Does Squad Consistency Breed Success in Football?

Does Squad Consistency Breed Success in Football?

The Football League published a list of the number of players used by each club in the Championship, League One and League Two this season earlier this week and it prompted me to do a statistical analysis of squad consistency versus success.

I have been having a similar argument with my friends, colleagues and people on social media for a while over whether International squads should be consistent or whether the players should be called up based on recent form. 

I am of the opinion that the former is the far better option.

I guess it's quite a traditionalist view, similar to the way some people think managers are not given enough time to make their mark 

I think that squads and starting elevens, at both domestic and international level, are too frequently chopped and changed.

Of course, sometimes managers must test and rotate their players when injuries and suspensions occur and also when the players they are picking just aren't up to scratch but on many occasions, I feel that players underperform simply because they don't have the chemistry with their teammates that can only come with time and experience.

Changes in management also unsettle squads with new managers likely to ring the changes until they decide on their own favourite line-up. 

The League Managers Association (LMA) announced recently that there have been 47 managerial changes in England's top four divisions this season with the Championship seeing a disproportionately high percentage of these, 43%.

There have been a total of 2,474 players used in the Football League this season, that's the 72 clubs immediately below the Premier League.

On average that's 34.4 players per team (I know you can't have 0.4 of a player!) but the range varies from as low as 24 used by Bristol City, Brentford and Wycombe up to as much as 50 used by Blackpool.

Just a quick glance at the teams that have made the least changes tells you that they've achieved more success but to what degree? 

I thought I'd do a statistical analysis using the Spearman's Rank test to check the correlation between Squad consistency and success.

Squad Consistency vs Final Position in the Championship

The number of players used by the club. 

• Brentford...... 24
• Bournemouth.... 25
• Wolves......... 27
• Middlesbrough.. 30
• Blackburn...... 30
• Ipswich........ 31
• Derby.......... 31
• Watford........ 31
• Norwich........ 31
• Birmingham..... 31
• Sheff Wed...... 31
• Charlton....... 32
• Leeds.......... 34
• Notts Forest... 34
• Brighton....... 35
• Huddersfield... 36
• Fulham......... 37
• Cardiff........ 38
• Reading........ 38
• Millwall....... 40
• Wigan.......... 41
• Rotherham...... 42
• Bolton......... 44
• Blackpool...... 50

In many ways just the briefest of scans over that table and you could be forgiven for thinking that's how the final season standings looked as well. 

The correlation between squad consistency and success immediately looks quite high but I thought I'd use some stats to make sure.

Spearman's rank is a trusted technique used for testing the correlation and significance of a relationship between two datasets in this case the success of a football clubs season, classified by it's position in the final table standings, in relation to the amount of football players used in that season. Of course this analysis only looks at the total number of players used and not how frequently they made appearances so it's not a perfect true assessment of the squad's consistency but it should be a very good indication.

I've calculated the Spearman's rank coefficient of these two datasets in the English Football Championship and it is +0.88. Where -1 is a perfect negative correlation and +1 is a perfect positive correlation, our figure of +0.88 shows that the correlation between squad consistency and final position is very high in this division for this particular season.

The ‘Degrees of freedom' technique also identifies the significance of this relationship and with these datasets producing a figure well below the 0.1% significance level this means we can be 99.9% sure that this relationship has not occurred by chance.

The analysis suggests that a more close-knit squad with fewer players having more consistent involvement in games will perform better. There are a few exceptions to the rule which is the reason why our coefficient is +0.88 and not a completely perfect +1. Cardiff City used 38 players, the joint 6th most yet finished higher in a respectable 14th position proving that using a lot of players doesn't have to end in abject failure. Whilst the likes of Cardiff buck the trend slightly they hardly make a dent in the overall relationship. If Bournemouth used the most players in the league but finished 1st that would make a big difference but they didn't; they used just 25, the 2nd fewest.

In fact 7 of the 8 teams finishing in the top 8 of the Championship this season also used 7 of the 8 fewest players, all 31 players or fewer. That shows just how significant this correlation is to me and should be food for thought for managers who like to rotate frequently. The same relationship is true at the bottom; it's no surprise to hear that Blackpool, who finished rock bottom in 24th, used the most players, a whopping 50, 6 more than any other side. Perhaps if they had more consistency in their squad they may have been a lot closer to avoiding relegation or even beaten the drop.

