Does seasonality exist in football?

In this piece I want to consider whether seasonality affects the number of goals scored and whether we can use this information to bet more intelligently on football matches.

Historically, the issue of seasonality in goal scoring rates has always been a factor to consider for bettors.

Typically we could expect more goals to be produced when pitches are in top condition and players are at peak fitness.

Alongside these two conditions, the importance of individual matches can also influence the number of goals scored. For example, it is a fair assumption to make that there are fewer goals scored in matches involving two title contenders or a pair of relegation candidates. It is also fair to suggest that more goals are scored in meaningless fixtures towards the end of the season, especially in matches between teams with little or nothing to play for.

Because pitch conditions are now considerably better season-round in the main European leagues, I looked into the number of goals scored in each month across the four “major” leagues (English Premier League, German Bundesliga 1, Italian Serie A and Spanish Primera Division) to see if the theory of seasonality still holds.

Although I have simply considered the number of goals scored in each calendar month, there are a number of different ways to test for seasonality. We can analyse the number of goals scored in each individual round of fixtures, in every four rounds, or even split the football season between the international breaks that are held at fairly consistent times of the year.

Each of the graphs below shows the average percentage of goals scored in each month compared to the total number of goals scored for that period. So, if there was an average of 2.8 goals scored in November over a ten-year period and the combined average number of goals over ten years was 2.8, the figure for November would be 100%.

I have also added in “Final Round” averages, which are the results from the last set of fixtures in a league season and traditionally see more goals scored:

Seasonality

The main takeaway from the graphs is that goal scoring rates begin at lower than average levels in August, before increasing around October as each squad presumably hits peak match fitness. There is then a dip during the business end of the season around February/March when there are more meaningful “must-win” matches being played.

Bundesliga 1 is the only league here that has a complete winter break and we can see from the graph that the scoring rate takes longer to return to average, almost replicating the phase at the very start of the season.

What is common across all four leagues is the vast increase in goals scored at the end of each season, especially in the final round of each league.

This can be explained by the amount of meaningless games that occur at this time, with titles won, relegations suffered and European places often sealed before the final set of fixtures take place. This then gives each team more freedom to play more relaxed football, which can breed creativity, cause mistakes and ultimately lead to more goals.

Germany, Spain and Italy’s top divisions all show an almost identical pattern over both the five- and ten-year time periods, though there is a small difference in the numbers for England’s Premier League during the middle months of the season.

To a degree this can be explained by the increase in average goals scored in the past five seasons compared to the last ten.

The ten-year average for this league is 2.66 goals, but the smaller sample of five years has a much higher average of 2.75. So far this season, the number of goals is back down to 2.64 and it may be the case that we are going through a low goals period in this league, having enjoyed a boom between 2009/10 and 2013/14.

In 2015/16 to date, Bundesliga 1 and Serie A have both followed their five and ten year period pattern well, indicating a normal season in these leagues.

However, there has been some difference in the Primera Division, where the goals in August were exceptionally lower than average, and in the Premier League, where we saw a higher number of goals in September but a significantly lower number in December.

In drawing conclusions from the data, it is fair to suggest that seasonality does still exist in the four major European leagues and I think this pattern will also hold for most other leagues around the world. Although we may see some anomalies, it remains my firm view that seasonality should always be considered when betting on football matches.

Mark McAfee

Total Goals Trading: One Bettor’s Process

This piece is designed to provide a series of criteria that can be used in combination with computerised models and regular statistics to assist betting selections in the Total Goals market.

I will focus predominantly on Spain’s Primera Division because of lengthy experience of covering this league, but the concept and thinking behind the blog should be transferrable to any league under StrataBet’s umbrella.

In order to keep things simple, when the terms under and over are used, they always refer to a goal line of 2.5. However, much of the analysis herein will cover thought patterns that are applicable to more or less any goal line.

The first step is to categorise all twenty Primera teams as either under or over teams. These are the ones I think most likely to have fewer than three goals or more than three goals in their matches.

The key question when making this selection is: “If this team’s tactical plan goes perfectly, would the end result see over or under 2.5 goals?”

When answering this question, I have considered the objective of each team going into a normal league game, but I have ignored whether that team might be home or away.

Here is my selection, with explanations where necessary:

Primera Div Over

*Valencia are in this list because of my view of their style since Gary Neville was appointed. They currently look like one of the most attack-minded teams in the division and have a weak, error-prone defence.

