An article was found on the internet dealing with the declining home advantage in the 1st Bundesliga. You can find this article at
http://www.n-tv.de/sport/dossier/Auswaertssiege-nehmen-zu-article646784.html.
There are a number of remarkable things about it. For example, that extensive studies were made about it. Well, these were not “studied”, so a real judgement is not possible. Nevertheless, the author of this text has observed football intensively and objectively (which simply means not as a “fan of a team”) from early childhood, which means that many phenomena have been known for a long time and their development has been followed. The fact that there are certain observations that allow the conclusion that home advantage is really on the decline should, on the one hand, be readily stated. On the other hand, however, we will also look at what is happening from the point of view of the betting market, which always has a say in the matter. First, however, a few facts on the subject:
1) Stocktaking
The specially created football programme with its built-in database provides some very illuminating information about this “development”.
Here is a diagram showing how the figures have developed in percent. All seasons of the first Bundesliga from the 1993/94 season to the 2009/2010 season are taken into account:
A brief interpretation of this diagram: The home victories have not really regressed dramatically in this picture. Even if it was still 50% in the first year and later developed in the direction of approx. 48%, one would rather describe the first value as an “outlier”. In the case of the undecideds, it is somewhat more “dramatic” in the sense that they have regressed, although it is certainly gratifying and desirable. From 30%, then even just above, down to 27%. The interpretation for this is quite simple: the three-point rule was introduced for the 1995/96 season. The intention behind it was to increase the attractiveness of the game by giving an extra bonus for increased risk when the score is even (provided you score the winning goal). One would only have to add here that the coaches and players first had to learn and use this effect. Because in the 95/96 season, there were initially more draws, only then did the figure drop.
Now that both figures are well interpreted, the final assessment of the development of away victories is no longer difficult. The home victories are generally correct, i.e. they are constant, the draws are becoming fewer, so the away victories can only have increased. The fact that draws and away victories are now more or less in balance is pure coincidence.
By the way, the computer does an even better job. The author has programmed it to make forecasts. The comparison of these figures is also gladly shown here in chart form:
Since the data range is broken down much more precisely here (from 43% to 51% as opposed to 0 to 60 before), you see a very different effect than in the coarser-meshed diagram above: the home wins jag much more than it appears above. If you initially only follow the dark blue line, you can see that after the first outlier result, it first goes down significantly after that, only to rise again relatively steeply, and then fall off fairly steadily towards the end. It suggests that you simply have to accept larger movements. It depends on so many coincidences how a single match turns out, so that there may well be relatively large deviations over the period of a season. Consider that there are only 306 games per season, so a deviation of a whole percent is achieved by three games that “coincidentally” end in a home win or a draw. And three games can always “accidentally” be overturned by a goal that is not given or is given after all, a penalty, a wrongly recognised offside. The jump from 50% to 46% is based on 12 games (per season). And even these could still be passed off as accidental.
Overall, the percentage of home wins over this period is 47.5%. It was still around 50% in the first season, but then dropped dramatically to 46%, then stagnated at 47% for a while, only to end up well above 48% again. If one wishes, one can safely regard the development as purely coincidental, but there is also the possibility of taking a more detailed look at it and considering a systematic development to be possible. Before doing so, however, let us briefly interpret the development of computer forecasts.
The computer has behaved much more moderately. It always takes current developments into account, so that it also makes adjustments, but never so drastically that it could represent reality. However, one would simply have to certify here that it does it rightly. Because in the end he has hit reality pretty accurately with the entire forecast. He mimics the current developments quite well, albeit more gently. Only: it is simply impossible for him to anticipate the extreme fluctuations. That would only be the case if one could forecast a development, for example, by changing the rules. Then, as a programme developer, one could intervene manually, of course.
All in all, the development of the last few years can be seen as a trend. After all, there have been 5 seasons where the value has moved downwards more or less continuously, from about 48.5% to about 47.5%. One would also accept this immediately and think about the causes accordingly. That is a matter of course for a statistician. Whether one should now react, outside of the normal adjustments made by the computer itself, is an open question.
2) Causal research
If one now wants to describe the number of home wins as “significantly falling” – from the author’s point of view, one should always bear in mind that there is the task of predicting as well as possible, since it is a matter of forecasts for betting and errors have a pecuniary effect –, then one must go into cause research. However, one should always consider the randomness of the development as possible.
The research results cited in the article above neglect an intuitive value, which one can only detect by reflection or intensive observation. The critical point is the referees. And in connection with this — be it possibly even a prerequisite — the increased public attention.
