Lies, lies, damn lies and statistics

[Mundell]

Lies, lies, damn lies and statistics.

I don't know whether anybody on this forum has any interest in how data is now being used to analyse the many different aspects of player and team performance, but I thought I'd post a speculative comment.

With every pass, move, tackle, dual, shot, save etc. in many games, week in week out, being recorded, measured and analysed there is, potentially, a huge amount of data, analysis and opinion out there for us fans to assimilate and discuss. One particular area of interest might be analysis which attempts to measure underlying performance to assess whether a team's results are better, or worse, than might be expected and hence sustainable. We might do this simply out of interest, which is my focus, or to predict future results and make money betting on them. Matthew Benham, the owner of Brentford, has made a lot of money this way, of course.

The "Fink Tank" website has produced this type of analysis for a number of years with their ranking system and game simulator. Unfortunately, from next season, Fink Tank's analysis will not be available, directly. They've entered into a partnership with an outfit that calls themselves the Secret Betting Club and are selling what amounts to a betting advisory service based on their system.

More recently, Squawka, for example, have begun to do something similar. Based on Opta data, and their own algorithms, they produce individual player and team performance scores, while also providing much more granular performance analysis than Fink Tank. There are others trying to do much the same thing.

How does any of this help us and what does it mean for Charlton Athletic? While analysis of this kind needs to be interpreted with care, I would suggest that it can provide useful insight into our performances and, hence, how the season is likely to develop. For example, last season both Fink Tank's rankings and Squawka's performance scores had a very high correlation with final league table position. There'd be something wrong if that wasn't the case. However, there were some standouts, i.e. teams that didn’t finish where their rankings and/or ratings predicted, and we were one of them.

On Squawka, for example, our performance scores placed us 20th in the table, well below where we actually finished. Of course, no "model" is without error and there may be fundamental and sustainable reasons which explain why we did better than our performance scores predicted.

However, that would mean that we were unusual in some way; a very high chance conversion rate or an exceptional goalkeeper, for example. The risk, of course, is that even if these explanations are valid they may not be sustainable.

Two things struck me last season. First, while we enjoyed a very good start results wise our performance scores, i.e. the assessment of our underlying performances, never measured up. It seemed very unlikely, therefore, that we could sustain our early season form and push for a play-off place. The data said we just weren't playing well enough. And so it proved. That wasn't because the squad had insufficient depth or because it wasn’t strengthened. This probably didn’t help, but the real issue was that we weren't good enough to start with.

Second, though we produced some very good performances under Guy Luzon, Brentford and Huddersfield at home, for example, overall our performance scores under him were behind those achieved with Bob Peeters, despite the fact that Luzon enjoyed a full strength and more complete squad. Not what I expected, but a potential cause for concern.

On players by the way, Francis Coquelin was our best performer, based on scores which I adjusted for minutes played, while Berg Gudmundsson had the highest overall score. No surprises there. Ben Haim and Bikey-Amougou also did well and were key to our early season form.

I'm sure this post is already way too long, but I was prompted to write it by something I came across in the last few days so I’ll briefly cover that.

At the end of last season the guy who runs the "Experimental 3-6-1" site produced a somewhat damning analysis of our performances. We really were a real outlier last season. The format and graphics are easy on the eye and the messages are very clear. I thought it was an interesting and provocative piece. Here's the link.

experimental361.com/2015/05/10/championship-attack-defence-201415/

He's now produced his own rating and ranking system which, in essence, is based on the number and quality of shots, for and against and, hence, expected goals for and against. The key here is the way "quality" is measured, i.e. it’s not just about numbers, but he also makes an adjustment for the quality of opposition played so that the assessment of “form” is not distorted by runs of “easy” or “difficult” matches.

For what it’s worth, based on his analysis we were rock bottom of the Championship table last season. A bit extreme, I'm sure, but thought provoking nevertheless. Moreover, as was the case with Squawka’s assessment, his graphic very clearly illustrates the deterioration in our underlying performances after Luzon's arrival, based almost entirely on what happened with our defence. Here's the link.

experimental361.com/2015/07/14/e-ratings-championship-201415/

As I’ve already said, this type of analysis needs to be interpreted with a degree of caution, perhaps even skepticism, but is probably worth being aware of. Looking forward to the forthcoming season, I'd have the following observations.

