played aiming to intercept.

Now, all of that is tactical instruction but hopefully the underlying numbers indicate what type of defending players get through. For example, I see Tackles/Blocked Passes as being someone who would aim to win the ball more. Therefore maybe someone better at man marking? Whereas interceptions and recoveries indicates someone who pounces on lots of second balls and lose passes, so maybe someone whose positioning and reading of the game is strong.

I think more than any other area, this is the one we will need to make some adjustments to the numbers for them to make sense as these will be mostly based on team, system & tactics more than the individual when comparing them. For example, Liverpool are high in tackles, recoveries, blocked passes and low in interceptions. Hull are low in blocked passes & tackles, high on recoveries as people play balls into the low block which are not completed. This will mean Liverpool players will show as very active defensively (pressing) and Hull players will show as very passive (low-block).

Here is the table of data to work with:-
Player      Recoveries/90    Interceptions/90   Tackles+Blocked Passes/90      Recoveries+Interceptions/90    Total Actions/90
Jon Flanagan      5,4    1,9   5,97      7,3    13,27
Alberto Moreno      4,87    2,23   5,47      7,1    12,57
Christian Fuchs      5,95    2,06   3,8      8,01    11,81
Nacho Monreal      5,42    2,51   3,74      7,93    11,67
George Friend      5,33    2,57   3,31      7,9    11,21
Erik Pieters      5,46    1,54   4,14      7    11,14
José Holebas      6,96    1,7   2,35      8,66    11,01
James Milner      5,9    1,25   3,76      7,15    10,91
Gaël Clichy      5,59    2,24   2,54      7,83    10,37
Allan-Roméo Nyom      5,15    2,01   3,21      7,16    10,37
Patrick van Aanholt      6,12    2,05   2,2      8,17    10,37
Andrew Robertson      6,02    1,81   2,43      7,83    10,26
Joe Average      5,24    1,7   3,24      6,94    10,17
Danny Rose      4,76    1,29   4,11      6,05    10,16
Kyle Walker      5,29    1,26   3,2      6,55    9,75
Stephen Ward      4,93    2,06   2,62      6,99    9,61
Marcos Alonso      4,88    1,47   3,21      6,35    9,56
Leighton Baines      5,14    1,44   2,81      6,58    9,39
Ryan Bertrand      5,06    1,61   2,62      6,67    9,29
Nathaniel Clyne      4,71    1,16   3,17      5,87    9,04
Charlie Daniels      4,88    1,5   2,27      6,38    8,65
Aaron Cresswell      4,34    1,06   1,61      5,4    7,01
Ben Davies      3,16    0,62   2,64      3,78    6,42

More and more, Danny Rose and Walker look very similar players. That makes sense and passes the eye test. You can see also here how they are both in the same system. However, Davies looks like he plays in an entirely different team/system entirely. Not sure what that means. Is Davies being used to shut down a flank in certain game situations? The same with Clyne vs Milner. Also Clyne/Milner vs Flanno/Moreno.

As you can see here, this is where I have a big problem with traditional stats which look at high activity as a good thing. If you are going from challenge, to interception, to tackle, to recovery... it suggests that player runs around a lot and might struggle to stay in his position and could be leaving big holes in the formation. Moreno and Flanagan come out of this scoring well and yet from watching them both I know they are both weak links defensively.

So please, donīt use numbers here as a score where more = good or even less = good. I think we will need to adjust these for each team to negate systems & tactics from the numbers and then we should be able to make better sense of them.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline redmark

  • RAWK Scribe
  • Legacy Fan
  • ******
  • Posts: 21,395
Re: An unbiased look at our full backs
« Reply #178 on: July 24, 2017, 10:15:59 pm »
So I mentioned yesterday some problems with how the data was being looked up/imported into the table I am working on. One of the drawbacks of living in a 3rd World Country is that I would need to sell a kidney to get a new computer right now such is the price of technology here and the computing power of the ones I have isnīt sufficient to really work on a 54mb spreadsheet which is what you end up with when you have loads of opta data from loads of leagues, most of which you donīt need.

I'm sure I speak for everyone when I say that our eagerness for your analysis updates and possible improvements probably outweighs your need for a spare kidney.

;)

Fascinating thread.

Stop whining : https://spiritofshankly.com/ : https://thefsa.org.uk/join/ : https://reclaimourgame.com/
The focus now should not be on who the owners are, but limits on what owners can do without formal supporter agreement. At all clubs.

Offline nellyp

  • hant. All Hail the old new Baldrick!!!
  • Anny Roader
  • ****
  • Posts: 273
  • Life's not black and white, it's red and white
Re: An unbiased look at our full backs
« Reply #179 on: July 25, 2017, 07:46:58 am »
I'm sure I speak for everyone when I say that our eagerness for your analysis updates and possible improvements probably outweighs your need for a spare kidney.

;)

Fascinating thread.

