Topic: *Internal Spyin' Kop - Dan Kennett's Dashboard - 2013-14 in one infographic

[royhendo]

Edit - the final infographic for the entire season is attached to the o.p. It neatly encapsulates the progress, and shows where we've room to grow as clearly as is possible really.

-


Dan Kennett uses stats well.

Take a second to digest that sentence. Everyone with a passing interest in the modern game uses stats, but very few use stats well.

You can follow Dan on twitter here: http://www.twitter.com/dankennett - do that alone and if your head's screwed on you'll get more insight from that timeline of bite-sized chunks than you would from the rest of the collected football media combined. If, on the other hand, you can't see what he's driving at, you might miss the points he's making.

This season, Dan's published a dashboard for us all to consume, free of charge, that measures Liverpool's performance against a number of key metrics. The metrics are listed in the following posts, and it'd be nice to see your thoughts on their significance.

You can find Dan's Dashboard here: http://public.tableausoftware.com/views/LFC_2012_13_Perf_Dashboard/LFC

Key
Red lines = LFC this season
Gold lines = LFC last season
Purple lines = Opposition performance
Grey shading = Top 4 to Champions over preceding 2 or 4 seasons

[royhendo]

1. Final Third Pass Completion

So what does this graph represent. Well, divide the pitch into thirds, and measure the passes we attempt and complete in that final third. There tends to be a correlation between being at the top of the table and doing this well (see Simon Kuper's article on Manchester City and stats in football here).

Gavin Flieg, Head Of Performance Analysis at Manchester City: “The top four teams consistently have a higher percentage of pass completion in the final third of the pitch. Since the recruitment of Carlos Tévez, David Silva, Adam Johnson and Yaya Touré to our football team, in the last six months alone, our ability to keep the ball in the final third has grown by 7.7 per cent.”

It's an important metric.

Over the last 2 seasons, here's how the best teams have done.

76.6% <- Champions
73.3% <- Top 4 teams

This season.

73.3% <- Us

The first graph shows how it's fluctuated game-to-game (the 'target range' is shown as a grey band), while the second shows it represented as a scatter plot.

[royhendo]

2. Total Chances Created and Clear Cut Chances

So we're keeping the ball in the final third roughly the way a top four side keeps the ball in the final third. How are we doing once we get the ball in there? Are we creating chances? And if we are, are those chances clear cut? And how do we measure up against the numbers of the sides who are already where we want to be?

The first graph shows the total number of chances we've created in each game, week to week, again with the benchmark figures shown as a grey band, and average shown in a line (I'm colour blind - is it gold?) horizontally across the middle.

Last year's benchmarks for chances per game.

14.7 <- Champions
13.9 <- Top 4 sides

This season to date.

13 <- Us


Last season.

10.8 <- Us


The second graph shows "Clear Chance Conversion".

The first thing that jumps out at you is that our opponents have been converting more of their clear cut chances than we have. We've also been allowing our opponents to convert more clear cut chances than the 'top four to champs' band.

The benchmark figures for clear cut chances created over the last two years.

48.5% <- Champions
38.1% <- Top four
29.7% <- Us

This season.

33.3% <- Us

So it's trending up, and we're now above where we were for our average last season, but we make life too easy for our opponents at the other end.


The third graph, "Clear Cut Chances", is a little more 'busy'. No benchmark figures are included in this graph. Above the '0' line, you have our figures. Below the '0' line, you have the figures for each side we've played (the clear cut chances we've given up).

For the two teams, the graph shows two numbers: clear cut chances created (red dots), and clear cut chances scored (gold dots). So there are either one or two dots shown on each line, because if you score the only clear cut chance you create, or if you don't create any full stop, you'll just see one dot.

The "W", "L" and "D" mean win, lose, and draw, obviously.

[royhendo]

3. Save Percentage

This one measures the percentage of shots on target that are saved.

The grey line in this one shows Liverpool's average from 2008/09 to 2010/11. That's the benchmark.

74.8% <- us in the recent past


This season.

58.6% <- us
72.9% <- our opponents

Note that ours is trending up. The other side's is pretty steady, and too high.