First, our performances will need to improve from last season simply to ensure that we can be confident of avoiding a relegation scrap. They'd need to improve very significantly indeed for there to be any hope of a play-off place. This is not simply about “pushing on”. It's good that there's been a major restructuring of the squad.

Second, though results appeared to improve dramatically under Luzon, he's yet to prove he can deliver more than the sum of the parts.

Third, I believe (and hope) that the bookmakers are wrong, but they're no mugs!! There are, it seems, very good reasons for their pessimism about our prospects.

Finally, I’ll be looking at the analysis produced by Squawka, and others, during the course of next season, especially over the first 10-15 games, to assess whether it can tell us anything about how we are faring which is not obvious from the stands and the results we are producing. If there is any interest I’d be happy to post the occasional update.

Apologies again that this is a bit long. Hope the links work.

[shg]

Great article Mundell

[jonkool]

A very interesting and thought provoking piece which may explain why we were relegation candidates pre season both last season and this.

As you say the bookies and their analysts can get it wrong but they are no mugs.

Thus coming Saturday is going to be important as a good performance and 3 points will raise morale and hopefully see the squad gel together.

[reamsofverse]

Great write up Mundell enjoyed reading it.

[Deleted]

Well I was going to get out early but up pops Mundell with another must read article. Reading through, the one factor you cannot input with any accuracy is "Luck" I hear fans say "If only we'd have been given that penalty" etc. But of course, this is the view and mindset of the supporting fan...yet if you stand back and forget about what your heart says, you can see how lucky we were, match after match.

Unlucky Charlton? Nope, I think the opposite is true.

Thanks MF...I can now go out 

[seriouslyred]

Indeed, luck together with the culture of a club cannot be factored in but I would venture that the player acquisitions and disposals every year and skill of coaching and player motivation account for most of what we see every Saturday.

Looking at this type of analysis, the games in August and the financials we can all see how our club is doing. We don't have to leverage individual errors or snippets of ITK rumours to show how knowledgeable we are.

Whether our preference is for short words or long ones, twitter 140 character bullets or posts of essay type proportions, we can take a balanced view. Without going into hysterics we can all have a view of the board and football management by say end September - are we better or worse than 12 months before? And what is the gap between us and the top eight.

This type of sober reflection on progress is far more likely to hold sway with the club than rants about ex players, coaches and "we want our Charlton back"

I really hope the board improve the squad with final purchases and FAPL loans. This way the bookies can recalibrate based on Squawka and 3-6-1 analysis etc so that we know we will finish higher than last season and not lower.

[seriouslyred]

Aug 2, 2015 10:24:51 GMT Mundell said:

Lies, lies, damn lies and statistics.


More recently, Squawka, for example, have begun to do something similar. Based on Opta data, and their own algorithms, they produce individual player and team performance scores, while also providing much more granular performance analysis than Fink Tank. There are others trying to do much the same thing.

On Squawka, for example, our performance scores placed us 20th in the table, well below where we actually finished. Of course, no "model" is without error and there may be fundamental and sustainable reasons which explain why we did better than our performance scores predicted...

Wiggins exits!

Mundell aluded to it the other day but just take a minute to look at Squawka stats for last season.

Every single low performer except Harriott, Piggott and Fox have now left the club! Unfortunately high performers like Bulot, Ben Haim and Buyens have also left but that's part of the model risk when older players are only on one year deals. They are there in the main to support the team to hit a standard and assist the younger players develop.

Wiggins is 27. The only players left who are 27 or older are Jackson (who probably won't start too many) Diarra, Bikey and Reza.

If the 15 retained players improve AND most of the new intake hit the ground running then we should have a big improvement on last season in performance?

[Mundell]

Ten games into the season and with an international break upon us, I thought now might be a good time to update the Lies, lies, damn lies and statistics thread I started before the season kicked off.

Before I get into the discussion, I should perhaps stress that I'm no expert in the area of football data analysis and I confess to being a little puzzled by what I've found. That probably means either that the data is telling us something interesting or, alternatively, that there is something wrong with it!! I guess that's potentially a subject for discussion and debate.

To state the obvious, the idea here is that random chance, good and bad, can have an important impact on football results, even over a prolonged period, so that the league table doesn't always give us a true reflection of a team's strengths relative to others in the league. It follows that if we can identify a robust measure of underlying performance, i.e. something that goes beyond the actual results so far, then it might be possible to predict future results with more confidence. We know where Charlton are in the league table, but are we playing more like relegation candidates or play-off contenders?