I agree. Though it is a tad confusing to a mere mortal like me  :'(
In a world of uncertainty and worry, it is calming to know that as a fan of the greatest club in the world I will never walk alone

Offline Rush 82

  • Seth Iffricans don't take the dog out for a walk - they take the line out!
  • Lead Matchday Commentator
  • Legacy Fan
  • ******
  • Posts: 22,105
  • From Cape Town to Anfield
Re: An unbiased look at our full backs
« Reply #180 on: July 25, 2017, 09:08:37 am »
Block Refining
So I mentioned yesterday some problems with how the data was being looked up/imported into the table I am working on. One of the drawbacks of living in a 3rd World Country is that I would need to sell a kidney to get a new computer right now such is the price of technology here and the computing power of the ones I have isnīt sufficient to really work on a 54mb spreadsheet which is what you end up with when you have loads of opta data from loads of leagues, most of which you donīt need.


What's the specs of the computer you're currently working on?

Offline Klippity Klopp

  • LFC. For life! Yes! Not just for Kloppmas.....nooo.....
  • Legacy Fan
  • ******
  • Posts: 3,438
  • We all Live in a Red and White Kop
Re: Pressing per 90?
« Reply #181 on: July 25, 2017, 10:27:46 am »
More and more, Danny Rose and Walker look very similar players. That makes sense and passes the eye test. You can see also here how they are both in the same system. However, Davies looks like he plays in an entirely different team/system entirely. Not sure what that means. Is Davies being used to shut down a flank in certain game situations?

I don't think Spurs, see a big threat on the opposition right wing, like a Mane or Mahrez and feel they should play Davies to shut down that position over Rose.

Davies got lots of minutes because Rose was out last season with an injury.  Otherwise Rose would have played every game over Davies, as Poch prefers him. 

So I don't feel he's being used to shut down a flank in certain situations.  He's just a different type of player to Rose.

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: Pressing per 90?
« Reply #182 on: July 25, 2017, 12:31:20 pm »
I don't think Spurs, see a big threat on the opposition right wing, like a Mane or Mahrez and feel they should play Davies to shut down that position over Rose.

Davies got lots of minutes because Rose was out last season with an injury.  Otherwise Rose would have played every game over Davies, as Poch prefers him. 

So I don't feel he's being used to shut down a flank in certain situations.  He's just a different type of player to Rose.

That is great news then because it shows - like Clyne v Milner / Gibbs v Monreal - that the individual shows through more than the style/formation/system. I was concerned the opposite might be true at times. But then it makes sense that managers would have similar full backs as he is buying what he needs.

Interestingly Pochetino inherited Walker & Rose. Yet signed Davies & Trippier. That should give us an idea what Pochetino* is looking for in full backs for that position and the differences may show us which weaknesses are acceptable for that position for Pochetino.

* Hopefully our Spurs fan following the thread can chime in here. I have no idea how recruitment works or used to work at the time both were signed. I believe at most clubs the manager has minimal input into transfers other than choosing who he wants to sign. How that choise is arrived at can range from him speaking with a team of scouts to something similar to what Liverpool has - analysts, coaches & scouts forming a committee to discuss and agree on transfers, managers say is final.

What's the specs of the computer you're currently working on?
Itīs a Gateway ne56r08b.

I agree. Though it is a tad confusing to a mere mortal like me  :'(

You'll be fine mate. We are basically just trying to look at data in a different way to anything I have seen on the internet. Trying to think of specific tasks players undertake on the pitch and imagining what that would look like in terms of data. Then seeing if creating a metric out of all that data (e.g. tackles vs fouls+dribbled past) can give you an idea of how good a certain player is to all others in terms of 1-v-1 defending. So instead of just looking at tackles, interceptions, dribbles, passes etc, breaking the position up into situations and trying to score based on that. Balls into the box, 1-v-1 defending, pressing/second balls, etc

It will start getting more complicated though once we get to the end. At the moment I am just using raw data but I plan to start making adjustments for things to reflect how playing in a certain team/tactical shape/system/style/league etc can impact on the numbers you produce. For example, Spurs as a team produce 1/2 as many inceptions per game as West Ham. Therefore adjusting Spurs players towards some sort of global average and West Hamīs also should show us how much Cresswellīs strange numbers are due to the fact he simply plays for West Ham.

It should also maybe give us an idea of predicting what sort of numbers Robertson should produce at Liverpool (and others who change clubs during the window) and answer that age old question "But what would he be like in a better team". Then over the course of the season we can see how close our predictions are to what Robertson actually does on the pitch. Then maybe refine what we are doing if our predictions are way off in January.

I also want to try and do this for center backs and look at Lindelof v Van Dijk (assuming he moves) v Keane and try to see if there is a way to predict where they will succeed and fail in a change of system/league/team.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline JCB

  • Sponsors of Digger Barnes.....
  • Legacy Fan
  • ******
  • Posts: 2,021
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #183 on: July 25, 2017, 01:22:37 pm »
Thought I'd quickly visualise Babu's data set for easier reading in my lunch break...

I know he said not to assign relevance to the scores here, but the conditional formatting helps visually distinguish the data in the table.


And I threw these into Tableau quickly....
Basically I let Tableau automatically cluster them into groups so it's somewhat arbritrary and therefore not carefully considered but should aid it somewhat:


edit: the above was quickly done to check if indeed there was a correlation between Recoveries and Interceptions as Babu suggested. It does seem to

« Last Edit: July 25, 2017, 01:31:44 pm by JCB »

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #184 on: July 25, 2017, 02:22:39 pm »
Thought I'd quickly visualise Babu's data set for easier reading in my lunch break...

