Money well spent? : Do £’s make places?

With the Championship season going down to the wire, as far as 2nd place is concerned, I wanted to assess who has earned their place in the top 6 and who has bought it.  These two scenarios aren’t necessarily mutually exclusive, and there are a number of ways to ‘play’ the system, however, lets see what the headlines are.

Looking at the teams that can finish in the top 6, and therefore are in with a shout for promotion, the transfer spends for the 2012/13 season are as follows, (in order of current position):

So in terms of spending money, 5 out of the 8 are in the order (if not the position) they should be in.  The most interesting are Cardiff, who have spent the most (£10,159,600) and  are top of the league and Hull City who have spent the 2nd most (£7,427,200) and are in 2nd.

The anomalies are caused by Watford, who have a +ve transfer expenditure due to the fact that they did most of their business in the loan market (did I hear someone say Udinese Reserves?) and Crystal Palace who sold Zaha to Manchester United for £10,340,000 and then loaned him back.  Brighton should also be mentioned as they appear to be outperforming Bolton, Leicester & Forest.

Now that we know what they have spent, it makes sense to see what difference the spend has had on the season v last season.  Or, has the money spent resulted in an improved performance?

Every team, with the exception of Bolton (who were relegated from the Premier League), have improved on last seasons finish so far, suggesting that any money spent on transfers is money well spent.

A better way to calculate the level of improvement is to measure how much each place that the teams have climbed has cost, or transfer spend / place improvement = value:

Cardiffs climb to 1st has cost them over £2m per place, Leicesters £1.5m transfer expenditure has helped them climb 1 place whilst Brighton are still looking great.  Their rise up the table has only cost them £98k per place which is less than 5% of Cardiffs.  Bolton are saddled with their total expenditure as they came from a higher division.

The average £ per place figure is £799,828.  If I multiply this by the place variance and compare it with the actual £ per place figure we will get a picture of who is over / underperforming against the average.

Clearly Watford & Crystal Palace are overperforming hugely, however if we give them a pat on the back and then move them aside, the next highest overperformers are Nottingham Forest who, using the average, should have spent £9,597,940 to climb their 12 places yet have done it with an expenditure of only £2,763,200 (a variance of £6,834,740).

Cardiff, Hull City, Bolton & Leicester have all underperformed against the average.  I’m sure that fans of each team will be able to offer explanations of how this has happened (Nick Proschwitz anyone?), and Cardiff fans are unlikely to care anyway.

The variance chart, and the League Table based on transfer performance, is as below:

Of course, with a minimum of £120m at stake, none of the above means a jot to the teams that are promoted.

*All transfer expenditures courtesy of transfermarkt.co.uk

Whats Going On?

It looks as though I’ve taken a holiday from blogging, however I haven’t.  I have been asked to provide some posts for EPL Index and so have been busy over there:

http://www.eplindex.com/30415/what-if-half-time-didnt-exist-premier-league-analysis.html

http://www.eplindex.com/30756/do-substitutes-win-games-stats-analysis.html

http://www.eplindex.com/31388/relegated-premier-league-what-happens-next-stats-analysis.html

I will continue to post articles onto this site when time (and inspiration) allows.

Thanks,

Andy

What if…half-time didn’t exist?

Continuing my fascination with Managers and whether or not they actually have any impact on games, I have taken a basic look at which teams, (and by association, Managers), are better at coming back from losing positions at half-time.

I would love to be able to attribute the turnaround of games (or lack of it) to the Managers half-time team talk, however it feels like a bit of a stretch to do this.  If you want to apply the results to Managers performances, feel free.

Using the Championship, I wanted to look at teams that were losing at half time but went on to salvage at least a point at the end of the game.

Below is the chart showing the final result after teams were losing at half-time.

Next is the chart showing just the number of draws & wins that the teams have managed after losing at half-time.

The average number of games won overall is 1.13 and games drawn is 2.25.  Clearly some teams are more proficient at turning games around than others, Blackburn being particularly poor.

The breakdown of won, drawn & lost is:

Whilst some teams are able to salvage points from losing positions, and even grab the full 3 points, statistically it is unlikely to happen.

So, which teams are performing best when it comes to flipping games in the 2nd half?  The starting point is to view the total points gained at full time.