[royhendo]

4. Penalty Box Conversion

The first graph is about penalty box shot conversion.

14.2% <- Our average over four years


Last two seasons (again, the shaded area).

17.5% <- Champions
16% <- Top four
8.5% <- Us last season

This season.

8.3% <- Us this season
19.7% <- our opponents

So it's not been great. But the slopes of the lines tell a story. Week on week, the opponents' average figure gets lower, whilst ours inches higher. We need those two trends to continue until the lines criss cross and beyond.


The second graph shows the number of attempts in the penalty box it takes to score a goal.

4 year average.

7 <- Us


Past 4 seasons.

5.7 <- Champions
6.3 <- Top four


Last season.

11.8 <- Us


This season.

12.11 <- Us
5.08 <- Our opponents

So again, this is bad. Again, it's gently trending the right way, but it needs to improve.

[royhendo]

5. Total Defensive Errors Per Match

This one's pretty clear.

4 year Premier League average.

0.85 <- Us

This season's average so far.

2.182 <- Us

Too many.


Aside from Dan.
The errors line chart tracks total errors in the game by LFC and their opponent against the 4 year average. The breakdown for us is in the circles chart. Basically you want to be making 1 error per 3 games.

[royhendo]

6. Punishing Defensive Errors

This is another 'busy' graph, representing both our and our opponents' tendency to make defensive errors. The red dots show the number of errors. The gold dots show the number of errors that led to conceding goals. You basically want a row of gold zeros along the '0' line for this one.

[royhendo]

7. Match Share

The last graph shows the 'Match Share' figures. For all the dots, you want the numbers to be above the 50% line, because that suggests that you've been in control, or maybe even dominant.

Note - I used the word 'suggests', because the stats used by people like Gavin Fleig show that there's a correlation between strong numbers here and strong league position.

The dots represent the percentages 'won'.

Orange - ground duels
Blue - aerial duels
Green - possession

[royhendo]

Note.
The stats are drawn from the standard Opta dataset (via the tremendous site www.eplindex.com). There are a few boo-boos in there (how they've counted errors and clear cut chances - sometimes there are grey areas), but it's the standard dataset that many of the clubs use to drive their analysis.

-


So some good and some bad. But some good trends, and some obvious issues that we must address in January. We all knew it, didn't we?

Excerpts from Dan's twitter timeline on the subject make for interesting reading.

"you could conclude that we are actually in the middle of the turnaround now".

"Opponents PBC and CCC also at season lows. All relative though and still way below target."

"Penalty Box Conversion, Clear Chance Conversion and Save % now all at season high's. But still too many errors"

"lots of trends in the right direction"

-

Over to you lot. What do you make of the hard numbers?

[No666]

The hard numbers provide a frame on which to hang the beautifully educated observation. It's interesting to read these stats and remember this:

http://www.redandwhitekop.com/forum/index.php?topic=279444.msg11052401#msg11052401

[BabuYagu]

Thanks for this Roy. I think this is the best way to measure progress using data sets like this. I remember someone (Pako?) once said something along the lines of that you need to first coach into the players the right ideas to increase your chances of winning more often than not. Only after this can results improve.

If I am honest, we are pretty much where I thought we would be. Rodgers was not taking a short term, results approach to this job. Had he, then I am sure we would be in a better league position but perhaps would be near to our ceiling on what we could achieve. Instead he has sacrificed our results to give us a sturdy foundation to work from going forward. These stats indicate that we are at that point I referred to where lessons have been learned and we should now start to expect some results from this. Dan seems to think the same "you could conclude that we are actually in the middle of the turnaround now".

The big problem in achieving the results is that we have just 1 player who can finish off the moves. That will sorely hurt our ability to achieve results. Could be a long wait until January.

[royhendo]

Bump, just to flag up that this is here. Keep your agendas out of it please and focus on the numbers and their significance or otherwise.

[[new username under construction]]

Didn't someone say that the techniques etc would kick in around xmas and we'd start seeing the "real" Liverpool then?