I know he said not to assign relevance to the scores here, but the conditional formatting helps visually distinguish the data in the table.


And I threw these into Tableau quickly....
Basically I let Tableau automatically cluster them into groups so it's somewhat arbritrary and therefore not carefully considered but should aid it somewhat:


edit: the above was quickly done to check if indeed there was a correlation between Recoveries and Interceptions as Babu suggested. It does seem to



Here is something I cannot get my head around still. There is a correlation between interceptions and recoveries at player level and an inverse correlation at team level :D

We are top in recoveries and 2nd bottom in interceptions.

The way I arrive at this as a possibility is this.

Recoveries = Passes played directly to an opposing player. Therefore they include forced pass errors. Therefore pressing the ball carrier will result in more recoveries as players are rushed to play the ball. Therefore I see recoveries (+ blocked passes) now as a sign of how aggressively you press teams. Whereas interceptions is where a player moves into the line of a pass to intercept. Therefore a side that focuses on shape (Hull) or focuses on pressing passing lanes (Man City) will have higher Interceptions vs Possession than a side like Liverpool who press the players.

However players who read the game or have better reactions will always be the ones who pick up interceptions & recoveries for a team. Therefore this is how we can differentiate between player skill sets and numbers generated by tactics/system. And this is where I have a major problem with how data is used now for players. Manchester United are more passive in how they defend and therefore rack up interceptions, the likes of Spurs and Liverpool do not. Yet this is one of the key categories used on the radars. On top of this, Interceptions is possession adjusted which means the two highest teams in interceptions (Arsenal & Manchester United) get their players numbers increased even further as they are two of the highest in terms of possession.

Basically, tackles & interceptions volumes are used to measure defending at the moment and I see no direct correlation to how good a defender is at those numbers at the moment. For example, before I even adjust for possession my data shows that Moreno and Flanagan are the two best full backs in the Premier League in terms of T+I. Does anybody believe that to be true? At the other end of the scale, Davies, Bertrand & Clyne are in the bottom 5. Does anybody believe that to be true either? How many times have you seen Flanno & Moreno have defensive problems? For me itīs a lot. Yet they are the best two in the league. In contrast how many times have you seen Bertrand, Clyne or Davies struggle? Itīs very rare. So this measurement simply HAS to be wrong. It cannot be as simple as the people who do the least = better either but we can at least agree that tackles + interceptions tells us literally nothing at all other than the players are chasing the ball a lot perhaps. When I word it like that, you can see how it stops being a good thing immediately. Defending isnīt about chasing the ball at all. So why are two metrics that involve you trying to win the ball the only two used as a measurement of how good you are defensively?

This is pretty much why I am doing this. Stats are telling me that Moreno is a good defender. Flanagan is a good defender. I donīt believe "stats are wrong". I never thought the stats were wrong about Downing. I simply thought they were being interpreted wrong with too much emphasis put on the wrong numbers and little to no emphasis put on the right ones. I think the same thing is happening here. But simply saying "that is wrong" is only part of the task, now I want to find how to interpret them correctly which is where the hard work is.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline JCB

  • Sponsors of Digger Barnes.....
  • Legacy Fan
  • ******
  • Posts: 2,021
  • We all Live in a Red and White Kop
Re: Pressing per 90?
« Reply #185 on: July 25, 2017, 03:35:32 pm »
snip..

It will start getting more complicated though once we get to the end. At the moment I am just using raw data but I plan to start making adjustments for things to reflect how playing in a certain team/tactical shape/system/style/league etc can impact on the numbers you produce. For example, Spurs as a team produce 1/2 as many inceptions per game as West Ham. Therefore adjusting Spurs players towards some sort of global average and West Hamīs also should show us how much Cresswellīs strange numbers are due to the fact he simply plays for West Ham.

It should also maybe give us an idea of predicting what sort of numbers Robertson should produce at Liverpool (and others who change clubs during the window) and answer that age old question "But what would he be like in a better team". Then over the course of the season we can see how close our predictions are to what Robertson actually does on the pitch. Then maybe refine what we are doing if our predictions are way off in January.

I also want to try and do this for center backs and look at Lindelof v Van Dijk (assuming he moves) v Keane and try to see if there is a way to predict where they will succeed and fail in a change of system/league/team.


Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: Pressing per 90?
« Reply #186 on: July 25, 2017, 04:00:29 pm »

LOL

one of the problems I have is that I do this while working as a translator and often have my spell & grammar checkers running while proofreading.

I donīt "work" while I am RAWKing and never proofread my own stuff. Therefore my checkers often change stuff as I am writing it and I never even notice. Thanks for that though. Laughed loud!
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline Thats So Dimitar

  • Kopite
  • *****
  • Posts: 608
  • We all Live in a Red and White Kop
Re: Pressing per 90?
« Reply #187 on: July 25, 2017, 05:14:12 pm »

Interestingly Pochetino inherited Walker & Rose. Yet signed Davies & Trippier. That should give us an idea what Pochetino* is looking for in full backs for that position and the differences may show us which weaknesses are acceptable for that position for Pochetino.