At first glance it appears that Crystal Palace, Blackpool & Watford are the Kings of the Comeback, however, if a team is consistently losing at half time they are more likely (based on the averages) to be higher in the chart.  For example, a team that is behind at half time in 30 games should, on average, gain 18.8 points from those games.

A way around this skewed view is to overlay the number of points gained v potential points available.

Thus, whilst Blackpool look as though they are doing well in respect of gaining points from a half time losing position, this is because they are in this position more often than 20 other teams in the Championship.

As a side note,  is it a coincidence that both have been managed by Ian Holloway this season?  The answer is yes.  It is.  In fact Holloway is responsible for 2 of the 4 wins & both of the draws at Palace and 1 each of wins and draws at Blackpool or, to put it another way, 8 of the 14 points won by Palace and 4 of the 12 points won by Blackpool.

By viewing the above chart as a percentage (Points as % of Potential Points) we get a better view of who is actually overperforming.

The clear winner of King of the Comeback is Crystal Palace, managing to claim a huge 47% of potential points when losing at half-time.  The average is 17% so all the teams up to Leeds United are overperforming and anyone from Peterborough down are underperforming.  Even by removing Crystal Palaces’ figure, the average only drops to 16%.

Barnsley, Birmingham & Wolves fans should be concerned that they are losing at half-time in 46%, 44% & 39% of games respectively, however, the teams with the biggest issues are Blackburn, Huddersfield & Millwall, none of whom have won a game after being in a losing position at half-time.  Is it a question of Managers being unable to motivate their team.  Interestingly, Blackburn have been through an incredible 5 different managers (including caretakers) this season and Huddersfield have changed their manager recently too.  Millwall seem happy to continue with Kenny Jackett.  For now.

Now its time to answer the original question of the post, “What if half-time didn’t exist”.

The assumption is that, without half time (and the various impacts that it brings), the position that teams are in after the final whistle is the same as at 45 minutes.

Below is the actual Championship table as of 10/04/2013.

The table below shows what the Championship table would look like without half-time (or with points gained in the 2nd half after being in a losing position, removed).

As expected, Blackpool & Crystal Palace are the hardest hit, whilst Huddersfield climb comfortably out of the relegation zone, Millwall climb to 11th & Leicester sneak back into the play offs.

Of course, there is another way to look at impacts of half-time, the teams that have gone on to lose after winning at half-time.  That”ll be coming soon.

*All data with thanks to http://www.soccerstats.com

Hull City v Derby – By numbers

It seems that every Hull City game these days is hailed as ‘crucial’, therefore, having being disappointed by the result in Citys’ crucial game v Brighton, here’s a look at the numbers for this crucial game v Derby at the KC, the crucial first of 3 crucial home games that will be crucial to Citys’ success.  Crucially.

Form

Taking form over the last 6 games, City have dropped to 15th in the Form tale whilst Derby sit in 10th.  The 5 places between the two are deceptive though as there is actually only a 1 point difference between Citys’ form and Derbys’.

Derby are unbeaten in 3 games, winning 1 and drawing 2.

Citys’ defensive form seems to be better than Derbys’ as they have managed 3 clean sheets in the last 6 whereas Derby have only kept 1, although Derby have scored in all of their last 6 games whilst City have failed to score in 2 of their last 6 games.

It is worth mentioning that the top 3 in the form tables are 3 of City’s main rivals (Leicester 1st, Cardiff 2nd, Watford 3rd).

City are currently earning over 1 point per game less than both Leicester & Cardiff.

Goals

The following charts show Goals information for City (home) and Derby (away).

Derby concede 1.73 goals per game away, whilst City concede 1 goal per game at home (this is against a league average of 1.28).  As Derby score less than 1 goal per game away from home we can assume that away goals will be minimal.

Derby have conceded 26% of their goals between 16 – 30 minutes, however City have scored the majority of their goals in the first 15 minutes of the second half.

A massive 36% of goals conceded by City are home are scored between 16 – 30 minutes and this combined with goals between 61 – 75 minutes account for 64% of goals conceded.

Of the 6 home games that City have scored first in, they have gone on to win them all.  When the opposing team scores first, City are more likely to go on and lose (67%).

Derby haven’t lost an away game this season after scoring first, winning 2 and drawing 2.