[rafathegaffa83]

Although I'm sure the anti-statistics brigade will have their say, I think it pretty much confirms some basic issues: we make too many defensive errors, our opponents are gifted goals with minimal opportunity and to counteract that we don't score enough. I would assume the first two are partially the product of playing a slightly different defensive system, including playing a higher line and rotating in some younger players at the back. It's a work in progress, but it's encouraging to see things are improving in some areas, while work clearly needs to be done eliminating other issues.

[Funky_Gibbons]

Some interesting work here.

So we’re keeping the ball in the opposition third as much as any of the top four teams but we are creating fewer chances and need more chances to score. Is the system to blame for the lack of chances created or is it the players?

We know that Rodger’s likes to keep possession of the ball and search for defensive weaknesses so are we passing too much or do we simply not have players with sufficient creativity in the final third.

The chance conversion is no surprise really and we have all been saying for a long time now that we need to add some more quality to our front line. This, coupled with our mistakes at the back, is a shocking combination when it comes to winning games.

It will be interesting to see how these stats ‘settle’ once we have played a few more teams around the middle/bottom of the league as in our first 11-games we have played 6 of the 7 teams who finished above us last season. These games can be deemed ‘tougher’ and may distort these initial findings as chance creation will be harder.
 

[Guz-kop]

Very interesting.

Would like to know what's define as a clear cut chance though? Are we saving more shots now because our shot stopping is improving or are we simply giving our opponents less clear cut chances?

It did appear to me last year that some games we'd have lots of the ball and look like we're controlling things but
a) We were very vulnerable when teams ran at us and looked short on numbers but
b) At the same time we weren't getting enough men into the box

I feel we're still plagued by this problems. We don't make enough clear cut chances sometimes but give them away at the other end and put our defence under a lot of pressure. Why is this? Where are our players positioned on the pitch at key moments when the ball breaks as a chance for us or our opponents?

The final set of stats are interesting. It'd be interesting to know how many times in a game we get the ball into the box with a certain number of players in the box and how that compares with other teams.

[Eeyore]

Some interesting work here.

So we’re keeping the ball in the opposition third as much as any of the top four teams but we are creating fewer chances and need more chances to score. Is the system to blame for the lack of chances created or is it the players?

We know that Rodger’s likes to keep possession of the ball and search for defensive weaknesses so are we passing too much or do we simply not have players with sufficient creativity in the final third.

The chance conversion is no surprise really and we have all been saying for a long time now that we need to add some more quality to our front line. This, coupled with our mistakes at the back, is a shocking combination when it comes to winning games.

It will be interesting to see how these stats ‘settle’ once we have played a few more teams around the middle/bottom of the league as in our first 11-games we have played 6 of the 7 teams who finished above us last season. These games can be deemed ‘tougher’ and may distort these initial findings as chance creation will be harder.
 

The matches where we have completed the most final third passes have been against teams that have dropped off and defended their penalty area. Passing the ball infront of a team 35 yards from goal is counted as a final third pass, a more useful stat for me would be how many times we have actually got behind teams. Park the bus teams let you into their final third but stop you getting in behind them. Completing thirty passes infront of team is nowhere near as productive as completing three passes when you get in behind teams.

The difference between a square ball across the six yard box and a square ball 35 yards out or the difference between the kind of final third passes United make and the kind of passes Arsenal make. United make final third passes against 3 or 4 defenders Arsenal tend to make final third passes against ten defenders.

If you manage the transitions better you make less final third passes but make more clear chances. The vast majority of our league goals have come after less than three or four touches.

[royhendo]

Didn't someone say that the techniques etc would kick in around xmas and we'd start seeing the "real" Liverpool then?

You would expect the metrics to start to back that up wouldn't you? Worth keeping tabs on.

[Aristotle]

Quite an interesting article I stumbled upon recently, it's posted down there for the German speaking population (and those brave enough to face internet translations). It was an interview with Jürgen Klopp, mainly about himself but also a brief inside as to how a top manager views statistics. I found it quite interesting that Klopp basically said statistics only confirm what he already knew and that having statistics helps mainly to keep things fresh when his memory fails. But what I found most interesting is that they focus much on their own game as Dortmund's top scout is in the stands, analyzing the game and drawing up how the first 45 minutes went and any following team talk and tactical adresses are made as a result of that. He's very old school (I like that about him personally) but, and this is something that I think really sets the Germans apart, is the emphasis on physical condition and health. To him the most important stats are the ones relating to a players ability to perform, not how he performs in each game.
Training every day with the ball is utter "Quatsch". You can not play the game unless you run, run, run. More stamina = more concentration over the duration of the game. The tempo of the game doesn't change, just because the player is getting tired.