* Hopefully our Spurs fan following the thread can chime in here. I have no idea how recruitment works or used to work at the time both were signed. I believe at most clubs the manager has minimal input into transfers other than choosing who he wants to sign. How that choise is arrived at can range from him speaking with a team of scouts to something similar to what Liverpool has - analysts, coaches & scouts forming a committee to discuss and agree on transfers, managers say is final.

Here I am to chime in.

What the previous poster said about our Rose and Davies is basically correct. Poch very rarely adjusts tactics for the opposition team. Most times we have our own tactics or strategy and we will either execute it well enough to win or we will lose. Whether that is good or bad is another matter.

But basically Rose would always play unless there were two games in a week, if there were then both fullback positions were rotated every time, Rose/Davies and Walker/Trippier playing once a week whenever possible. Became less possible once Rose was injured, which led to overplaying Davies which led to the disaster of Son at lwb in the FA cup semi once Davies got injured too.

In terms of what the signing of Trippier and Davies mean for what Poch wants in a fullback, it may not tell you a ton. Until last season transfer dealings were done through what was essentially a transfer "committee" of Paul Mitchell, Levy and Poch. Paul Mitchell also worked with Poch at Southampton but left us before the beginning of the season. So how much of a hand Poch had in their signings and what attributes of theirs he wanted is debatable, one thing I will say though is that Davies and Trippier are both better in possession, but it is not always realised as they don't have the stand out pace of Walker and Rose that catches the eye. Rose and Walker both better defenders out of the two imo, but that may also be a product of teams preferring pacy wingers which the slower Davies and Trippier struggle against.
We beat Real Madrid

Offline JCB

  • Sponsors of Digger Barnes.....
  • Legacy Fan
  • ******
  • Posts: 2,021
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #188 on: July 25, 2017, 05:32:29 pm »
Here is something I cannot get my head around still. There is a correlation between interceptions and recoveries at player level and an inverse correlation at team level :D

We are top in recoveries and 2nd bottom in interceptions.

The way I arrive at this as a possibility is this.

Recoveries = Passes played directly to an opposing player. Therefore they include forced pass errors. Therefore pressing the ball carrier will result in more recoveries as players are rushed to play the ball. Therefore I see recoveries (+ blocked passes) now as a sign of how aggressively you press teams. Whereas interceptions is where a player moves into the line of a pass to intercept. Therefore a side that focuses on shape (Hull) or focuses on pressing passing lanes (Man City) will have higher Interceptions vs Possession than a side like Liverpool who press the players.

However players who read the game or have better reactions will always be the ones who pick up interceptions & recoveries for a team. Therefore this is how we can differentiate between player skill sets and numbers generated by tactics/system. And this is where I have a major problem with how data is used now for players. Manchester United are more passive in how they defend and therefore rack up interceptions, the likes of Spurs and Liverpool do not. Yet this is one of the key categories used on the radars. On top of this, Interceptions is possession adjusted which means the two highest teams in interceptions (Arsenal & Manchester United) get their players numbers increased even further as they are two of the highest in terms of possession.

Basically, tackles & interceptions volumes are used to measure defending at the moment and I see no direct correlation to how good a defender is at those numbers at the moment. For example, before I even adjust for possession my data shows that Moreno and Flanagan are the two best full backs in the Premier League in terms of T+I. Does anybody believe that to be true? At the other end of the scale, Davies, Bertrand & Clyne are in the bottom 5. Does anybody believe that to be true either? How many times have you seen Flanno & Moreno have defensive problems? For me itīs a lot. Yet they are the best two in the league. In contrast how many times have you seen Bertrand, Clyne or Davies struggle? Itīs very rare. So this measurement simply HAS to be wrong. It cannot be as simple as the people who do the least = better either but we can at least agree that tackles + interceptions tells us literally nothing at all other than the players are chasing the ball a lot perhaps. When I word it like that, you can see how it stops being a good thing immediately. Defending isnīt about chasing the ball at all. So why are two metrics that involve you trying to win the ball the only two used as a measurement of how good you are defensively?

This is pretty much why I am doing this. Stats are telling me that Moreno is a good defender. Flanagan is a good defender. I donīt believe "stats are wrong". I never thought the stats were wrong about Downing. I simply thought they were being interpreted wrong with too much emphasis put on the wrong numbers and little to no emphasis put on the right ones. I think the same thing is happening here. But simply saying "that is wrong" is only part of the task, now I want to find how to interpret them correctly which is where the hard work is.

Inspired by this post I've decided to look at Int/Recoveries at League and Team level. It might move away from fullbacks (which we've been isolating so far) but it might go some way to contextualise how the metrics form part of a more cohesive overall strategy and how as BY suggests a low/high metric needs to be considered as part of the overall plan

Here's the League Level data. You can see how Tottenham and Liverpool's high-output pressing game put them well out of the cluster that is the rest of the League.



And here are the team level viz's for Tottenham, Liverpool and Man U. 
You can see how team strategy affects this metric: 






Please remember that all Axis are not the same. But it's the level of the metric within each team that I'm curious about.
Incidentally I had to exclude the goalkeepers Recovery stats from the above team graphs (which tend to be  quite high). I accidentally left Vorm in the Tottenham one.
 