In the original post above I referred to three different sources of analysis, "Experimental 3-6-1", "Squawka" and "Fink Tank". As I noted at the time, Fink Tank's analysis is no longer available - they're now selling advice based on it - so I've replaced it with "Whoscored.com". As far as I can tell, Whoscored is similar to Sqauwka. In essence, they take Opta data and then develop algorithms which generate performance measures which, in turn, correlate with results, at least on average. Whoscored is an impressive source of match analysis.

As I explained above, it's very important to bear in mind that even if these measures of performance work on average, they may not do so for a particular team. There may be something about that team, it's players or its style of play which doesn't fit the standard model. In such a case, a team may continue to out/under perform expectations throughout the season. In other cases, a team's results may "mean revert" to the predicted outcome and, of course, that's the more likely scenario, by definition.

I'll start with Experimental because the analysis is very intuitive and easy to interpret. The output is an expected score based on an analysis of shots taken and defended. For example, the number and type of shots taken in the home game against Hull City would, on average, be expected to lead to 1.3 goals for Charlton and 1.0 for Hull. In the away game against Derby the "result" was 0.6 v 2.4.

We don't rate very highly on this type of analysis. It suggests that our home games against QPR, Hull City and Huddersfield were very close and perhaps should all have been drawn. Rotherham we should have beaten easily (2.4 v 0.7) while we probably shaded the match against Fulham (2.1 v 1.6). We've been second best in each of our five away games. The simple message here would seem to be that we're not creating enough chances and while we might be solid at the back we are getting opened up away from home.

If we took this analysis at face value, we might be forced to agree with the bookmakers and conclude that we are clear relegation candidates. However, fortunately, for what it's worth, Squawka reaches a very different conclusion. Last season Squawka, whose rankings produced a very high correlation between each team's performance score and their final league table position, had us as a real outlier. Based on their analysis we ranked 20th in the performance table, well below our actual finish, but this time around the reverse is true.

We are currently ranked sixth based on Squawka's analysis with an average performance score per game which is not only very significantly higher than we produced last season, but which if sustained over the entire season might be good enough for a play-off place. This is puzzling, to me at least, so what might be going on?

As might be expected, our performances have been quite erratic, though to be frank I don't know to what extent that might be typical of many teams. Our best performance, viewed through the Squawka lens, was against QPR. It appears we battered them and fully deserved our 2-0 win. Our next best performance was in the home game against Rotherham. Perhaps we were all so disappointed not to win such a winnable game that we overlooked the fact, notwithstanding our lack of creativity and guile, that we did more than enough to get all three points.

On the other side of the coin, we were well beaten at Cardiff and especially at Blackburn - no wonder Guy Luzon was upset with those two performances - and had the worst of it at Wolves too. It appears we may have been fortunate to get a draw at Derby while having marginally the better of it at Forest. Back at the Valley, we just shaded the games against Hull and Fulham, though not significantly so, but Huddersfield did a job on us and just about deserved their win it seems.

Before trying to make some sense out of this, it's worth taking a quick look at Whoscored. Quite encouragingly, they rank us 10th in the table on performances (compared to 19th last season), but echo the Squawka message about inconsistency. Indeed, based on both their analysis and Squawka's the average performance score of our opponents is higher than ours - not what might be expected given our overall performance rankings. This is an outcome which appears to result directly from the fact that we've had some real off days. Indeed, it turns out that Whoscored's analysis fits quite well within intuition and, I suspect, for the most part, the post match views expressed on this forum.

Whoscored conclude that we dominated QPR and then just shaved the next three games. We had an excellent start, but then fell away at Wolves where we deserved to lose. We should have beaten Rotherham, but deserved to lose against Huddersfield and were then very poor at Blackburn and Cardiff. We then recovered to just edge the game against Fulham.

Lies, lies, damn lies and statistics indeed, but what do we make of this? For what they are worth, here are my broad, but tentative conclusions;

1. When at full strength we are a very capable side. Almost certainly better than last season, but that doesn't mean we'll finish higher in the table, of course. Last season we should probably have finished lower than we did.

2. We lack a cutting edge though and in only two games have we created enough chances to "expect" to score more than once, against Rotherham and then against Fulham. That's potentially a real problem and it could cost us as the season progresses. (Note, again for what it's worth, that Squawka has a slightly more optimistic read on chances created).