Just think it was an interesting aspect of the game since we went through so much with Brukner and Burgess who are now gone and Rodgers has brought in his own people. Of course these are statistics that never reach public eye, due to the fact that you don't want to let the opposition know who to isolate and target. Let them do their own homework there. But if purely out of curiosity I'd love to see how we handle the metrics of physical condition in our squad.

And before I forget, the article sauce: http://www.tagesspiegel.de/sport/juergen-klopp-im-interview-ich-sehe-das-leben-als-geschenk/7015154.html

[Funky_Gibbons]

The matches where we have completed the most final third passes have been against teams that have dropped off and defended their penalty area. Passing the ball infront of a team 35 yards from goal is counted as a final third pass, a more useful stat for me would be how many times we have actually got behind teams. Park the bus teams let you into their final third but stop you getting in behind them. Completing thirty passes infront of team is nowhere near as productive as completing three passes when you get in behind teams.

The difference between a square ball across the six yard box and a square ball 35 yards out or the difference between the kind of final third passes United make and the kind of passes Arsenal make. United make final third passes against 3 or 4 defenders Arsenal tend to make final third passes against ten defenders.

If you manage the transitions better you make less final third passes but make more clear chances. The vast majority of our league goals have come after less than three or four touches.

Doesn’t the second (scatter graph?) in the ‘final third pass completion’ section show that the only game in which we never had passing statistics comparable with the top four was Everton? I agree with you that the percentage is higher against the teams that have dropped off (Stoke being the obvious example) but we have managed to maintain a high level of possession in the final third against different opposition in all but one game.

I don’t think we have been completely tied down to the whole pass, pass, pass philosophy and have tried mixing it up at times. Against Newcastle we attempted a few counter attacks and should have scored from one where Allen and Suarez set-up Sterling. We have also hit it long a few times to good affect such as the Norwich and Newcastle games.     

[TitanTrigger]

The matches where we have completed the most final third passes have been against teams that have dropped off and defended their penalty area. Passing the ball infront of a team 35 yards from goal is counted as a final third pass, a more useful stat for me would be how many times we have actually got behind teams. Park the bus teams let you into their final third but stop you getting in behind them. Completing thirty passes infront of team is nowhere near as productive as completing three passes when you get in behind teams.

The difference between a square ball across the six yard box and a square ball 35 yards out or the difference between the kind of final third passes United make and the kind of passes Arsenal make. United make final third passes against 3 or 4 defenders Arsenal tend to make final third passes against ten defenders.

If you manage the transitions better you make less final third passes but make more clear chances. The vast majority of our league goals have come after less than three or four touches.

That's spot on, I wonder if it would be possible to pull the average number of defenders in the final third while the ball is in the final third from the opta dataset.

It would probably give a good indication if we or the opposition are getting behind the midfield via quick transition.

[royhendo]

Al - the notion you're putting forward there re teams that retreat. Everton is our only outlier that isn't at top four level.

[PhaseOfPlay]

The matches where we have completed the most final third passes have been against teams that have dropped off and defended their penalty area. Passing the ball infront of a team 35 yards from goal is counted as a final third pass, a more useful stat for me would be how many times we have actually got behind teams. Park the bus teams let you into their final third but stop you getting in behind them. Completing thirty passes infront of team is nowhere near as productive as completing three passes when you get in behind teams.

The difference between a square ball across the six yard box and a square ball 35 yards out or the difference between the kind of final third passes United make and the kind of passes Arsenal make. United make final third passes against 3 or 4 defenders Arsenal tend to make final third passes against ten defenders.

If you manage the transitions better you make less final third passes but make more clear chances. The vast majority of our league goals have come after less than three or four touches.