**I feel I must justify Las Palmas appearing in this list based on their strong unders bias this season. They are here due their recent change of manager, who is less set on playing for clean sheets. Indeed, this is a team that now wants to entertain and score goals.

Primera Div Under

To ensure the best possible chance of winning in this very tricky market, it is important to ask a series of questions about both teams involved in a game you wish to bet over or under on.

These questions should give you a good idea of how a game is going to pan out and help you determine the ideal scenarios for more/fewer goals.

Here are the questions I ask myself, along with a brief description and/or example of why they are important:

Do both teams have a desire to win the game?

On the face of it this looks like a silly question, but often one team will be more than happy to settle for a draw.

If a team begins a game with the aim to keep it at 0-0 for the duration, or even just for as long as possible, it shouldn’t necessarily put you off betting overs. However, ideally if you bet this way you want both sides looking to win the game.

Conversely, if you are on unders then you want both teams to fear losing and in particular to fear conceding the first goal.

Where are the goals coming from?

Matches most likely to have overs are those in which one team that struggles to defend and keep clean sheets faces one that has good finishers.

Teams that defend well, like Atlético Madrid and Villarreal, are typically risky bets to have when betting overs. Both of these teams focus their plan on stopping the opponent and as they are so effective, particularly Atlético, I would usually avoid considering them for overs.

In the Primera Division there are plenty of good finishers, but also lots of players who tend to thrash at chances due to a lack of composure and/or technical ability.

Naturally, the more names you see in a projected or actual starting XI that you feel confident can finish chances, the better you should feel about backing overs. This is especially important when the mutual respect between two teams is high, because often in tightly contested matches the goals come from great finishers

What happens if Team A or Team B scores first?

This is another vital question to assess how a game is likely to pan out. While it relies on a large amount of speculation, in any game you must look at whether you envisage the game being a high or low scoring one.

If Team A scores first are they going to sit back and defend? Indeed, most teams outside of Barcelona and Real Madrid will do this to an extent, but how effective are they with this approach?

Let’s look at Atlético Madrid again. They are an incredibly good team at scoring first and killing the game to win to nil. They are so effective at this I wouldn’t even consider betting overs in any of their games prior to kick off. Villarreal are also strong at protecting a lead and winning to nil.

It’s also important to consider how effective both teams are on the counter-attack here.

If it’s a game where mutual respect is reasonably high but both sides can counter effectively, there is real benefit to overs backers. This is why strong counter-attacking teams like Sevilla and Villarreal, two teams that I deem as unders ones, often have games which ultimately end up having over 2.5 goals.

There is only a small group of clubs who I do not rate as counter-attacking dangers right now: Athletic Club, Getafe and Real Betis. These teams all have little to no pace in their attacks and tend to focus on killing time or keeping possession to protect leads.

How much do the two teams respect each other?

This question is important when considering overs or unders, but if the answer is no for both teams, then it is a big bonus if your model (or your gut instinct!) likes the overs. When there is little-to-no mutual respect between teams, they are likely to attack each other instead of focusing on defending, which is what would be more likely to happen is there was great respect between them.

This can be a great angle to get good value on overs when the smaller clubs play each other.

It clearly does not mean that goals or even an open match is assured, but it is much more likely that the game will open up with the first goal, which will ideally arrive in the first half.

Conversely, if there is a lot of mutual respect between two teams, this should generally deter an overs bet. This also applies when discussing goals during derby matches and specifically those that come in circumstances where neither side wants to lose. Put simply, when the fear of losing outweighs the desire to win it can be a good signal to go with an unders bet.

What does each teams’ immediate schedule look like?

If there is another significant fixture just days after a league game, this can affect the predictability and probability of goals.

The rotation risk is the most obvious factor and will depend upon how the club/manager weighs the following game versus the current one. Even if rotation does not occur there is a danger of a team not wishing to expend too much energy during the ninety minutes, which could negatively impact on goals and favour an unders bet.

As always knowledge is power here, and you can improve your analysis by checking the projected line ups offered by StrataBet’s Previews, both three days and one day before kick-off.

Will one team try to kill the game if they go ahead?

With the exception of Barcelona and Real Madrid (when not playing each other or another top four side), teams in Spain are prone to reducing their attacking output in favour of wasting time, especially in the final games of the season when points really matter.

These tactics are most likely to be employed by teams struggling at the bottom of the table and the main culprits are currently Granada and Real Betis, followed by Deportivo La Coruña, Getafe and Sporting Gijón.