Hull City have gained more Points per Game (2.76) after scoring more than 1 goal than any other team in the Championship.

Dangermen

Theo Robinson is Derbys’ main threat, being their joint top scorer on 8 goals with Jamie Ward. However, 4 of Robinsons goals have come away from home.

Without Sone Aluko, Jay Simpson is Citys’ next top scorer at home with 4 goals

Citys’ defence will be under pressure from a Derby side that have scored 11 goals in their last 6 games, with City managing only half that over the same period.

Based on recent form, there’s very little between the 2 teams.  If specific home and away data is to be believed, then City are likely to record a narrow win.

City have won 4 out of the last 7 games v Derby, drawing 1 and losing 2.  Interestingly, neither Hull City nor Derby have scored more than 2 goals in their previous 7 meetings.

Hull City v Brighton – By numbers

Hull City take the long trip down to the Amex this weekend to play Brighton so I’ve had a look at some of the numbers to try and find a statistical way of winning the game.

First Goal Counts

Brighton have conceded first 13 times this season, Hull City 11 times.  As Brighton haven’t yet won a game after conceding first it would be fair to assume that the first goal will be very important.

If Brighton score first a draw is unlikely.

City have scored first in 8 of their 15 away matches so far, going on to win 7 of them and drawing 1.

Brighton haven’t lost a home game when they’ve scored first (5 wins and 2 draws).

When to score

The charts below show Brighton at home & Hull City away.

The charts show that the best time for Hull to score is between 16 & 31 minutes as this is the time that Brighton concede most and City score most.  Brighton are unlikely to concede in the first 15 minutes but should be vulnerable between 61 & 75 minutes.

Chart 2 shows that, in general, Brightons’ most dangerous goal scoring times are when City are defensively strongest although during the first and last 15 minutes City will need to be on their game.

Brighton have scored in each of their last 7 games whilst City have only have failed to score in 1 of their last 4 away games.

Hull Citys’ most common score away from home is 1-2 (26.7%) whilst Brightons most common home scores are 2-2 and 1-1 (14.3% apiece).

Clean Sheets & Failed to Score

Very similar results for Failure to Score, however Brightons’ Clean Sheet count at home is almost 10% better than Citys’ away count.  Interestingly (and spookily) City score exactly 26.7% in both categories.

Overall performance v League Average

Dangermen

The obvious dangerman is Craig Mackail-Smith who has scored 11 goals with a scoring rate of better than a goal every 2 games (1 every 177 minutes).  He has, however, only scored 1 goal in the last 7 games.

Robert Koren is Citys’ main threat as he has scored the most away goals (4).

The overriding message from the data is that whoever scores first has the best chance of winning, and in any case Hull City need to keep it tight for the first 15 minutes.  Whilst statistics can help spot patterns and offer prediction models, they are of course just a tool to help with decisions.  And they don’t take into account the curse of being live on Sky Sports.

The Next 6 – A statistical comparison

In my opinion the next 6 games for Hull City may well be the key games of the season, with a number of questions being answered:

• Can they win consistently without Aluko?
• Will the 2 new ‘Pharaohs’ deliver what is required / expected?
• Will a striker ever score again?

In order to gauge where we are against our next 6 opponents I have had a look at various data & statistics, predominantly based on the form of the last 6 games.

First though, a view of the head to head table from games so far this season:

Clearly a good recent record against the next 6, the only loss being the early season game v Blackburn.

Form overall

The current form team out of the next 6 are Blackburn who are actually 4th in the form table.  City are in 9th place in the full form table with Charlton one place ahead and Derby one place below.

This suggests that, based on form, the games v Bolton, Birmingham & Brighton should be comfortable, however, the other 3 will be much harder to call.

Goal Difference

Hull City lead the way in terms of Goal Difference for the season, however, Blackburns’ recent results give them the advantage in the form table.

Don’t watch Hull City if you want to see goals: Only 71 goals have been seen in their games so far, the lowest in the Championship

Average Goals Scored

Not very pleasant reading for Hull City fans.  Hull City and Bolton are the only teams that are scoring less than 1 goal per game currently.  This chart suggests that, for 5 out of the next 6 games, City will have to score 2+ goals to achieve a win.