The vast majority of all goals are like this though. It doesn't matter whether you are Barca or Stoke. Most goals are scored in less touches and less passes.

[PhaseOfPlay]

That's spot on, I wonder if it would be possible to pull the average number of defenders in the final third while the ball is in the final third from the opta dataset.

It would probably give a good indication if we or the opposition are getting behind the midfield via quick transition.

You could probably do a very rough correlation between average positions and percentage of time in the attacking third. It would be very limited from this though

[royhendo]

That's spot on, I wonder if it would be possible to pull the average number of defenders in the final third while the ball is in the final third from the opta dataset.

It would probably give a good indication if we or the opposition are getting behind the midfield via quick transition.

Very good question that.

[PhaseOfPlay]

Doesn’t the second (scatter graph?) in the ‘final third pass completion’ section show that the only game in which we never had passing statistics comparable with the top four was Everton? I agree with you that the percentage is higher against the teams that have dropped off (Stoke being the obvious example) but we have managed to maintain a high level of possession in the final third against different opposition in all but one game.

I don’t think we have been completely tied down to the whole pass, pass, pass philosophy and have tried mixing it up at times. Against Newcastle we attempted a few counter attacks and should have scored from one where Allen and Suarez set-up Sterling. We have also hit it long a few times to good affect such as the Norwich and Newcastle games.   

The games where we didn't dominate time of possession in the attacking third were Norwich, Everton and Chelsea. Two of these games are also games where we pulled away with a lead early on. WBA was an even game, as was Reading. The rest, though, were dominant positions for Liverpool, including United, Newcastle and Arsenal. So we do dominate possession in the final third, but as Al says, we maybe don't dominate in penetrating passes. However, when a team sits in its own box and defends a zone, penetrating passes aren't easy to come by, and the most effective weapons are dribbling and long shots. Luis excels at one, but doesn't convert, and I don't really remember anyone being dominant from our midfield for shots from outside the box.

[Vulmea]

interpretation is the key I guess

Al makes a good point about final 3rd passing. Its akin to what the context of the game is. Park the bus team would happily conceded possession even in the final third provided its not threatening  and likewise you'd expect teams that are ahead to be content keeping teams at arms length. It would mitigate against teams that hit the byline and cross cross the ball rather than tiki taka and is it just saying of those passes made how many are completed rather than how long you spend in the final 3rd?

the classic misinterpretation was the long ball school which derailed much of english football for a decade or more

Some of the bench marks are against last season - that was an epic year but it set records for poor conversion and hits on the woodwork.  I'm not sure last year is a fair comparitor even taken across 38 games

One of the joys of football is just how dynamic it is drawing conclusions from one game to another is difficult. General trends may be easier but how mucgh of comparing against the top 4 is self fulfilling. Do these things make a top four team or have top four teams simply played that way .........

The physicist Werner Heisenberg developed the Heisenberg Uncertainty Principle, which suggests the more precisely you measure the momentum of a particle the less precise your measurement of its position. Again, in Heisenberg's interpretation this wasn't just a measurement error or technological limitation, but an actual physical limit.

potentially that could apply to any system and therfore football ' the more precisely you try to define something the less precise its value will be'

dont get me wrong as alluded to above with Klopp stats are often useful to check your interpretation and challenge your own ideas. I'm just not sure how much value can be placed on the raw data. The important thing is the intelligence thats applied to them  and the intelligence which identifies what needs to be measured in the first place, why these stats?

if the germans are suggesting the physical ability of players is key then if that is more important, what if statistically top 4 teams have a +10% stamina is that detrmining factor in winning, does that give you more final 3rd passes. more wins

some form of pattern recognition thrown at Opta may throw up some whacky connections coaches have never dreamed of but its all down to whats caaptured - if you dont look for how many spints a player makes in a game it'll never show up as a key factor. I guess I'm saying we dont know what we dont know. Football is harder to analyse than a simple if interesting set of numbers

the other key thing is you get what you measure. If we are looking at things like this and measuring ourselves against them they will improve - that is the way of human nature. Set a target to complete 90% of jobs on time and you will, you may well throw more resource or more money at it but you'll get it in. Unless all you are interested in is time then the measurement isn't good enough because you may well be introducing waste in cost and resource to meet a deadline. The suggestion was this is exactly what happened last year with the purchase of 'chance creators' rather than players that could play for LFC.

the other basic human trait is that we are lazy bastards, give us a number and we think we know what it means, we dont necessarily look further

dangerous things them stats. I know that comes across as a bit why botherish, thats not really the point merely that you can see what you want to see in those stats, reasons for optiism I'm not convinced by them.  I still think its whats between the players ears that the biggest determinant of success or failure.