Atlético Madrid and Villarreal are teams who can focus almost exclusively on defending after taking the lead. Athletic Club and Deportivo La Coruña do this as well.

It is important to understand that it is much easier to kill time in Spain because free kicks can be earned with a minimal amount of contact and officials are often slow to recognise various other methods of time wasting.

This is clearly a big negative for overs backers, but a big positive for those on unders. A home team wasting time can often be gold for an unders bettor.

Other Trends

To close I will inform you of some clear patterns and approaches of certain teams that are relevant to the number of goals scored in their matches.

Home/Away

Tactical approaches change dramatically for most Primera clubs when they play at home, compared to when they play on the road.

Indeed, a team who are very bold at home will often set up more conservatively when away, which may favour unders bets in certain circumstances.

Almost all clubs in the division fall into this category, perhaps excluding Barcelona, but the biggest differentials I personally observe are Eibar, Espanyol, Málaga, Sevilla and Sporting Gijón.

The Big Two

In games involving Barcelona and Real Madrid the goal line will normally rise significantly, depending on their opponents and whether they are at home or away.

Both clubs generally seek to win every game comfortably and play an attacking style, but away from home there is always more tendency to tighten up defensively. Barcelona have been particularly effective at doing this away from home. Naturally, opposing teams who can find a way to breach the defences of the big two are great for beating the goal line in an overs bet, rather than relying completely on the favourite to score four or five goals without reply.

Teams that are strong on set pieces or have composure/speed in attack are ideal for getting goals against the big two. Both Barcelona and Real Madrid tend to switch off and leave spaces in defence, particularly in games they know they are likely to win well.

Barcelona and Real Madrid drop their intensity in almost every game they are leading comfortably, which gives the opposition a chance to have periods of attacking and obviously increases the probability for goals.

Pitch Quality

Poor pitches where the ball does not roll predictably can cause problems for attacking football and tend to be bad for overs outcomes. Celta Vigo, Espanyol, Las Palmas and Rayo Vallecano all have below-average/poor pitches where ball movement is less predictable and teams need to work a little harder to create.

Las Palmas’ pitch is easily is the worst in the division and this has clearly affected the number of goals scored at the venue this season to date.

Managerial Change

New managers can bring fairly dramatic changes in levels of commitment, personnel performance and tactics.

At Real Madrid, a team that has now scored eleven goals in two matches under Zinedine Zidane, you can now expect extremely offensive and high scoring matches at home, but a much more controlled approach away.

Seemingly every player (with Gareth Bale a notable exception) has made it known that they prefer Zidane to predecessor Rafa Benitez, which confirms prior suspicions that the players weren’t giving maximum effort under the Spaniard. With that in mind, this is a team who should continue to look better and improve results from the first half of the season. Real Madrid are something of an exception here, because managerial moves tend to favour unders betting initially.

We are likely to see this with Real Betis, the most recent club to change manager (at the time of writing). In the short-term, I would expect them to focus on becoming stronger defensively. As they are already a team who struggle to score goals, they should be even more unders biased while the new manager tries to rebuild confidence with a safety-first approach.

At Las Palmas and Valencia, both clubs are seeing the new managers bring about a much more expansive approach. However, Valencia in particular are paying the price defensively. They were one of the most unders biased sides at the start of the season but with Neville in charge they look very open and happy to take risks, especially when at home.

Despite the aforementioned poor surface in Las Palmas, I would be actively looking to back them on the overs both home and away in the near future. My reasoning here is that there is immediate market value to be had after all but one of their home matches went unders in the first half of the campaign.

In addition to this Las Palmas have generally become much more of an overs side in the small period of time since new manager Quique Setien arrived. One reason for this is that previous manager Paco Herrera was terminated because the fans hated his defensive tactics. So now we have a Las Palmas team with an error-prone defence trying to play open, expansive football. More than most sides in the Primera, they are happy to risk a lot when going behind and as a weaker team they go behind often!

To conclude, I hope that this piece gives an insight into one bettor’s thought process when attacking the Total Goals market in football. As I said in the opening, while everything here has been very Spain-specific, many of these principles should apply to other StrataBet leagues very comfortably indeed.

Good luck if you decide to venture into this notoriously difficult market!

Dan King

For access to the in-depth match data featured in this blog, betting recommendations and much more visit StrataBet to get a 30 day free trial.

Has there been a “Klopp effect” at Liverpool?