« Last Edit: July 25, 2017, 05:36:32 pm by JCB »

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #189 on: July 25, 2017, 05:45:57 pm »
Inspired by this post I've decided to look at Int/Recoveries at League and Team level. It might move away from fullbacks (which we've been isolating so far) but it might go some way to contextualise how the metrics form part of a more cohesive overall strategy and how as BY suggests a low/high metric needs to be considered as part of the overall plan

Here's the League Level data. You can see how Tottenham and Liverpool's high-output pressing game put them well out of the cluster that is the rest of the League.



And here are the team level viz's for Tottenham, Liverpool and Man U. 
You can see how team strategy affects this metric: 






Please remember that all Axis are not the same. But it's the level of the metric within each team that I'm curious about.
Incidentally I had to exclude the goalkeepers Recovery stats from the above team graphs (which tend to be  quite high). I accidentally left Vorm in the Tottenham one.
 

Interesting fact on that first table, Liverpool & Spurs concede by far the fewest number of shots per game in the Premier League. Which indicates they protect the final 3rd of the pitch best yet are low-average on defensive metrics. So how are they doing this? How is this measured. And are there players crucial to this who arenīt racking up numbers in tackles, interceptions, etc but that are in some other metric that is going unnoticed?

What I am basically wondering, is there a formula using the data we have that will say how well, or not, a team protects is goal? If there is, what numbers are plugging into it and which are not? What is the importance of those numbers? Can we break that formula down on a player level to see how well, or not, they protect the goal?

All the metrics we look at seems to involve the ball, yet protecting space seems far more important. Is there some way to determine thing from the numbers we have and identify players good at it?
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline redmark

  • RAWK Scribe
  • Legacy Fan
  • ******
  • Posts: 21,395
Re: An unbiased look at our full backs
« Reply #190 on: July 25, 2017, 07:03:22 pm »
Interesting fact on that first table, Liverpool & Spurs concede by far the fewest number of shots per game in the Premier League. Which indicates they protect the final 3rd of the pitch best yet are low-average on defensive metrics. So how are they doing this? How is this measured. And are there players crucial to this who arenīt racking up numbers in tackles, interceptions, etc but that are in some other metric that is going unnoticed?

What I am basically wondering, is there a formula using the data we have that will say how well, or not, a team protects is goal? If there is, what numbers are plugging into it and which are not? What is the importance of those numbers? Can we break that formula down on a player level to see how well, or not, they protect the goal?

All the metrics we look at seems to involve the ball, yet protecting space seems far more important. Is there some way to determine thing from the numbers we have and identify players good at it?

Is that quite right? Or it is a seemingly obvious/reasonable conclusion but the wording of which gives us a slight problem?

If 'space' is primary - which would seem to validate 'shape' - then Roy Hodgson would be a football genius. What does the opposition need to create a goalscoring opportunity? The ball - obviously. 'Space', in that they need to be in those spaces on the pitch (the final third) that provide goalscoring opportunities. But don't they also need 'time'? - to spot a run, identify space, pick a pass, prepare a shot.

The 'disruptive' nature of our game (and Spurs) would suggest we give opponents less time to do those things around the box. We may not always protect 'space' particularly well*, and we're certainly not as shape-focused as most teams. Though notably in both our and Spurs maps, you can see that defenders fall to the left (interceptions/shape/space) of an 'average' line, and midfielders to the right (recoveries/press/time).

How do you measure how much time you give (more to the point, don't give) to the opposition? It's probably implied in recoveries vs interceptions, but perhaps 'distance covered' gives us a clue too (as long as it's not entirely aimless), as a measure of 'busy-ness' and disruption to the time the opposition are allowed.


* Is this why many of the goals we concede feel so 'chaotic' - our shape and coverage of space looks terrible, because it's dependent on us pressuring the ball. Infact, it might indicate that the best tactic to employ against us is a 'chaotic' one (the specialism of several lower mid table clubs) - run at us, throw balls into the area or down the channels, overload areas of the pitch and see how the ball falls. We do well against 'good' teams because they're trying to score 'good' goals. We don't give them time to do such things without disruption.

« Last Edit: July 25, 2017, 07:16:00 pm by redmark »
Stop whining : https://spiritofshankly.com/ : https://thefsa.org.uk/join/ : https://reclaimourgame.com/
The focus now should not be on who the owners are, but limits on what owners can do without formal supporter agreement. At all clubs.

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #191 on: July 25, 2017, 07:26:47 pm »
Is that quite right? Or it is a seemingly obvious/reasonable conclusion but the wording of which gives us a slight problem?

If 'space' is primary - which would seem to validate 'shape' - then Roy Hodgson would be a football genius. What does the opposition need to create a goalscoring opportunity? The ball - obviously. 'Space', in that they need to be in those spaces on the pitch (the final third) that provide goalscoring opportunities. But don't they also need 'time'? - to spot a run, identify space, pick a pass, prepare a shot.

The 'disruptive' nature of our game (and Spurs) would suggest we give opponents less time to do those things around the box. We may not always protect 'space' particularly well, and we're certainly not as shape-focused as most teams. Though notably in both our and Spurs maps, you can see that defenders fall to the left (interceptions/shape/space) of an 'average' line, and midfielders to the right (recoveries/press/time).