[royhendo]

But your reasons for worrying are there in the graphs, clear as day. The numbers are neutral. They should be used to illustrate.

[Eeyore]

Al - the notion you're putting forward there re teams that retreat. Everton is our only outlier that isn't at top four level.

I asked if we could park agendas. Please do.

We set out to retain possession Roy, a hell of a lot of that possession is in front of the opposition and skews the stats regarding penetration. You might think Allen, Wisdom, Sahin and Gerrard retaining possession 35 yards out in front of the opposition is great and it looks great in the stats but how many completed passes in the final third have we actually scored from. If you have one player in the box and six or seven outside the box then you can keep possession with ease but cannot penetrate, that for me is where we are at.

In comparison how many completed final third passes where there in the goals against Reading, Swansea, Chelsea, Everton and Newcastle. Games are won by taking advantage of 1v1's, 2v2's, 3v3's, 4v4's etc and not passes in front of teams. Have a look at how many final third passes Mourinho's Inter conceded or how many final third passes Di Matteo's Chelsea conceded on the way to their respective CL wins.

[Eeyore]

But your reasons for worrying are there in the graphs, clear as day. The numbers are neutral. They should be used to illustrate.

The numbers are far from neutral when you set out to retain safe possession in front of teams. The numbers should be as a result of how you play we are going about it back to front we are playing by numbers instead of doing the right things and letting the numbers take care of themselves.

[PhaseOfPlay]

We set out to retain possession Roy, a hell of a lot of that possession is in front of the opposition and skews the stats regarding penetration. You might think Allen, Wisdom, Sahin and Gerrard retaining possession 35 yards out in front of the opposition is great and it looks great in the stats but how many completed passes in the final third have we actually scored from. If you have one player in the box and six or seven outside the box then you can keep possession with ease but cannot penetrate, that for me is where we are at.

In comparison how many completed final third passes where there in the goals against Reading, Swansea, Chelsea, Everton and Newcastle. Games are won by taking advantage of 1v1's, 2v2's, 3v3's, 4v4's etc and not passes in front of teams. Have a look at how many final third passes Mourinho's Inter conceded or how many final third passes Di Matteo's Chelsea conceded on the way to their respective CL wins.

All correct. But one of the conclusions we can draw then is that in terms of personnel, we don't have enough players willing to take a shot. Although it's not difficult to conclude that BR wants players to assess the risk of each pass and dribble, all coaches will happily take the risk of losing possession from a shot in return for the reward of a goal. Perhaps it is a case of analysis by paralysis from the players, because they are so eager to adhere to the system that they look for higher reward/lower risk shots. This would explain why BR sees Shelvey as important, because he is willing to take a risk on a shot in return for the possibility of scoring. The build-up isn't the problem, and the defending isn't even much of an issue, because we have the same defensive record as the current league leaders. What is a problem is getting a variety of shots on target, from a variety of sources. I daresay that if we improve this part, the trend will be dramatically upward in terms of results

[Draex]

Quite an interesting article I stumbled upon recently, it's posted down there for the German speaking population (and those brave enough to face internet translations). It was an interview with Jürgen Klopp, mainly about himself but also a brief inside as to how a top manager views statistics. I found it quite interesting that Klopp basically said statistics only confirm what he already knew and that having statistics helps mainly to keep things fresh when his memory fails. But what I found most interesting is that they focus much on their own game as Dortmund's top scout is in the stands, analyzing the game and drawing up how the first 45 minutes went and any following team talk and tactical adresses are made as a result of that. He's very old school (I like that about him personally) but, and this is something that I think really sets the Germans apart, is the emphasis on physical condition and health. To him the most important stats are the ones relating to a players ability to perform, not how he performs in each game.
Training every day with the ball is utter "Quatsch". You can not play the game unless you run, run, run. More stamina = more concentration over the duration of the game. The tempo of the game doesn't change, just because the player is getting tired.

