Friday, 13 May 2016

Is it better to be patient or to take long shots?

Watching a lot of Celtic games recently got me thinking about how they play, why Ronny Deila has come under so much pressure and why they are different to many of the other sides who dominate their respective leagues.
The answer struck me when watching Barcelona versus Atletico Madrid in the first leg of their Champions League Quarter-Final.
Despite facing ten men and with Atletico sitting very deep Barcelona kept the ball and continued to try and work positions for Great and Good Chances, which have a significantly higher conversion rate than Attempts.
While I am in no way comparing Celtic to Barcelona in terms of quality, they do face similar problems when playing domestically. Opponents tend to sit deep and be very compact against them, allowing them to control the ball 30 or 40 yards out and restricting the space in behind. However, in contrast to Barcelona I had observed that Celtic are very happy to have lots of Attempts, sacrificing the possibility of creating better chances by keeping the ball.
This led to further thoughts about how Attempts might affect the Total Goals Line in games. Given the conversion rates for Great and Good Chances it means that leagues with more of these should have higher starting lines. Indeed, I would expect to see a trend where games with low lines have a higher proportion of Attempts than Great or Good Chances.
The graphic below shows the difference in the most common total goals line across 15 competitions covered by StrataBet. These competitions are generally taken from the start of the 2015/16 season to date, though Japan J League and Norway Eliteserien look at the entire 2015 season:
Patience or Long Shots Chart 1-01
From this table it is easy to see that Norway has a much higher goal line than the English Championship, meaning that more goals would be expected to be scored in their games.
If more goals are expected, then the proportion of Great and Good Chances should also be higher to account for this.
Using StrataData I investigated this theory and there does appear to be some correlation. Of the leagues with low goal lines, three of the four lowest (those with 2.25 as the common line) also appear as three of the four lowest in terms of the number of Attempts per Great Chance created:
Patience or Long Shots Chart 2-01
The outlier here is France Ligue 1, where Great Chances are created but the goal lines are low and the fewest goals per game are scored of the big five leagues.
The English Premier League and Championship also fall into this category, but in the Premier League 32% of games had a goal line of 2.25. The difference between that and the most frequent of 2.5 (37%) is a relatively small amount, making this understandable.
Austria, Switzerland and Germany Bundesliga appear to be the polar opposite of this, with fewer Attempts taken and more Great Chances created. This fits well in Switzerland, with 64% of games having a starting goal line of >2.5 and 49% of German Bundesliga games having the same.
The graphic below shows this in more detail, so that we can see the leagues where more goals are expected on a regular basis:
Patience or Long Shots Chart 3-01
Germany and Switzerland are only behind Netherlands and Norway in terms of the number of games where over 2.5 goals would be expected.
So how does this relate back to Celtic?
Looking at teams who dominate their league we can see if they typically play like Barcelona in terms of keeping the ball and waiting for better chances, or if they simply shoot often and from wherever possible.
For this I have chosen teams from the top three of all the leagues mentioned who are at least 15 points clear of the team in fourth and the table below brings up some interesting points to note:
patience or longshots table 4-01
As suspected, Celtic lead this table by quite a distance from the rest. For every Great Chance they create they have ~3.5 Attempts, which is significantly more than Barcelona, Dortmund and PSG, who have ~1.5 Attempts per Great Chance.
This does not necessarily mean what Celtic are doing is wrong, but the question would be is this repeatable? We would expect the number of Attempts Celtic have had to result in ~6 goals, though they have actually scored 14. This says that they are either lucky and doing better than we’d expect or that they are confident the quality of strike they can produce from poor positions can beat the quality of goalkeeper the opposition has.
Ultimately there is no singular proven way of playing successful football regardless of personnel. Surprisingly it is Juventus who lead this list of teams if you include the number of Attempts taken compared to both Great AND Good Chances created, despite being a fantastic side. However, what it does show is the style that teams like to play in:
patience or longshots tabke 5-01
The teams at the lower end are more content to wait for better chances to come, knowing that they have the quality to keep creating rather than taking longer shots. The teams at the higher end are apparently confident in their ability to shoot from range, which can be a useful tactic in many situations. It can keep opponents guessing and also opens up opportunities for strikers to get on the end of rebounds and to win more corners.
What should be noted here is that almost all the teams in these tables are at the top end of creating Great and Good Chances in each of their respective leagues. As such, even though they may be registering a lot of Attempts they are still creating better quality chances on a regular basis than most other teams in their league.
Of course Attempts are not always a bad thing but teams that favour these over Great and Good Chances may struggle when the standard of opposition increases, such as in European competition when chances can often be at a premium. Celtic are the prime example of this as they dominate the Scottish Premiership but have failed to make much of a dent during their recent Champions League and Europa League campaigns.
It will be interesting to see how much Ronny Deila’s successor changes their style of play and attitude towards chance creation and if this improves their fortunes on the grander stage.
This blog was originally available at StrataBet Trading Expert.