How do you measure how much time you give (more to the point, don't give) to the opposition? It's probably implied in recoveries vs interceptions, but perhaps 'distance covered' gives us a clue too (as long as it's not entirely aimless), as a measure of 'busy-ness' and disruption to the time the opposition are allowed.

Sorry, yes that was really badly worded on my part. Iīll think on how to word that properly.

However how we defend is about defending space, which we do by attacking the ball carrier.

For example, we play a high line (to compress teams vertically) and we play in a compact shape (compress further vertically and also horizontally). We also have a layered formation (e.g. 4-2-3-1) rather than straight lines. Also even within any formation Klopp uses he does so in quite assymetric ways. One full back will be higher than the other. One CB attacks threats the other covers threats. Our 3 midfielders will all be at different depths on the pitch too. The one place this leaves spaces on the pitch down our sides and behind that people can play into. We encourage the ball wide and then use pressing traps to try and regain possession. E.g. ball to centerback, we press from the front and the side, the ball goes to the full back and we continue to that press from the side and also from the front.

The then canīt go back the way the ball came, so he has a dangerous ball to play over or past the press from that side. He also canīt go forward without a dangerous pass over/past the press from the front. Likewise diagonal. The ball then tends to go backwards (where our forward who is pressing the side then presses the keeper) or they take the least risky but more attacking ball usually down the channel into the area of the full back/dm (hence an explosive full back is important), a high long ball where your DM and CB need to attack (hence why height at DM is important).

As for protecting the space behind, we have two things going on.
1. Offside trap
2. Pressing

One without the other doesnīt work. If you press without a high line in means there is far too much vertical space on the pitch between your lines (we see City using this to draw team out and then crush them between the lines). High line without a press is even more dangerous. For example look at the Vardy chance from Saturday. There is no pressure on the full back, that means he has time and space to play a ball to Vardy. In our system is the full back has no pressure on him it means we are exposed over the top so you have to drop and cover. Klavan instead pushes up for an offside that wonīt be coming. Likewise go back and watch that Crystal Palace 3-3 and you see the same thing. We donīt press the ball and yet play a high line and it costs us a goal.

So while you are trying to win the ball, pressing is more about trying to control the options the carrier has and make them play into areas you want them to play - typically down the sides. There they have only 180 degrees of passing options rather than the 360 degrees in a central area. The touchline counts as a defender in pressing them too. If you want an example of this being used masterfully - the 1-4 vs United. We encouraged balls to Ronaldo where Aurelio, Lucas, Masch, Riera would all be pressing that side of the pitch at different depths making Ronaldo have two options.
1. take on Aurelio (but this resulted in one of Carra, Lucas, Masch stopíng him)
2. play into midfield (but that meant playing around or through Lucas/Masch)
3. Go back (but Riera was shutting that option down).

We effectively used their best player against them in a pressing trap. Everytime the ball went out there that day we looked odds on to get possession back. Itīs a brilliant game to rewatch and also shows pretty well how controlling the space on the pitch can be the key to defending.

Hodgsons problem is that his players maintain shape but donīt control space. Staying in shape means you can be overloaded easily. There is spaces between your lines. Switches of play stretch you out. etc.

Hopefully that makes more sense and explains what I mean when I see that controlling space is more important than trying to win the ball. For me, pressing is more about trying to control where the ball goes and making sure itīs into spaces we have more control over than the opposition rather than as a means to win the ball directly.

« Last Edit: July 25, 2017, 07:31:02 pm by BabuYagu »
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: Pressing per 90?
« Reply #192 on: July 25, 2017, 07:45:12 pm »
Here I am to chime in.

What the previous poster said about our Rose and Davies is basically correct. Poch very rarely adjusts tactics for the opposition team. Most times we have our own tactics or strategy and we will either execute it well enough to win or we will lose. Whether that is good or bad is another matter.

But basically Rose would always play unless there were two games in a week, if there were then both fullback positions were rotated every time, Rose/Davies and Walker/Trippier playing once a week whenever possible. Became less possible once Rose was injured, which led to overplaying Davies which led to the disaster of Son at lwb in the FA cup semi once Davies got injured too.

In terms of what the signing of Trippier and Davies mean for what Poch wants in a fullback, it may not tell you a ton. Until last season transfer dealings were done through what was essentially a transfer "committee" of Paul Mitchell, Levy and Poch. Paul Mitchell also worked with Poch at Southampton but left us before the beginning of the season. So how much of a hand Poch had in their signings and what attributes of theirs he wanted is debatable, one thing I will say though is that Davies and Trippier are both better in possession, but it is not always realised as they don't have the stand out pace of Walker and Rose that catches the eye. Rose and Walker both better defenders out of the two imo, but that may also be a product of teams preferring pacy wingers which the slower Davies and Trippier struggle against.

I look at  Trippier and his crossing and it looks a nightmare to defend. Iīve never looked at his number to see how much it creates in terms of chances, shots, goals etc... but crossing can also be a means to disrupt shape. You cannot allow a player like Trippier to tee off on you like that. Which means the starting position of players on that side starts to drift wider and wider making more chances centrally. So while he lacks Walkers physical abilities I imagine he does a good job against low block sides in stretching them out? I suspect Walker might struggle here as his attacking is more limited. The threat there is pace, but if you donīt allow space vertically for him and have a good layering for defenders to cover if he beats his man you have taken a lot out of his attacking game.