3. Our lack of strength in depth is a real concern. We failed to turn up at Blackburn and Cardiff and our performances appear to drop off quite quickly when without key players. This could also prove to be costly.

4. We are, perhaps, going to be at our best this season when playing at home against teams who come to the Valley to attack. These games may prove to be very entertaining. A welcome bonus following the more attritional football of recent seasons.

5. On the other hand, the ability of a very average Huddersfield side to frustrate us and secure three apparently deserved points has to be a concern. It could be another tough season if we can't break down the more defensively minded teams.

6. We began the season with two creditable away draws, but we don't look likely to score more than once away from home and we certainly don't look like winning. We'd be likely to go down if things don't improve. The good news here is that we've probably had tougher than average opponents in our away games, just.

7. The other side of the lack of strength in depth coin is a dependency on a small number of key players, Gudmundsson, Cousins, Makienok, Bauer, Diarra, and, of course, Henderson. Keep those six players fit (or get them fit) and we should have enough quality to be comfortable. Lose two, three or four of them for an extended period and it could be a very stressful season.

8. Perhaps it's easier, and much more fun, simply watching the matches!!

Any comments welcome. If anybody has got this far, that is.

[somerton]

Very interesting Mundell, and you seem to point to the fact that these stats companies are onto something. For example last season even when we won a game I never felt we had been in control of the game, now I see why.
I have noticed on OPTA stats that in most games we win less than half are tackles, which is why I feel we are nevr in control of a game as we cannot get the ball back or break up attacks.
The on definable is luck though, which of course Napoleon new, when he said give me a lucky general.

[Deleted]

had a look at the Germany stats, interesting, take a look at Stuttgart, going on this model they should be challenging for honors !

experimental361.com/2015/10/04/season-so-far-bundesliga-4-oct-2015/

uk.soccerway.com/national/germany/bundesliga/20152016/regular-season/r31545/?ICID=TN_02_01_04

[seriouslyred]

There are a number of aspects which are interesting about the stats approach, not least the fact that they are put together by people who aren't Charlton fans and probably not Championship club fans. They are trying to add up all of the chances, tackles and passes and then, irrespective of goals actually scored, one can see a pattern of progress, or not!

What adds to this is that a number of journalists have revealed that club boardrooms up and down the country are studying Prozone and Opta and using this to judge player acquisitions, targets and coaching performance. Rumours abound of Peeters losing the dressing room etc are just that - our performances went downhill after the first ten games and I believe that this is what the board examines when attempting to steer the club.

If anybody is interested in Squawka then here are a few pointers
1) they suugest that a score of 20 for a player in a game is good but there is no maximum nor minimum limit.
2) Negative scores are possible and last season we had a number of players on negative scores for the season!
3) If 20 is good then 40 games with an average of 20 (800) presents a decent player in the Championship. Take a look at the top clubs last season and all their regular players will range from 400 to 1,000

As Mundell states, we simply do not have strikers who will attain that level unless Watt and Makienok delivers their best every game. And we absolutely do NOT have the strength in depth. Both of these can be addressed in the next two windows if that is the path the board chooses.

Also from above there is a "dependency on a small number of key players, Gudmundsson, Cousins, Makienok, Bauer, Diarra, and, of course, Henderson." I would add that we also have Solly and Kashi delivering respectable performances and scores. And Bergdich looked OK at left back the other day and I'd like to see more of Sarr.

I won't detail scores as it's too early, the sample is too small and I really need to talk to someone who knows football inside out first before committing to what would be a data based assessment of our club. But it does look like we might have quite a few players who might score 500-900 this season.

A word of caution as it all started off so well last season and our Squawka stats are not dissimilar this time around after ten games. The difference is that we have Henderson, Makienok and Kashi about to return and Luzon as head coach. So give it another five-ten games watching the team and glancing at the stats and we may well see a departure from the path trodden last time.

And we can all be well informed as to whether we are better or the same as last season. That doesn't mean we all agree but there might well be a consensus on many aspects as well as violent disagreement on others! As posted elsewhere if the Squawka data is a true representation and the squad is more consistent than last time then we live in interesting and exciting times.

Time will tell!

[reamsofverse]

Enjoyed that last post Mundell however if you were to go back over the post match views threads on every game I have basically told it how squawka see it. Not blowing my own trumpet here at all but when you have been a scout in the Professional game, whether that is watching upcoming opponents or individual players, both of which I did while at Cardiff, you just see the game differently to other fans.