Average Goals Conceded

Not only have City got the best defensive record over the season, their defensive form is even better.  Only Blackburn have conceded less than 1 goal per game in their last 6 games.

Points Won – Home v Away

The chart below shows Citys’ points gained for both the last 6 home games and last 6 away games.  Only the relevant form (home or away) is shown for opponents.  Obviously for comparison purposes, use red v blue and blue v red.

Whilst City have only taken 8 points from a possible 18 at home, none of the next 6 have better away form than Citys’ home form.

Relevant Goals scored & conceded (average)

The final chart shows goals scored & conceded for relevant home or away form only.  City have scored the same number of goals away as both Brighton & Bolton have at home, however they have conceded fewer.  The story for the home games is slightly different as 3 out of 4 teams have better away scoring form than Citys’ home form, although, again, Citys’ defensive record outperforms the competition.

Readers will react differently to the data.  Some will take the many positives that statistics suggest, others will read only doom & gloom.  Others won’t care as the only stat that matters is the points at the end of the game.

There is nothing to suggest that City can’t replicate the first table, and I would be happy with 14 points from the next 6 games.  Blackburn are coming in to some form and I think the Derby game will be tough, however, there are many unknowns yet to be established that can have both positive & negative influences on games.

*data courtesy of http://www.statto.com

The Hunt for the ‘Proven Striker’

Anyone who follows me on Twitter will probably know that I’ve got a real, albeit irrational, problem with the phrase ‘proven striker’.  Throughout January a recurring theme on the #hcafc hashtag, and I assume on other teams hashtags, was the need for a proven striker.  Whilst I understand the sentiment, it became clear that most people can’t actually define what they mean by proven.

Even though its a widely used phrase throughout football, often used in the media, including the Hull Daily Mails’ very own Phil Buckingham (@PJBuckingham), what does it actually mean?  How do we quantify ‘proven’?  What’s the difference between a proven striker and a striker?

And more importantly, do they actually exist and how easy are they to sign?

Having asked the question “What is a proven striker?” on twitter I received a number of responses.  Some examples were:

• “Consistent goalscorer at a particular level” @promooch
• “Someone who has excelled at a level for a consistent number of seasons” @tigernelo28
• “(someone who) scores goals, knows where the net is and someone you feel can score a goal” @tockaaa

All seem perfectly good explanations but none actually enable us to quantify or prove ‘proven’.

@promooch expanded a bit further by saying ” I suppose the stereotypical that we look for is a ‘consistent 20 goals a season’ striker, that’s what we all say, isn’t it?”.

So now we have a measurable number of goals.

@brunobrookes definition, “A player with a track record of scoring in the Championship, perhaps a ratio of 1 goal per 2 games over 2/3 seasons”, now gives us a ratio and a timespan.

Therefore I am going to use the following definition as the starting point for finding ‘proven strikers’:

A player who has scored 1 goal every 270 minutes, with a minimum of 15 goals, over the past 2 seasons (2010/11 and 2011/12)

By increasing the goal ratio to 1 in 3, reducing the required number of goals to 15 and only looking at the past 2 seasons I hope to achieve a higher number of candidates for each league.

Ah.  Not a huge choice for the Football League teams to go at then.

Take into account that Holt & Van Persie would be hugely unlikely signings for a Football League side and neither Jordan Rhodes or Billy Sharp can play for another team this season, that leaves 2 ‘proven strikers’.

So, lets cast the net a little wider.  A ‘proven striker’ now becomes:

A player who has scored a minimum of 15 goals last season and a minimum of 10 goals so far this season.

The players that qualify as ‘proven strikers’ using this, more relaxed, definition is below:

Of the final list, I would suggest that there would have been very little chance of 6 of the 11 ‘proven strikers’ in English football moving clubs in January, never mind moving to a Football League club.

To summarise, there are very few ‘proven strikers’ in England and even less ‘available’ for clubs in the bottom 3 divisions.  There are approximately 2400 players in English football.  Roughly 450 of these are strikers.  2.4% of these are ‘proven strikers’.  Proven strikers are very rare and very hard to sign.

When you are tweeting for your club to sign a ‘proven striker’ have a look at the 2 lists above.  Which would you want your club to sign?  Which could they realistically get?

Maybe a better request would be for a striker with potential or a striker on form, rather than a ‘proven striker’.