One of the reasons I liked Moreno is he has a lot of attacking things going on that are a problem. Firstly, he runs in diagonals a lot. This is horrible to defend. If someone is running down the side of the pitch they have to go 90 degrees to get past me. If someone is running diagonally at me they only have to turn 45 degrees which means even the slightest of feints and I am fucked. Look at that goal Martial scored against us. He runs diagonally at Skrtel and just turns him inside out. A shit winger would have carried that to the byline and tried to get the ball into the box, or went across the front of the goal to shoot. He also has that early, hard cross that will land between keeper and penalty spot that are horrible to defend. Touch it and itīs an own goal. Donīt and who knows what happens. You cannot look around you because of the pace on the ball. He packs a powerful shot so you cannot just stand off. He loves a 1-2. He makes unselfish runs to pull players out or across. He makes very good and well timed runs in curves behind defences that play high. He is just a nightmare.

However... he is brainless. That totally outweighs everything else unfortunately. A team that has so much tactical structure as ours cannot have someone who doesnīt "get" it. Which makes him somewhat of a liability when he plays. But Moreno has everything to be an elite full back, except the brain of one.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #193 on: July 25, 2017, 08:19:30 pm »
Is that quite right? Or it is a seemingly obvious/reasonable conclusion but the wording of which gives us a slight problem?

If 'space' is primary - which would seem to validate 'shape' - then Roy Hodgson would be a football genius. What does the opposition need to create a goalscoring opportunity? The ball - obviously. 'Space', in that they need to be in those spaces on the pitch (the final third) that provide goalscoring opportunities. But don't they also need 'time'? - to spot a run, identify space, pick a pass, prepare a shot.

The 'disruptive' nature of our game (and Spurs) would suggest we give opponents less time to do those things around the box. We may not always protect 'space' particularly well*, and we're certainly not as shape-focused as most teams. Though notably in both our and Spurs maps,1. you can see that defenders fall to the left (interceptions/shape/space) of an 'average' line, and midfielders to the right (recoveries/press/time).

How do you measure how much time you give (more to the point, don't give) to the opposition? It's probably implied in recoveries vs interceptions, but perhaps 'distance covered' gives us a clue too (as long as it's not entirely aimless), as a measure of 'busy-ness' and disruption to the time the opposition are allowed.


2. * Is this why many of the goals we concede feel so 'chaotic' - our shape and coverage of space looks terrible, because it's dependent on us pressuring the ball. Infact, it might indicate that the best tactic to employ against us is a 'chaotic' one (the specialism of several lower mid table clubs) - run at us, throw balls into the area or down the channels, overload areas of the pitch and see how the ball falls. We do well against 'good' teams because they're trying to score 'good' goals. We don't give them time to do such things without disruption.

1. That might be down to pressing traps. I would need to compare that map to the actual games to see but it feels that way. It makes sense though. The defence slightly left midfield slightly right creates a natural hole for people to play into.  One on which we can ambush the carrier from all sides, pull his pants down and take the ball away.

2. Yes, taht is exactly it to be honest. We can do a lot better here. Our clearing headers often arenīt clearing. Our keeping doesnīt doesnīt do enough to prevent shots. We donīt win enough second balls etc. But itīs simply mathematics. If the ball is in the danger areas and you have 4 players in there to our 8, there is a 1/3 chance that a loose ball will fall to you. They arenīt trying to create goals against us, just beat us down with probability. Then there is the fact that I am not sure our shot stopping is anywhere good enough which means the limited amount of shots we concede result in a much high ratio of goals than we should reasonably expect.

The interesting thing is we might feel that we have one of the solutions here already in Karius. is shot stopping is outstanding. If his general goal keeping level could get up to just league average levels then he would be a beast for us. VVD dominates the air like nothing I have seen for a while and his reading of the game is top notch too. He would be our Hyypia but with pace. And so he fucking should be for 300m or whatever we end up spending :D Then you have Keita, the guy who hoovers up second balls.

There is  alot we can do better to be honest. Our pressing doesnīt look right. Iīm not sure what it is but I would watch Dortmund and if you pause the game ever time a player is taking his first touch against them, you struggle to find a dangerous ball to play. They never really let space open up. We still seem easy to play from by comparison. I donīt know if this metric is available but I would love to know the number of passes per shot against teams. If you eliminate instances sides go direct against you and set pieces, it should give you an idea how hard a side is to play though. I imagine we are probably the hardest in the league at the moment but there are moment where we donīt get it right and are very open. I never saw that at Dortmund. It was much more machine like. Same as that Rafa Valencia side. There was never any space to play through them. This is where coaching is far more important than transfers. That all comes from repetition and every time you replace a part it can be like starting again. Itīs also why I am quite calm about transfers. I can see how the two big names we want would be huge, I can also see how purely because our squad is still together we will be a lot harder to break down, attacking patterns better drilled and everyone just 1 more year closer to their physical peak.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline redmark

  • RAWK Scribe
  • Legacy Fan
  • ******
  • Posts: 21,395
Re: An unbiased look at our full backs
« Reply #194 on: July 25, 2017, 11:32:37 pm »
Hopefully that makes more sense and explains what I mean when I see that controlling space is more important than trying to win the ball. For me, pressing is more about trying to control where the ball goes and making sure itīs into spaces we have more control over than the opposition rather than as a means to win the ball directly.