Just think it was an interesting aspect of the game since we went through so much with Brukner and Burgess who are now gone and Rodgers has brought in his own people. Of course these are statistics that never reach public eye, due to the fact that you don't want to let the opposition know who to isolate and target. Let them do their own homework there. But if purely out of curiosity I'd love to see how we handle the metrics of physical condition in our squad.

And before I forget, the article sauce: http://www.tagesspiegel.de/sport/juergen-klopp-im-interview-ich-sehe-das-leben-als-geschenk/7015154.html

Last year Swansea's team were extremely fit and lasted a full 90minutes as an 11 more often than not - When we played them we looked knackered after 60mins and they we're still fizzing around.  It's a big juggling act to combine the long distance stamina with the requirements of burst power/pace, I think Stevie for example has looked far better nearing the end of a game this season, but how much is that down to being injury free? Let's not forget, it's far easier to retain your fitness by leading games and controlling them.

About the statistics, they pretty much reinforce what the optimistic fan is trying to argue - we are improving, but it's a slow process. We've changed from Rafa's strict 4231, to Roy's 442 hoofball, to Kenny's mishmash/expressive football to Rodgers disciplined possession based game. That's 4 distinc styles in little over 4 years - compare to Barcelona who have the system ingrained in their DNA, it is going to take time, extensive drilling/repetiveness to stop the players defaulting to an old style and thus breaking up the gameplan. It's no wonder our youth players fit pretty seemlessly into the 1st 11 squad, they've been playing this way (albeit a slightly varied 4231 version) since young and it shows.

I can certainly see we are improving - remove those mistakes (a result of a foreign system) and we would of beat City for example.. Once the system is like second nature and the mistakes are brought to a minimal level, we'll turn the draws to wins and losses to draws..

[royhendo]

Al. The numbers back your point. Stop being defensive and simply use the chances created and converted stats alongside possession.

Please can we just use the numbers rather than being defensive and polarising what should be a useful debate. We get to the opposing team's third, but we don't penetrate well enough to create clear cut chances, and we don't convert enough of the ones we do create.

Meanwhile we make the opposing teams' numbers look like top four performance which means we make life too easy for them.

[BreakfastPercy]

Thanks very much for this Roy. I follow Dan Kennett on Twitter and I'm not ashamed to say you've helped explain a few points there better than I could in my own head, especially the graphs less commonly used (above and beyond a bar chart).

Did you read Richard Whittall's 'Soccer Analytics' article as retweeted by Rory Smith?
http://blogs.thescore.com/footyblog/2012/11/07/special-state-of-analytics-this-machine-kills-pundits/

Richard Whittall goes along with Aristotle's Klopp article (going on Aristotle's own summary, will read the translation later) in the suggestion of statistics as a means to either affirm or contradict subjective ideas, and ultimately as a 'pundit killer'. The issue with that line of thought (Klopp's not Aristotle's) is that there is always a danger that the data is never truly 'neutral' because you only really pay attention to the data you are looking for.

These metrics do have the potential to pokes holes in opinions, and I'm certainly glad the data doesn't disprove the idea we are performing well and improving at the moment, but I'm hoping its value will lie somewhere between 'pundit killer' and 'winning formula' in the long run.

It'll be interesting to plot the line graphs as a curve at the end of the season to see when and at what rate our performance has changed. There's obviously the school of thought that says certain metrics perpetuate each other (converted chances lead to a position of confidence, which leads to more chances created as the opposing team chases the game) and therefore several graphs may share a steep upturn or downturn. The classic striker's 'scoring run' interpreted throughout the team.

[Warks Moustache]

This is brilliant Roy, thanks for sharing. Quick question, are chance conversion and penalty box conversion numbers overlapping metrics?

[Eeyore]

Al. The numbers back your point. Stop being defensive and simply use the chances created and converted stats alongside possession.