Friday, 15 April 2016

Who will make the Championship Play Offs?

This blog was originally posted on StrataBet Trading Expert (08/04/16)
The English Championship continues to be a fiercely competitive division. At the top end of the table the recent poor form of Hull and Middlesbrough has created what now looks to be a four-way fight for the two automatic promotion spots. At the bottom the recent resurgence of Rotherham under Neil Warnock has given them a great chance of avoiding relegation. Bolton and Charlton look doomed, but there are other teams who could still be dragged into the fight. Indeed, MK Dons look set to join them after poor results in their last two games.
The focus of this piece will be the race for the two remaining play-off places. Derby secured an excellent victory over Hull on Tuesday and it looks like four teams will be competing for two places, just like in the race for automatic promotion. Birmingham seemed to be in contention before the weekend but defeats to Charlton and Brighton have left them facing an almost insurmountable task, despite the fact that they have a game in hand.
The Contenders
Screenshot 1 - Current Positions
Sheffield Wednesday
After a huge improvement last season under Stuart Gray, this season saw a new and largely unheard of manager Carlos Carvalhal arrive, and with the backing of fresh investment Wednesday have been excellent this season. Their aim at the beginning of the campaign was a top half finish, but it will now be disappointing if they don’t make the top six. They have a squad high on quality but they can be overly reliant on the talismanic Fernando Forestieri, with few players capable of acting as a direct replacement. Wednesday have a very strong spine and will be the favourites to finish sixth.
Derby have history with missing out on the play-offs after last season’s dismal finish and they will be determined not to let the same happen again. They have already had a bad spell this term when they dropped from first to fifth and Paul Clement was sacked after less than eight months in charge. Despite sound bites from the club that promotion is not their primary aim they will not want to miss out, but need to get the team playing a cohesive style. They can often look disjointed and tend to struggle against more physical sides.
While Cardiff fans have recently had a taste of the Premier League they have emerged as unexpected challengers this season after cost cutting seemed to hinder any kind of push. Resentment towards owner Vincent Tan, plus a general dislike of Russell Slade and his style of play seems to have diminished slightly as a togetherness within the team has seen them only lose twice at home all season. Their lack of strikers could be their major downfall in the end, but they have been able to spread the goals around the team so far.
Having missed out on the 2014/15 play-off final after losing to local rivals Norwich in the semi-final, another shot at the top six was the aim for this campaign. Despite losing key players like Tyrone Mings and with very little investment Mick McCarthy has again worked wonders to get them competing at the top end of the division. Their style of play is still direct and aggressive, but there is perhaps not quite as much long ball play as there was last season and they still remain effective. Daryl Murphy remains their most dangerous player, but he has regressed from last season when he hit an unprecedented 27 goals, which was more than his total in the previous four years combined. Murphy’s saving grace is that he adds a lot of value to the team even when not scoring, so Ipswich will hope the injury he picked up on international duty is a minor one.
Sheffield Wednesday ~1.11
Cardiff City ~4.5
Ipswich Town ~15.0
5th-12th placed opposition – Draw
13th-18th placed opposition – Win
19th-24th placed opposition (home) – Win
19th-24th placed opposition (away) – Draw
6th – Derby County (74pts)
7th – Cardiff City (73pts)
8th – Ipswich Town (70pts)