Offensively I have said on here a hundred times that we are poor in the final third, lack a cutting edge and have no killer instinct. What I see when at games is not us trying to rectify that but I witness in the flash how other sides know it too and set up to take advantage of it.

[Deleted]

Oct 6, 2015 14:12:03 GMT seriouslyred said:

There are a number of aspects which are interesting about the stats approach, not least the fact that they are put together by people who aren't Charlton fans and probably not Championship club fans. They are trying to add up all of the chances, tackles and passes and then, irrespective of goals actually scored, one can see a pattern of progress, or not!

What adds to this is that a number of journalists have revealed that club boardrooms up and down the country are studying Prozone and Opta and using this to judge player acquisitions, targets and coaching performance. Rumours abound of Peeters losing the dressing room etc are just that - our performances went downhill after the first ten games and I believe that this is what the board examines when attempting to steer the club.

If anybody is interested in Squawka then here are a few pointers
1) they suugest that a score of 20 for a player in a game is good but there is no maximum nor minimum limit.
2) Negative scores are possible and last season we had a number of players on negative scores for the season!
3) If 20 is good then 40 games with an average of 20 (800) presents a decent player in the Championship. Take a look at the top clubs last season and all their regular players will range from 400 to 1,000

As Mundell states, we simply do not have strikers who will attain that level unless Watt and Makienok delivers their best every game. And we absolutely do NOT have the strength in depth. Both of these can be addressed in the next two windows if that is the path the board chooses.

Also from above there is a "dependency on a small number of key players, Gudmundsson, Cousins, Makienok, Bauer, Diarra, and, of course, Henderson." I would add that we also have Solly and Kashi delivering respectable performances and scores. And Bergdich looked OK at left back the other day and I'd like to see more of Sarr.

I won't detail scores as it's too early, the sample is too small and I really need to talk to someone who knows football inside out first before committing to what would be a data based assessment of our club. But it does look like we might have quite a few players who might score 500-900 this season.

A word of caution as it all started off so well last season and our Squawka stats are not dissimilar this time around after ten games. The difference is that we have Henderson, Makienok and Kashi about to return and Luzon as head coach. So give it another five-ten games watching the team and glancing at the stats and we may well see a departure from the path trodden last time.

And we can all be well informed as to whether we are better or the same as last season. That doesn't mean we all agree but there might well be a consensus on many aspects as well as violent disagreement on others! As posted elsewhere if the Squawka data is a true representation and the squad is more consistent than last time then we live in interesting and exciting times.

Time will tell!

The stats approach by the Brentford chairman has worked well so far.

Warburton must be having a wry smile to himself.

[seriouslyred]

Hey @elfsborgaddick , do you mean the guy who made this happen in just 12 months?

www.theguardian.com/football/2015/jul/27/how-fc-midtjylland-analytical-route-champions-league-brentford-matthew-benham

[Deleted]

Oct 6, 2015 18:59:18 GMT seriouslyred said:

Hey @elfsborgaddick , do you mean the guy who made this happen in just 12 months?

www.theguardian.com/football/2015/jul/27/how-fc-midtjylland-analytical-route-champions-league-brentford-matthew-benham

Yep that is fair enough.

BUT Brentford are 20th in the division and have sacked their manager after a few games.

Weir and Warburton were not having any of it either.

My money says Brentford will be bottom six come the end of the season.

[seriouslyred]

Two things:
I look forward to help making that happen when we play them soon.
I don't think you can run a club on data analytics alone - it needs hard cash and leadership.

We saw this with Curbs/Varney/Murray 20 years ago. Incidentally I've just started Sir Alex Ferguson's book on the same subject.

[Deleted]

Oct 6, 2015 19:43:31 GMT seriouslyred said:

Two things:
I look forward to help making that happen when we play them soon.
I don't think you can run a club on data analytics alone - it needs hard cash and leadership.

We saw this with Curbs/Varney/Murray 20 years ago. Incidentally I've just started Sir Alex Ferguson's book on the same subject.

And decent managers left alone to do their job.

[Deleted]

Some fantastic work has gone into this thread, but I'll admit all it gives me is a headache.