It has now been three months since Liverpool decided to replace Brendan Rodgers with Jürgen Klopp.

This move instantly transformed the atmosphere around the club and was welcomed by most supporters. However, since a trio of high scoring wins away to Chelsea, Manchester City and Southampton the shine has begun to wear off a little.

So has Liverpool’s move paid instant dividends, as the market suggested it would when the price on a top four finish spiked following Klopp’s appointment?

I undertook some further investigation from a purely statistical perspective to try and answer this question. It is important to note that only league data is considered.

Klopp vs Rodgers 1

In terms of a direct comparison, Liverpool won 37.5% of league games played under Rodgers this season and 41% under Klopp, although the German has now overseen four more games.

Despite this slight incline, the goal output of the side has actually risen only very slightly since the change in manager. Rodgers’ side scored an average of 1 per game, while Klopp’s apparently “more attacking” side is registering 1.1 per game.

Another marginal sign of improvement under Klopp is again evident in the goals against column. Rodgers was conceding an average of 1.25, whereas the number drops to 1.16 under Klopp. This is hardly a monumental shift, but in correlation with the attacking output it is definitely a change for the better.

To get a clearer picture of how things have changed since Klopp’s arrival it is appropriate to look at deeper metrics like chance creation, to see if there really has been an alteration in style.

Klopp vs Rodgers 2

Breaking it down into Great and Good Chances, we can actually see that Liverpool were more productive under Rodgers, creating on average 1.5 Great Chances and 2.6 Good Chances per game.

They did have some injury problems with strikers early in the season, so the poor conversion rate may be affected slightly by that. However, the same is also true for Klopp, who has arguably been hit harder with injury problems since taking charge. His side have averaged just 1.3 Great Chances and 2.4 Good Chances per game so far.

This points to Liverpool being a more efficient side under Klopp. They are winning more points by converting fewer chances, which signifies another minor improvement under the new manager.

This seems to be the main difference in output, because there is very little between the two managers in terms of chances conceded:

Klopp vs Rodgers 3

Under both managers the team has given away 1.5 Great Chances per game, although Rodgers saw his side afford opponents slightly more good chances at 1.8 per game. Unsurprisingly this correlates with the average goals per game conceded.

However, it is notable that Liverpool have conceded more goals than they have scored, despite both managers seeing their side create more chances than they concede, but is this down to poor attacking or poor defending?

Indeed, there is evidence to suggest that both are to blame.

Under Klopp, Liverpool average 4 shots on target per game, very marginally better than Rodgers’ 3.9 per game:

Klopp vs Rodgers 4

Once again this shows a slight improvement in shot conversion under Klopp, but it really is very minimal.

Klopp vs Rodgers 5

Defensively, Klopp sees his side give away just 2.5 efforts on goal per game, whereas under Rodgers 3.25 shots on target were being allowed.

Going back to the goals conceded, it means that Liverpool are conceding at a higher rate under Klopp than Rodgers, with opponents finding the net from just under half of their shots on target.

Liverpool are actually the third strongest team in the league when it comes to chance prevention, with only Arsenal and Tottenham giving away fewer opportunities per game. If they can somehow slow down the rate at which they concede goals from chances allowed, their outlook will greatly improve for the rest of the season.

So where does this leave Liverpool moving forward?

With just five teams having scored fewer goals than Liverpool their top four prospects do not look great, especially when all of those sides sit below them in the table. The fact that they have not conceded many compared to others in the Champions League race will give them hope, though.

Overall it seems that while there has been a very slight “Klopp effect” at Anfield, Liverpool’s metrics still make them an outside bet to achieve a return to Europe’s premier club competition this season.

However, it still looks as though the German will be a shrewd long-term appointment at Anfield. He has overseen a slight upward trend despite very limited playing resources. Some January incomings should help the new man continue this by putting more of his stamp on things.

Alec Payne

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):

Graph 1 update

I principally cover the Scottish Premiership and it appeared to me that many of the teams seemed to do better when away.

St. Johnstone lost against Motherwell at the end of December but it was the first time they had not won away since the end of September (winning five in a row, though they did lose four in a row before that).

Inverness Caledonian Thistle picked up a significant away point at Aberdeen (only drawing due to conceding a controversial 90th minute penalty) and won at Hamilton.

Partick Thistle have been in great form and, after also drawing at Aberdeen, they were only denied another excellent point by a 90th minute winner from ten-man Celtic on Saturday.