Please can we just use the numbers rather than being defensive and polarising what should be a useful debate. We get to the opposing team's third, but we don't penetrate well enough to create clear cut chances, and we don't convert enough of the ones we do create.

Meanwhile we make the opposing teams' numbers look like top four performance which means we make life too easy for them.

Apologies if I come across defensive Roy, it's just managers looking to affect one stat pisses me off it lead to the prevalence of Pomo and the destruction of the English game, it lead to Brazil trying to play like a European side and it leads to neanderthals like Allardyce and Pullis.

After Comollis chance creation catastrophe we need to look a bit deeper and as Aristotle said judge the game with your eyes.

[Vulmea]

But your reasons for worrying are there in the graphs, clear as day. The numbers are neutral. They should be used to illustrate.

they aren't neutral because they've been quoted against previous seasons and what constitutes a top 4 side -  I agree they should be used  to show whatever you think -but with those contexts  their 'relevance' has already been determined.

you say 'these are trending upwards' thats means we are improving - no it doesn't it means some arbitary factor in a game measured against another arbitary assessment of what constitutes a top 4 side is 'improving' - its only progress if you place any weight on the stat being useful and relevant but without context it isn't.

The context of each of our games will be different as will the outcome. We've played different formations and personnel in those games , we've had players sent off and been ahead and behind, we've played home and away, good teams and bad and for every factor thats impacted us there has been a similar one for the opposition. Does the stat mean anything of itself? Does it have any real value? Does it even represent a trend given the small sample I'm not sure.

Come up with a stat that takes this all into context and you've actually got the stats for one game and what can you learn from one game?

Put these stats as your measurement for performance and as I said above they will 'improve' because consciously or subconsciously you'll adjust to make them improve  and it wont mean a tap. Are these things the building blocks to a good side or are they a result of being a good side.

If Rodgers goal is death by football then he will be monitoring possession stats and that must influence our risk taking and pass selection. Thats the concern for me. Death by football is fine if your leading, a fitter side and dont wont to use the adrenalin of the game or the crowd. It strikes me as particularly passionless.

Are there any stats which specify the pace of the passing, how incisive those passes are, how many players they bypass or take out of the game, how much space they create, how much time they give the recipient? Any stats which show how many opportunities we could have had if the 'right' pass selection had been made? Any stats on the right pass with teh wrong execution, the right pass with a poor touch by the receiver? Any stats on how many options each man with the ball had to choose from? But do that and then you have to consider yes but he was playing against this type of team, this type of defence, this player. All of these would be useful in context but basically winning football matches is the only genuinely significant stat. To do that you need to score more than the opposition. How you get there is largely subjective there is no single way of playing not least because the opposition have their own say on what will work,  so even an ambitious model of fluid dynamics is going to struggle to be meaningful.

As I tried to say early stats are interesting but too much importance is dangerous and if as Al implies they dictate what you do then thats just wrong.

[BreakfastPercy]

winning football matches is the only genuinely significant stat.

That has to be one of the most close-minded clichés in football, and to me suggests you've not truly read the OP.

[Vulmea]

not sure what you mean by cliche

true , factual or untrue and not longer valid, cliche has so many meanings...............

are you suggesting there is a more significant stat? if not I'd guess you haven't truly read my response.

[BreakfastPercy]

Cliché as in an overused expression. I've seen that comment a lot lately on here, which obviously isn't your fault,  but it just blocks debate. 'Winning is all that matters' is used to end a range of ideas people are uncomfortable with. If we truly lived by it would never watch the game, skip straight to the result, and forums would be defunct. Your post is a critique of statistics in football that could have been written without ever looking at the OP, which to me suggests an agenda, which the author requested be left at the door. I apologize if I'm wrong, but that's how it read to me.

The data is, and can be, neutral. Look at the data, and compare it to your own subjective appraisal of improvements or deficiencies in performance over the course of the season if nothing else. Test it rather than pre-judge it. DK and the OP don't say 'Stat A' results in a win. What they have done is devise metrics that correlate with our performance targets (namely top four league position), and we are now getting the first lots of data to see how we're doing by those metrics. There are no definitive conclusions yet, so there's no need to write-off the metrics yet.