Just to reiterate the problem of inconsistency in line-ups, we can see that 4 of the 5 teams to use 40 or more players in this season's Championships finished in the bottom 4, with Bolton the only exception.

Squad Consistency vs Final Position in League One

The number of players used by the club 

• Bristol City... 24
• MK Dons........ 28
• Preston........ 29
• Walsall........ 29
• Fleetwood...... 30
• Bradford....... 30
• Doncaster...... 31
• Port Vale...... 31
• Rochdale....... 31
• Crewe.......... 32
• Chesterfield... 32
• Swindon........ 32
• Gillingham..... 33
• Sheffield Utd.. 33
• Peterborough... 33
• Leyton Orient.. 35
• Coventry....... 36
• Crawley........ 36
• Oldham......... 38
• Scunthorpe..... 40
• Yeovil......... 40
• Barnsley....... 42
• Notts County... 44
• Colchester..... 44

The Spearman's Rank Coefficient for the relationship between squad consistency and final position in League One in the 2014/15 season is +0.68.

This shows a strong positive correlation, however not as strong as the Championship suggesting that some success can be had with less consistent squads and also that some sides have struggled even when keeping their teams consistent.

The Degrees of Freedom figure is bang on the 0.1% significance curve which tells us that this relationship is 99.9% certain not to have occurred by chance, similar to the Championship.

Some stand out figures are Barnsley who used a whopping 42 players yet still managed to finish 11th. Barnsley had to bring in Lee Johnson after sacking Danny Wilson so this could explain the inconsistency in the squad as Lee Johnson needed to find his favourite line-up.

The opposite is true of Leyton Orient who only used the 9th most amount of players yet finished 2nd bottom. With an average consistency in their squad you would have expected them to finish higher than that, especially considering they were 2nd favourites for promotion before the season began. Walsall only used 29 players, the joint-3rd least yet finished 14th showing that consistency doesn't always bring success though they did make it to a Wembley Final in the shape of the Johnstone's Paint Trophy.

Squad Consistency vs Final Position in League Two

The Number of players used by the club 

• Wycombe....... 24
• Southend...... 28
• Morecambe..... 28
• Exeter........ 28
• Plymouth...... 30
• Dagenham...... 31
• Shrewsbury.... 31
• York.......... 32
• Newport....... 33
• Luton......... 33
• Bury.......... 33
• Northampton... 35
• Wimbledon..... 35
• Burton........ 36
• Portsmouth.... 37
• Stevenage..... 37
• Accrington.... 38
• Hartlepool.... 38
• Carlisle...... 39
• Mansfield..... 40
• Oxford........ 41
• Cambridge..... 41
• Cheltenham.... 42
• Tranmere...... 48

The Spearman's Rank Coefficient for the relationship between squad consistency and final position in League Two in the 2014/15 season is +0.72.

This shows a strong positive correlation, not as strong as the Championship but stronger than League One suggesting that some success can be had with less consistent squads and also that some sides have struggled even when keeping their teams consistent but in general there remains a strong trend; the more consistent the squad the more success is had.

Southend and Wycombe used the least players in the League Two season, 28 and 24 respectively, fitting then that the Play-Off Final should be contested by those two clubs. 

Tranmere and Cheltenham used 48 and 42 players respectively and they finished bottom and 2nd bottom, the pattern continues.

Interestingly Burton, Shrewsbury and Bury, the 3 sides to win automatic promotion, were quite far from using the fewest players

The rest of the division tended to follow the trend much better though.

Conclusions

I hope you've found this interesting. 

It's been quite an eye-opener for me, I just wish I was a football manager and could put this sort of stuff into action! 

There is an undeniable correlation between squad consistency and success suggesting that ‘the more consistent the squad, i.e. the less players used in a season, should and does lead to more success in most cases'. As for how can we profit from this from a betting perspective I would suggest backing sides when you know they've kept their line-up similar for much of the season, however the most consistent line-up in the Premier League this season was probably Burnley and we all know how that ended so it doesn't always breed success.