How significant are early season streaks either at home or away?

At the start of each soccer season certain teams can perform better either at home or away, resulting in them tagged as 'home/away specialists'. This article examines historical data to see if bettors should incorporate this into their betting?

Teams perform better at home than away

Teams, regardless of their overall ability, will invariably record better results at home than on their travels. Poor sides may struggle to win matches, even with the help of familiar surroundings, but in the long term, they tend to perform even worse away from their own stadium.

The advantage enjoyed by the hosts can be demonstrated by looking at the percentage of league points a side wins at home compared to on the road.

In the ten completed seasons since 2003-04, home teams won 5,690 points at home, but just 3,713 on the road. Therefore 61% of a side’s total league points, be they title contenders or relegation candidates, are gained at home.

61% of a side’s total league points, be they title contenders or relegation candidates, are gained at home

Every season a handful of teams produce home and away splits that diverge from the league average. Teams either perform exceptionally well at home compared to their away performances, inevitably gaining the label of “home specialists” or they excel on the road, and the narrative dubs them as away “specialists”.

These unusual splits can appear early in the season. Bettors need to know if these effects are real and persistent, and therefore likely to alter future match odds for such teams, or whether they arise because of random fluctuations within small samples?

This season, Stoke, who have historically performed much better at home than on their travels in the EPL, have failed to win when expected – they have lost three home games against relegation-threatened Burnley, Leicester and Aston Villa.

Only seven points have been won at home compared to eight away from the Potteries and the disparity in performance widens further when we see that Stoke have won at defending champions, Manchester City, as well as defeating Spurs at White Hart Lane.

Proportion of points won by Stoke City (2008-2014)

Season

Home Points

Away Points

% of Total Points Won at Home

2008-09

28

14

67

2009-10

29

16

64

2010-11

34

12

74

2011-12

27

20

57

2012-13

35

10

78

2013-14

36

14

72

2014-15 (After 12 games)

7

8

47

Once the strength of their opponents is accounted for by using the match odds for Stoke City’s initial 12 games, Stoke were expected to win an average of 10 points from their six home games in 2014/15 and between 5 and 6 in the six matches on their travels. Typically, Stoke would have won around five more points at home than they have won on the road.

So is their current record far enough removed from their expected results to conclude that they may have become an “away” specialist, either through tactics or temperament and should bettors consider adjusting their upcoming home and away chances accordingly?

In the plot below I have used the odds available for Stoke winning or drawing each of their 12 matches played, to simulate how often they might have been expected to gain one more point away from home compared to at home, in 2014/15.

 

Although a home/away point differential of at least minus one is a less likely outcome, it still occurred in over 10% of the 10,000 trials. Therefore a Stoke team with typical Premier League home and away expectations, might expect to produce more points on the road than at home around once in every ten occasions over a 12 game period.

Appealing as it may be to construct a narrative explaining their poor home performance so far, bettors should note that random effects, unlikely to be repeated consistently in the future, are a more credible cause.

Are there really home/away specialists?

We can further confirm that the specialist tag is often erroneously attached to a side by gathering the most extreme examples of sides that have either atypically struggled or excelled at a particular venue early in a season and then see if this apparent preference persists into the remainder of the season.

Since 2003/04, 50 teams had won 75% or more of their points at home over the first 12 games of the year compared to the average 61% figure for the Premier League in the recent past.

For example, Liverpool began the 2006/07 by taking 17 points from their opening 12 league games, 16 of which came at home.

Teams that have over performed at home or away for a period of games rarely sustain the trait in the long term

However, this apparent preference, over and above usual league averages when playing at home didn’t persist, either for the Reds or for the group as a whole in the remaining 26 games.

Liverpool, having gained 95% of their total points at Anfield in the opening 12 games, finished with an unremarkable 59% of their total points coming from home matches.

And, as a group, the “home specialists” who had gained an average of 81% of their points at home in their first dozen matches, won only a combined 63% of their total points at home over the rest of their respective campaigns.

How early season Home & Away “Specialists” performed in the rest of the year (2003-2013)

Cumulative top 50 Home or Away "Specialists" after 12 Games

Cumulative % of Points Won in final 26 Games at Home

81% of Points Won at Home

63%

41% of Points Won Away

63%

Similarly, teams who found it difficult to win points in early season games at home, but won more points on the road, as Stoke have done in 2014/15, largely produced less extreme home and away splits, both individually and as a group from game 13 to game 38.

These apparently polar opposite groups of teams from 2003-2013, potentially dubbed home and away “specialists”, actually performed in identical fashion from December onwards, each winning their proportion of home points at entirely normal levels of 63% of their total points won. Bettors therefore should note that teams, which have over performed at home or away for a period of games rarely, sustain the trait in the long term.

It's often tempting for bettors to try and identify sides that appear to exhibit traits that seem to be profound and permanent, and incorporate these into odds estimations, with a view to gaining an edge. But these traits are almost always a consequence of a small sample size, and can often be classed as statistical noise.