And would you want one of the names at your club anyway (you can decide which one I mean)?

 

*valuations courtesy of transfermarkt.co.uk

What If……Pele signed for Hull City?

With Hull Citys’ desperate search for a striker still fruitless, being seemingly overlooked by strikers they wouldn’t normally look twice at, I started wondering where Hull would be should they have had a (fit) quality striker in the team. How about one of the most top quality strikers to ever play the game. With one of the best goal to game ratios ever.

Surely that would see us winning the league easily……

And so, Hull City signed Pele in July.  And i’m not talking about old, viagra-peddling Pele, I’m talking Pele in his pomp.  The Pele that, in Escape to Victory, had the line of the century;

Give me the ball here. I do this, this, this, this, this, goal!

Peles’ 620 goals in 661 appearances is a phenomenal return, coming in at 1.06 goals per game.  What if we applied these goal statistics to Hull Citys’ season so far?  Would there be a marked difference?

Obviously, by signing Pele, someone has to make way. Jay Simpson will be the unlucky one in this scenario and therefore any playing time that Jay Simpson had will be transferred over to Pele. This means that any goals that Jay Simpson has scored this season will be erased from the results and the goal difference altered accordingly.

Having calculated the exact minute that Pele will score his goals, based on 1.06 goals per game, his goals will be added to the relevant games final score, assuming that Pele (Jay Simpson) is on the pitch at the calculated time. If he isn’t, the goal will be discarded.

The table below shows what the result of each game would be if it was Pele and not Jay Simpson playing:

As we can see, whilst the goal difference is dramatically improved the points gain is only 3 points. But why?

The main reason is simply, timing.

For example, Simpsons goal v Brighton had to be deleted, however Peles’ first goal doesn’t come until 5 minutes into the second game (1.06 goals per game). This meant that a win turns into a draw and 3 points into 1.

Another example would be the 1-0 defeat by Cardiff. Pele was actually scheduled to score a goal in the 81st minute, however he (Jay Simpson) was substituted in the 56th minute so wasn’t on the pitch to score it.

Whilst this article is just a bit of fun, it does actually raise a few questions:

• What would happen if Nick Proschwitz was given more playing time?
• Is his lack of goals purely a result of timing?
• Would a 20 goal a season striker make a significant difference to overall points gained?

The answers will differ from fan to fan, from debate to debate.

My answer is to give Simpson the ball, then he’ll do this, this, this, this, this………

What If……Substitutes Didn’t Exist?

At around 10 past 4 every Saturday, at football grounds around the country, if you listen very carefully, you’ll hear a variation of the following call….

Time to bring Big Mick/ Little Mick/ Quick Mick/ Fat Mick on…we need a goal

….the fans knowing that Mick (whichever incarnation of him that your club has) will score the goal that will secure the points.

But will he?

During Euro 2012 I wrote a blog (here) about the use of substitutes and whether they actually make any difference to the final score.

Having analysed the use of substitutes during the group stages I found that, in the vast majority of cases, the answer is no.

Only 5.4% of substitutions resulted in an ‘impact’ – for the sake of this piece an impact is defined as a substitute having a direct, points changing, influence on the final result, e.g. a goal that changes the distribution of the points at the final whistle (think Jesus Navas’ introduction in the Croatia v Spain game).

What if substitutes were taken out of football?  Would there be any change to league tables?

I wanted to assess this in the most basic manner possible (a more in depth analysis will follow) and realised that this could be as simple as removing substitutes goals from the game.  So, using the nPower Championship, that’s what I did.

Before we go on, a few headline figures on Championship substitutes:

• 1783 substitutions from a maximum possible of 1980 have been made in the first 28 weeks of the league (90%).
• 111 goals scored by subs
• 6% of substitutes overall got on the scoresheet
• 8.76% of all Championship goals have been scored by substitutes
• On average, you are likely to see one of your subs score a goal once every 6 games.
• Blackpool ar the best (luckiest) team when using subs (13 goals from subs so far)
• Nouha Dicko is the most lethal sub so far, all 5 of his goals coming from the bench. 2 having a 3 point impact on the final result and 1 having a 1 point impact.  The other 2 were scored in games that were already won.