Derby County
Cardiff City
Ipswich Town
As shown in the table below the teams face run-ins of various difficulties. On paper it appears that Ipswich have the hardest set of fixtures with three away games against current top six sides, all coming in the final five games of the campaign. Cardiff have just come through a tough spell against teams in the top six themselves and took a very respectable four points:
Upcomming fixtures2.png
It should be noted that because Sheff Wed and Derby currently reside in the play-offs they obviously count as members of the top six. Wednesday still have all three play-off rivals to face, with two of these games being at home, and it is these three games that should be most vital. Derby have perhaps the easiest run in of the four sides, purely due to their games against teams in the bottom six.
Ipswich’s trio of difficult away games sandwich two games at home against teams in the bottom three. While this may seem like a massive bonus, at this stage of the season it may actually be a huge banana skin compared to playing a mid-table side with nothing to play for. This was shown on Tuesday night when they struggled at home to a determined Charlton side.
In previous seasons 70 points has been seen as the benchmark needed to make the top six. However, this would only have seen a team make the play-offs in three of the last ten seasons:
previous points needed for 6th place.png
75 points actually seems a much more reasonable target, especially with the number of teams chasing down the play-off spots. 75 Points would be enough in nine of the last ten seasons, with only last season throwing up an anomaly as Wolves (78 points) and Derby (77 points) both missed out. These were the highest totals to do so since the Championship switched to its current format of 24 teams in the late 80’s.
Using StrataData we can look at some key indicators as to how each of the teams have performed over the season:
Table 3 CHANCE VOLUME1.png
From this we can see there are only marginal differences between the teams. Cardiff, Derby and Sheff Wed create a similar number of Great Chances (~1.5/1.6 per game), but Wednesday convert them at a much better rate. While this may be expected to decline, the small number of games remaining means that this is unlikely to happen before the end of the season.
Sheffield Wednesday actually have the best conversion rate across two of the three categories, which is a big reason for why they are the second highest scorers in the league (behind leaders Burnley). This is a real change from last season when they scored the fewest number of home goals in the division – even less than the bottom three.
Derby’s Great Chance conversion rate is the lowest of the group, but they make up for this by not only having the highest conversion rate of Good Chances, but also creating many more of these than the other sides. Ipswich’s rates are also quite good, though not at the level of Wednesday and Derby. This could be an indicator of some quality in their shooting but it could also be an indicator of an uncontrollable luck factor too.
Defensively, it is again Derby who look most impressive, giving up the lowest number of Great and Good Chances, with a low conversion rate on those Good Chances as well. Wednesday also seem to be skilled at preventing chances but the conversion rate on the Greats shows that if they concede one there is almost a 50% probability of it being scored. Naturally, preventing these situations is a must for Carvalhal’s men, but fortunately for him he has a solid defence.
Another thing to consider here is the fair outcome for each team in the table, which is basically StrataData’s “fair score” of their games. This is based on various metrics and is designed to show that if a team were to play the same game 100 times in the same situation would they be more/less likely to get something out of it. These scores are then added up in the same way a league table would be and the top four is quite representative here:
Fair Outcome Table 2 DB NOW 1.png
It seems that only Brighton are not truly worthy of their place in the top four, though even by this method they would still be 7th, which is not a big stretch giving the strength of the league. Chris Hughton’s side had a great start to the season, but drew a lot of games that they could have lost on many other days.
Using this technique to observe the four teams we are focusing on reveals a different story:
Fair Outcome Table 1 DB NOW 2
While Wednesday apparently deserve their fifth place, the other teams could be in much different positions in the table. Derby are ranked at the top and this shows that a combination of bad luck and poor form has seen them slip out of the running for the automatic spots. Cardiff and Ipswich look like they may be over-performing, as they are ranked 10th and 13threspectively. This shows that they are grinding out wins that they may not have got on any other day. However, this does not mean they are not deserving of a top six place, but that they have consistently achieved better results than they have been expected to.
This is far from perfect, but provides an indicator of which teams could be over- or under-performing. It is interesting that the other team who deserve to be in the top six is actually Blackburn who sit way down in 16th – while losing Gestede and Rhodes is not easy to cope with they have badly underperformed this season.
Mini League
One final thing to look at is the impact of these teams playing each other. We have already mentioned that there are still a number of games between the sides due before the end of the season. Cardiff have the fewest fixtures against the other teams competing for a play-off place and this could be a benefit, as if they do win when two of their rivals compete they have an increased chance of taking advantage of both dropping points at the same time.
Of course these games are destined to have a significant impact on how the race for sixth finishes up. If Wednesday were to lose their three games against close rivals but win the three other it would likely see a much tighter finish than they would like. As such, picking up points in these matches is key.
Previous results between these teams this season look like this:
Table 4 mini league
As there is quite a difference in how many games they have played against each of their rivals it should be noted that Sheffield Wednesday have picked up less than a point per game against the teams around them. Even with three matches to play this is not a positive statistic, especially when you see that they have also failed to take any wins from the top four sides as well. Derby look understandably strong and with two home games to come they will be looking to further strengthen their position. It is often breaking those teams who look to defend down that has been their undoing, so playing against opponents that might need to go for a win will surely help them.
Who has the edge?
The market has Derby and Sheffield Wednesday as clear favourites (odds as at 6/4/16):
Derby County ~1.08
These prices seem quite reflective of the information that has been covered here so far. As an example, Sheffield Wednesday are likely to need around seven points from their final six games to reach the previously mentioned 75 points. This could be achieved with wins against the two teams in the bottom six that they are due to play, leaving them needing just one point from the remaining four games.
Let’s finish with some rough estimates and assume the following:
1st-4th placed opposition – Lose
This leaves the table at the end of the season as follows:
5th – Sheffield Wednesday (76pts)
There could be some value in Cardiff as they are close to the teams above in the projected final table despite their much longer odds. Indeed, a win for Slade’s team away to Sheffield Wednesday would leave both clubs on 75 points.
We will have to wait until the 7th of May to see how accurate this is but there are sure to be plenty of twists and turns before then. Knowing The Championship, it will certainly be an exciting end to the season.
Dave Willoughby (@donceno)

Friday, 11 March 2016

Are there goalkeeping problems in the City of Discovery?