Thanks, definitely clarifies what you were talking about in the previous post and given me a few more things to pay more attention to :).
Stop whining : https://spiritofshankly.com/ : https://thefsa.org.uk/join/ : https://reclaimourgame.com/
The focus now should not be on who the owners are, but limits on what owners can do without formal supporter agreement. At all clubs.

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #195 on: July 26, 2017, 02:10:24 am »
Thanks, definitely clarifies what you were talking about in the previous post and given me a few more things to pay more attention to :).

Btw - another mechanism to look for is transtions. The wider your shape in your attacking phase then the longer it takes to transition back into your defensive phase. Therefore the longer you are vulnerable. One aim of gegenpressing is to remain compact so that our transitions are as quick as possible. In addition, if you are compact and win the ball in midfield you should have immediate overloads all over the pitch. I found this diagram of a dortmund press to demonstrate.



As you can see, we win that ball back and we have a 3 v 2 overload on the wing. Another 3 v 2 overload on the half space between RB and RCB. Plus in the immediate panic of winning the ball back and the LCB trying to move over to deal with those overloads we end up with Blaszcyzkowski available for a half space to half space switch and having an immediate 1-v-1 opportunity for a clear goal scoring chance.

Also with this diagram, look where the space is to attack Dortmund and look how hard it is to get the ball there. The term cover shadow means the area behind you (as in your shadow) where you are preventing the ball being safely played into. Literally the only "safe" pass here is a chipped pass to the goalkeeper. That doesnīt sound safe though and with Lewandowski lurking with intent to commit harm, I wouldnīt recommend it. Therefore the real safe option is to play the ball into the area either in front of or behind hummels and hope the forward can challenge and win it there. Basically, percentage balls.

Another interesting thing Klopp does is, imagine for example the ball ends up  going between the two blue forwards and swept up by Subotic. He can either recycle possession back or wide to Piszczek. Alternatively he can just drill a pass back into the exact area it came from knowing that the odds are favourable for us to either get to the pass or a blue to control it with an immediate counter press right on him again. Basically forcing transitions in favourable situations.

Lastly, hopefully that diagram above shows you why straight lines of formation and shape doesnīt effectively control space. That blue team is controlling nothing right there. Even if he somehow manages to turn someone and escape that immediate press, where would he go? Instead imagine that blue team with 5 lines of depth instead of just 3 and you can see how itīs much harder to press as with that many lines of depth, the center midfielder have a direct line to every player in his team rather than having them stacked up the way the LCM and LM are here.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline Flinstone

  • Whatever you like. Would like to slip it in without making it awkward... Chinese Information Minister
  • Legacy Fan
  • ******
  • Posts: 3,747
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #196 on: July 26, 2017, 02:35:12 am »
Why can't you just break that press by chipping the ball in the space b/w the keeper and the centre-backs and use rapid forwards to get it?
The West is finished, China is the future

Offline Flinstone

  • Whatever you like. Would like to slip it in without making it awkward... Chinese Information Minister
  • Legacy Fan
  • ******
  • Posts: 3,747
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #197 on: July 26, 2017, 02:36:43 am »
A sweeper keeper would solve that but was Weidenfeller like that? Mignolet certainly isn't
The West is finished, China is the future

Offline BabuYagu

  • It's Portuguese for 'BabyYoghurt'. The John Motson of RAWK. Or Barry Davies. Or Charley Boorman, even. Expertly silent fist-pumper. Needs to pay more attention. Repeatly analing goalkeepers.
  • Legacy Fan
  • ******
  • Posts: 17,350
  • wakelet.com/@BabuYagu
    • Wakelet of the Articles I have written
Re: An unbiased look at our full backs
« Reply #198 on: July 26, 2017, 03:50:17 am »
Why can't you just break that press by chipping the ball in the space b/w the keeper and the centre-backs and use rapid forwards to get it?

Because look at the position of the ball, the carriers body and Sebastien Kehl. He need to turn and face Kehl to play that ball. From his position now, he can see and play into the area in front or behind Hummels.

However chipped through balls over the time while being closed down are actually very hard to play. It becomes more a clearance than a pass. Unpressured is another thing (see Vardyīs chance the other day).

As for your keeper question, Weidenfeller wasnīt really. Karius is.
My first article on Anfield Index on Shaqiri. Enjoy. bit.ly/2mAq3Qd

Offline Flinstone

  • Whatever you like. Would like to slip it in without making it awkward... Chinese Information Minister
  • Legacy Fan
  • ******
  • Posts: 3,747
  • We all Live in a Red and White Kop
Re: An unbiased look at our full backs
« Reply #199 on: July 26, 2017, 06:45:22 am »
Because look at the position of the ball, the carriers body and Sebastien Kehl. He need to turn and face Kehl to play that ball. From his position now, he can see and play into the area in front or behind Hummels.



Fuck, sweet spot. Did not even notice that

Thanks for the explanation
The West is finished, China is the future