So is the decrease in home advantage significant in Scotland? The answer is actually no:

Table-2-Update1

There has only been a marginal increase in the number of away points won in 2015/16 compared to 2014/15 (0.33%). As we are only halfway through the season this is likely to change and could go either way compared to last year.

While there are some teams who appear to favour playing at home and some that favour playing away this is no different to any other league in the world.

Anecdotally I would have expected more teams to have stronger home records but this does not appear to be the case, with only Ross County & Motherwell picking up significantly more points at home:

Table 3 update

Table 4 update

Despite St. Johnstone’s fantastic recent away run they do not even rank in the top four teams in the league in terms of picking up a higher % of points away from home. This could suggest that they are just doing well generally. Indeed, in the midst of that five game away winning streak they only played three home games (1/1/1). It was their home record before they started to win away that was better.

Kilmarnock have a terrible home record, having lost 4-0 on three separate occasions as well as suffering heavy 5-2 defeat. It seems that when they lose they tend to lose big.

What is interesting is that of the two sides with artificial pitches in the SPFL (Hamilton and Kilmarnock) neither has won a significant number of their points at home, something that might be expected due to the unfamiliarity of their opposition to playing on such surfaces.

Given that home advantage seems to be fairly insignificant in Scotland we would then expect this to be reflected by the bookmakers’ Asian Handicap lines:

Table 5

The number of times the bookmakers have given the home team a supremacy (an AH line of -0.25 or greater) is roughly in conjunction with the % of points picked up by the home teams.

There is still a slight tendency to give supremacy to the home side, but not by much. However, there were a slightly higher number of games they could not call than those where the away team was considered favourite (22.4% DNB). In fact, of the teams given away supremacy, the current top three of Aberdeen, Celtic and Hearts accounted for over 80% of such occasions.

Looking for reasons why all of the above is happening and teams are becoming stronger away from home is tougher and many of the old reasons given are unlikely to hold up today.

Firstly, having to travel and stay in poor accommodation is no longer a big issue, with road travel improved and hotel conditions generally quite good. In Scotland only Inverness Caledonian Thistle and Ross County are a significant distance from the rest, with most other teams within ninety minutes.

Officials are a notorious source of consternation for fans and some believe a referee always has it in for their team. However, referee bias has been shown to be less evident, with only a slightly higher number of fouls and penalties given to home sides. When looking deeper into this, an easy explanation could be that home sides tend to have more possession and therefore a higher chance of being in positions to win free kicks and penalties.

Another point previously put forward was that home teams tended to put more effort in to impress the fans, as that was where they saw them more often. The amount of live football and highlight shows on television makes this extremely unlikely now and there is no longer a place to hide for most players.

That leaves me with one final theory: tactics.

In recent seasons many teams have stopped playing with two strikers. Indeed, in the SPFL only Hearts, Ross County and St. Johnstone set up this way, though County could be argued as playing 4-4-1-1.

Many teams now favour 4-2-3-1, a tactic known for stifling the midfield and counter-attacking quickly rather than playing an open game. Essentially ten of the teams play with five midfielders, if we include Ross County.

When you are the home side having to break down considerable numbers of players makes it difficult. As such, while the home side will still see a lot of possession it can tend to be in harmless areas with a mass of bodies between the ball and the goal.

The away team can then break and have fewer defenders to get beyond in order to create chances of their own. They may have less chances overall but could they expect the chances they do get to be of better quality? Using Stratagem’s definitions of Great Chances, Good Chances and Attempts I investigated further:

Table-6-update1

Aberdeen are the outliers here as the only side to create more Great Chances when away and they are also one of only four teams to create more Good Chances when away too.

More significantly though, eight teams are better at converting Great Chances and Attempts away from home, while Good Chances are split perfectly down the middle.

It seems to make sense that it would be easier to make more of chance situations when facing fewer defenders and playing more on the break as an away team.

TABLE 7 update

This final table shows that the average conversion of each chance level is better when away from home than when at home. In the case of Attempts it is actually more than double.

This appears to be playing the biggest part in why teams are faring better on the road than previously.

Home advantage may not be the force it was previously, but it seems that potentially unsustainable conversion rates are driving this, in the Scottish Premiership at least. Over time the expectation could be that these rates will regress to more usual levels, which may see an increase in the number of points picked up at home over the second half of the season.

I will keep track on how this progresses and update you throughout 2016.

Dave Willoughby