I recently wrote a blog about how Celtic had been performing this seasonand a small part of it focused on Craig Gordon. It turned out that the Hoops goalkeeper had the third highest save percentage in the division at the time of writing, behind only Aberdeen’s Scott Brown/Danny Ward combination and Neil Alexander of Hearts:
A throwaway sentence within that piece sparked a Twitter exchange that found its way to Dundee goalkeeper Scott Bain, a comment he has since deleted. He was understandably unhappy at being called out for having the second lowest save percentage in the league.
Funnily enough Bain has just been called up to the Scotland squad for the friendly with Czech Republic thanks to his performances this season. He could even make his debut this time, having not featured when last involved with the squad in May 2015.
I should start by saying that I think goalkeeper is one of the hardest positions to analyse. Using statistics to judge players is still in its relative infancy and it can be difficult to know which statistics are useful and which aren’t. Indeed, goalkeepers often have so few interactions that their individual data sets are quite small.
Key elements of being a goalkeeper such as organising the defence, commanding the box, sweeping behind the back line and judgement when catching or punching crosses are hard things to rate. The analytics community often decries this, but at times these so-called intangibles can be just as important as the performance elements that can be measured more easily.

The rest of this blog can be viewed for free at - sign up to StrataBet is also free, with paid packages available

What has gone wrong at Celtic?

This may seem like an odd question.
After all, Celtic are three points clear at the top of the Scottish Premiership with a game in hand. They are also well placed in the Scottish Cup; with a quarter-final against mid-table Championship side Greenock Morton to come. Indeed, they might have progressed to the final of the League Cup too but for an early red card in the semi, yet Ronny Deila is under severe pressure.
Celtic have seemed to underperform constantly. The perceived difference in quality between them and the rest of Scottish football means that many people expect them to win the league at a canter, triumph in both cups and most importantly make at least the group stage of the Champions League every season.
The main issue for Deila appears to be the improvement of Aberdeen. In fact in 2014/15 Aberdeen actually had a better record against the other 10 teams in the Scottish Premiership before the title was won:
It is common to hear that since Rangers’ demotion Celtic do not have any genuine competition. However, this is clearly not true at the moment, with Aberdeen capable of taking points off Deila’s side as well as the others in the division. Already this season they have taken six points from Celtic and if they can replicate their form from the rest of 2014/15 against the other clubs they would have a great chance of pushing the Glasgow giants all the way.
The rest of this blog can be viewed for free at - sign up to StrataBet is also free, with paid packages available

Tuesday, 2 February 2016

How unpredictable is the Championship?

The English Championship has a reputation for being a very competitive league where it is particularly difficult to predict the correct scores of games.
There is generally quite a lot of parity across the board and while you of course still get favourites and underdogs the gap is rarely as large as it is in the top divisions of England, France, Germany, Italy and Spain.
As an example, Middlesbrough were heavily backed for promotion in pre-season and have performed well, even going on an impressive nine-game run where they did not concede a solitary goal. This run was surprisingly broken by Bristol City, a team sitting third-bottom at the time. City actually managed to complete the double over Aitor Karanka’s high-flyers in the process.
The Championship has been like this for some years now and in 2012/13 the gap between relegated Peterborough and promoted Hull City was just 25 points. To compare, even this strangest of Premier League seasons already sees the gap between Newcastle United (third-bottom) and Leicester City (top) at 26 points.
To expand on this we can estimate some level of supremacy by looking at the Asian Handicap lines in the Championship.
In only 22 of 340 games played so far has the AH line been at least -1 or +1, which is just 6.5% of the time. The Premier League has already seen the same level of supremacy in 54 of 223 games, which is a relatively enormous 24% of the time in comparison.
Using the Brier Score method, I am now going to look at how many “shocks” have occurred in the Championship this season. The Brier Score uses a function that observes the accuracy of probabilistic predictions.

For the rest of this blog post please see the full version on the StrataBet Trading Expert page - along with plenty of other insightful pieces.

Friday, 8 January 2016

Do away teams have the advantage in Scotland?

Home advantage is generally taken for granted in football. In most cases no matter where they are in the table the home team will always feel they have a chance, even in a top vs. bottom scenario.
However, during this season the home advantage of some leagues seems to have experienced a significant drop.
As a result of this I decided to look at all the leagues that StrataBet cover, paying particularly close attention to the Scottish Premiership.
To begin with, the graph below shows the % of points won at home vs. % of points won away (data correct up to 04/01/16):

I principally cover the Scottish Premiership and it appeared to me that many of the teams seemed to do better when away.
To view the rest of this blog please see view the StrataBet Trading Expert - sign up is free and there are plenty of other useful blogs as well as previews for games, models and data available for the leagues covered by StrataBet.