Now i’m going to remove all substitutes goals from the Championship.  The starting point is the actual Championship table, as of 24th January 2013:

Next I went through all of the fixtures, removed any goals from subs and worked out the variance in overall points and goal difference.

I then redrew the table with the above changes included.

As can be seen from the new Championship table, if substitutes were removed from the game, the biggest losers would be Blackpool (dropped 7 places) and Middlesbrough (dropped 5 places) and the largest benefactors of this system would be Charlton and Ipswich, climbing 6 places apiece.

The alternative way of looking at the above table is that Blackpool and Middlesbrough are, in reality, the most effective at using substitutes.  The collected data (not displayed) shows that substitutes are more likely to score against Peterborough (13), Ipswich (10) and Charlton (9).  Huddersfield, Millwall & Leicester are the least likely to concede a goal by a substitute, conceding only 1 apiece.

It is worth pointing out that only Ipswich and Sheffield Wednesday have had no goals from substitutes so far this season and Peterborough subs have only scored once.

Hull City, Nottingham Forest and Blackburn are the only teams to have all conceded the same amount of subs goals as they have scored, weirdly, 4 each.

Interestingly though, substitutes have only had a points gaining impact in 6.2% of Championship games, which is very similar to the figure during Euro 2012.  The remaining 93.8% of games have either had no goals from substitutes or goals that subs have scored have had no relevance to the points at the end of the game.

Of course, those teams that have the bigger squads and better strength in depth should perform better at substitutions.  Blackpool & Middlesbrough have squads of 36 and 31 respectively and Ipswich only 25.  It might be argued that, whilst Blackburn and Huddersfield have both gained the same amount of points from substitutes impacts, Huddersfield are overperforming with a squad size of only 25 and Blackburn are underperforming with a squad of 35.

As mentioned earlier, this is a very basic way of looking at substitutes impacts, but the next time that the supporter near you is screaming at the manager to bring on Mick, you can confidently point out that Mick has a very slim chance of scoring the winning goal or the goal that will secure a draw.  Roughly a 1 in 17 chance.

Poor Mick.

Hull City – Better than last year?

The quickest, simplest of answering this question is to look at Citys’ points tally at the same point last year and compare it with this year.

The comparison shows that City are just 4 points better this season than last season, right?  Wrong.

Well, wrong (ish).

Last year I read an article by Simon Gleave of Infostrada in which he developed a different way of looking at, and quantifying, how teams have improved (or not, as the case may be) year on year. He called it the ISG coefficient (his post here).

So how does it work?

Simply by comparing results from last year against the same fixture this year. Fixtures that involved teams who were promoted or relegated are replaced by the team that took their place. For example, the top promoted team is replaced by the top relegated team, so Reading v Hull City becomes Bolton v Hull City and so on.

Rather than apply the ISG coefficient to the whole of the Championship, which would take me and my ZX Spectrum weeks to do, I have only applied it to Hull City, looking for some indication that the Championship table doesn’t tell the whole story.

The results paint a positive picture.

The figures show that, whilst City have improved by 4 points in the Championship table, they are actually 12 points better than the corresponding games of last season. They have also scored 12 more goals, suggesting a more attacking style but have also conceded 4 more than the same games last term. Interestingly, 5 of the 12 points were a gift from Leeds.

Some more headline figures:

• Citys’ biggest points gains have come as a result of turning 7 of last years draws into wins this time around (+14 points)
• Only 2 games have seen last years losses result in draws this term (+2)
• City have managed on only 2 occasions to turn losses from last season into wins this year (+6)
• 4 games have seen City drop points v the corresponding game last year (-10), with Peterborough taking 5 of these alone.

Thus the answer to the question “Are City better than last season?” is, yes, but actually by quite a bit more than is instantly visible.

Finally, to predict the total Hull City total points tally at the end of the season we simply take all of Citys’ remaining games and assign the same results as the corresponding game last season.

From this point to the end of last season, using corresponding fixtures, City picked up a further 30 points. This would see us end the season with a total of 80 points.

Will it be enough……….

- As an addendum, and using some licence, we can gently (grossly) manipulate the final tally of 80 points. Currently, using the ISG coefficient, City are approximately 31% ahead of last season. By applying the same level of increase, weighted over 18 weeks (19.8%), we can assume that City will get the 30 points plus a 6 point increase.

86 points sounds much better doesn’t it.