The thinking these days (about everything, not just football) is that any objective can be achieved if enough data is gathered and analysed. With our ability to gather data increasing exponentially, the logical conclusion would be that before long every variable will be accounted for.

Once that stage is reached, there will be no point in actually playing a football match. The result will already have been extrapolated from existing data.
Indeed there will be no point doing anything at all, for the same reason.

[Deleted]

Oct 6, 2015 21:22:51 GMT @lardiman said:

Some fantastic work has gone into this thread, but I'll admit all it gives me is a headache.

The thinking these days (about everything, not just football) is that any objective can be achieved if enough data is gathered and analysed. With our ability to gather data increasing exponentially, the logical conclusion would be that before long every variable will be accounted for.

Once that stage is reached, there will be no point in actually playing a football match. The result will already have been extrapolated from existing data.
Indeed there will be no point doing anything at all, for the same reason.

What a refreshing post...I also have the same headache.

[seriouslyred]

Sir Alex opens his book with an anecdote about the stodge players used to consume in a pre match meal and how he changed that as a manager.

Like data analytics, diet and fitness have come a long way in twenty years.

Neither will make a player out of you and I so there are limitations. And the beauty of football is that the outcome is never certain.

But having the right medical team, nutrition and analysis can all assist Guy Luzon get the best results over the season. And help the board get improvements without wasting money.

[Deleted]

Oct 7, 2015 21:41:33 GMT seriouslyred said:

Sir Alex opens his book with an anecdote about the stodge players used to consume in a pre match meal and how he changed that as a manager.

Like data analytics, diet and fitness have come a long way in twenty years.

Neither will make a player out of you and I so there are limitations. And the beauty of football is that the outcome is never certain.

But having the right medical team, nutrition and analysis can all assist Guy Luzon get the best results over the season. And help the board get improvements without wasting money.

Looking forward to reading that book, SR. Apologies for seeming ignorant but I read frequent references to 'the board'. Who is on it? Don't we just have an absentee owner and his 'Person' making the decisions?

[Deleted]

Oct 7, 2015 21:41:33 GMT seriouslyred said:

Sir Alex opens his book with an anecdote about the stodge players used to consume in a pre match meal and how he changed that as a manager.

Like data analytics, diet and fitness have come a long way in twenty years.

Neither will make a player out of you and I so there are limitations. And the beauty of football is that the outcome is never certain.

But having the right medical team, nutrition and analysis can all assist Guy Luzon get the best results over the season. And help the board get improvements without wasting money.

You've missed the point SR, all clubs have got to be successful NOW!!!  Damn this slow build stuff and financial stability nonsense.  Live now, pay later!

Keeps tongue in cheek.

[seriouslyred]

Win, lose or draw I've followed CAFC all my life. Those who want instant success follow Citeh or whoever. Those who choose our club or someone similar can wear it as a badge of honour. Or perhaps adopt a sense of martyrdom or some psychotic canibalistic tendancy to eat ones own kind!

Of course I want us to win and a favourite memory is the St Andrews game vs Leeds back in the 80s when we came back to win 2-1. Wrexham away in the cup when we conceded six comes to mind too.

As for our board, I think that's for another thread. I've already posted some clear views on challenges in this area. Many, many years ago, an anarchist friend suggested that the best form of governance is a benevolent dictatorship. When I read the manifesto of the "opposition" in our CAFC landscape, one might be inclined to agree!

[Mundell]

I created this thread just before the season started and then updated it in early October.

It's a bit of a depressing read, but I thought I'd add a link to the post Hull City (A) update from Experimental 3-6-1. His analysis before and during the season has been on the money, unfortunately. Both he and the bookmakers have been right. 

These Scatter plots are very nicely done and easy to read. This is a link to what the blogger call his E Ratings. It was with this metric that he predicted our plight at the beginning of the season, based on our performances last season. Also in this post you'll find his estimates of promotion and relegation. On this analysis we have a 93.9% chance of the drop. That's based on our points total to date, the fact that based on the so-called E ratings we are the poorest side in the Championship by some margin, and on our remaining fixtures.

A 6.1% chance of survival is better than no chance, but we clearly need to improve very significantly to have any real prospect of retaining our Championship status. I know that's not telling us anything we don't already know, but its striking to me just how damning this analysis is. 

[bigandy99]

We need a massive change in our results! Starting Saturday, we have a decent set of fixtures - the end of Feb will tell all in my opinion...

[Mundell]

Thought some may be interested in the Experimental 3-6-1 timeline for the Bolton game here. The stats tend to confirm that it was an encouraging performance.

The season to date data is a bit of a mess, but here it is for what its worth. Again, the scatter plot confirms that we've made a solid start. The match by match analysis shows we've been improving.  

[novice]

Despite being a bit useless with this sort of maths, I've really enjoyed these. I appreciate the value added nature of them as the sort of raw data that gets used on Sky can be hugely misleading. Although early days, it appears that our defence is pretty average (up from basket case last season) and our attack doing a bit better than that- we can hit a barn door this season! It would appear from my naked bud admittedly dodgy eyes that our defence is improving so this will hopefully translate into improved stats as he season moves on. The one interesting set of stats I find interesting and depressing at the same time relate to the Premier League and I wonder if they are available for EFL as well. There's a guy called Paul Tomkins who's a Liverpool fan and has a journalistic blog called The Tomkins Times. He had some data and I'm not sure whether it was his or third party that simply added up the transfer fees and salaries of all the players used in the premier league and ranked the clubs accordingly. He showed that almost without exception, you finish in the same league position as your financial outlay. Occasionally a team might end up one place better or worse off but that was about it. The number of exceptions who performed more than one place better or worse was something ridiculous like 10 clubs in 15 seasons. Of course last season may have blown that out the water. I'd be interested to know if this data is available for EFL leagues and whether the same applies.

[seriouslyred]

novice there are two places one can go to get an approximation of the salary bill for EFL clubs: Swiss Ramble blog and Transfermarkt.com.

Swiss Ramble has recently written two excellent pieces on Huddersfield and Millwall who have similar budgets and crowds to ourselves. Ironically the way forwards after Powell for Huddersfield was a foreign coach and a change of strategy - basically do things smarter and better than the opposition, not spend more money.

The graphs he produces show the differences in revenues and expenses between Barnsley and Newcastle.

As for Transfermarkt, the valuations by squad have a very high correlation with league position at the end of each season - often as high as 75% or more.

Right now we have the highest value squad in League 1 even if one takes off Teixeira so no surprises we are competitive.

Obviously there is more to it like coaching and finding the final parts of the jigsaw but in the main, player wages and valuations will determine league position.

Another way to look at it is to check the bookies odds which state we are 3:2 to finish top six or sixth favourites in this market. And that is where the board stated they wanted to position the club at the start of the season.

And that is perhaps why some choose to highlight these data factors rather than cry out endlessly about this player being sold or that player not working out.

We have stopped signing poor players and for me the ship has been stabilised. Bring on the final signing(s) and the next competitive match.

The manager has to work with the resources at his disposal and experimental 361 will show which clubs have spent wisely.

So yes, loadsa data out there which gives an indication of how we might do. It doesn't replace individual observations or the experience of watching live football in the EFL.

But it does give a framework, an envelope around expectations if you like.

[Mundell]

novice Many thanks for the reply. Thanks also for the reference to The Tomkins Times. I wasn't aware of the blog and it looks interesting. Bit disappointed that you need to pay a subscription to get the content though!

I read 'Pay as you Play' a few years ago and was a bit disappointed in all honesty. It struck me at the time that the authors had got hold of a bunch of data and thrown it into the hopper to see what came out without thinking very hard about the problem they were trying to analyse. There is a well documented relationship between wage bill and league table position, especially in the Premier League where, as you say, its close to 1-1. If you're interested, I can recommend Stefan Szymanski's books on football economics. For example, Soccernomics (with Simon Kuper) and Money and Football; A Soccernomics guide. Where the authors of Pay as you Play go wrong, in my view anyway, is that they conflate transfer fees, which are perhaps best understood as the cost of acquiring wage bill, with the more direct impact of wage bill itself. Just my recollection of my assessment at the time. 

If you want to read about the economics of football in the Championship; the relationship between revenues, including parachute payments, wages and losses, I can recommend the superb Swiss Ramble blog. Try not to get too depressed though! Its not a pretty or particularly encouraging picture. The manic battle for promotion to the Premier League creates a financial nightmare as clubs run wage bills well beyond their means, making up for the resultant losses with loans from the owner and profits on players sales.

PS Posted this before I saw the post above from seriouslyred hence some duplication. 

[novice]

Thanks both. I'll do some reading. Should make me entertaining at parties.....🙄✌️