The referees used to have a much easier task, although we are talking about much earlier. Very few games were followed by cameras, at least not permanently, but even if they were permanently, they were not broadcast publicly. In addition, there were no referee observers and no petty scorekeepers. Nor should the men in black from the earlier years really be accused of partisanship, let alone of having made bets themselves. On the other hand, it is well known that in the past the organisers, i.e. the hosts, were happy to court them, pick them up from the airport or put them up in a (good) hotel. This can have a small influence on a critical decision. The main influence for that fact from earlier times is nevertheless another one: Unless one is held accountable for every single decision, one might let the people’s voice decide here or there. And the people’s voice in this case is mostly the home audience.
Here is a short first-person account of the author’s experience:
I myself can remember very well the first real encounter with this phenomenon. I was nine years old. Up to that point, the 1967/68 season, Hertha played in the regional league (I was too small for their first first league appearance from 63 to 65). My father took me to all kinds of games, not especially Hertha, but they usually took place in front of crowds of 1000 to 5000 at the most (except for Hertha’s in the sold-out Gesundbrunnen, where they could fit 15000) . Hertha had gained the right to play in the promotion games by winning the Regionalliga Berlin for 1968. And this time the expectations were really high. So I was in the stadium for the opening game of the promotion round against Rot-Weiß Essen. With me in the stadium that day were about 80,000 expectant and hopeful Hertha fans. It was an atmosphere that had never been experienced before, and it simply gave you goosebumps. In the first minute, right after kick-off, a kind of foul action on a Hertha player Foul or not foul not the question: the whole stadium cried out indignantly, you just had to join in yourself. A gigantic concert of whistles almost burst the eardrums. The referee had no choice but to call a foul.
This experience taught me quite early on that this form of home advantage exists. Not only did we always talk about it in our parents’ house and were bothered by the injustices, which quite often went in favour of the home teams, no, I also illustrated the home advantage in the Tipp-Kick national leagues, which were always held, where I kept the tables, drew up the match schedules and printed them out on a typewriter, and my father only turned up to play. And he always got Hertha as his number 1, so he had priority (so much only on the subject of fanaticism versus objectivity).
Today it has changed. The referees are not professionals, but they are in the limelight and get paid a lot of money for their appearances. In addition, they are selected at short notice, neutrally accommodated. On top of that, at least one observer sits in the stands at every match and, last but not least, their performance is monitored by cameras placed everywhere. A biased decision, which even the public wanted to impose on him — like the one at Hertha – Essen 1968 — he cannot make at all. The referees are even trained to withstand this public, spectator pressure and to remain as “objective” as possible.
It should be readily admitted here that the points mentioned in the article play a small role. There are the often uniform court sizes, which are familiar to every team, not especially to the home team. That tactics in this sense also give the away team the small – in their own view undesirable – advantage that it is easier nowadays to counterattack than to make the play is certainly true. The home team, at least in this country, with nominally more or less equally strong teams, is basically forced by the crowd to attack. And that’s not really the (only, right, only right?) way to get to the goal anymore. The comment in the article that the falling goal average also reduces the number of home victories is intuitively correct, but it is not so directly reflected in the data presented here, since, as can be seen above, the draws, which logically should increase with fewer goals, have actually fallen.
What is missing from the list of possible causes, by the way, is the point of travel, which was certainly relevant in the past. In the past, for financial reasons, teams often travelled by bus or train, and on the day of the match itself, which definitely made the legs heavier. This used to result in a greater home advantage. Incidentally, this consideration still plays a role in the lower English leagues today. If the journey is more than 200 kilometres, it can still have a detrimental effect on the away team’s chances with the bus journeys that are common there.
A small phenomenon, which you can only recognise on closer inspection, which can then definitely annoy you, or at least make you wonder: The display of the injury time. It certainly has positive side effects, in that the arbitrariness of the referee is somewhat limited. However, it has many more negative aspects. One is that the injury time has been taking more and more of the drama out of the game lately. The reason for this is that a team that has achieved its goal only wants to run down the clock – for example, through the unspeakable substitution procedure, which is by no means added on at the end. In addition, however, there are far fewer opportunities for the home team, which often presses for the winning goal, to score. But that is not enough.
This season, something else has been added: this season in particular, the length of injury time shown is much shorter, often 0 minutes, maximum 2 (just look at England, Spain or Italy, where there is virtually never less than 4 minutes). Although this is not meant to be a definitive digression – the reason being that the referees are afraid of making mistakes anyway, and especially in injury time there are often hectic or critical decisions, which they avoid by shortening the time — the effect is less drama, less suspense, fewer goals — and thus logically fewer home wins.
3) The betting market
If Ranga Jogeshvar is to be believed, the mass intelligence he tested on the audience is a reliable indicator of what is right. This can also be observed in Günter Jauch’s million-dollar quiz, where the audience almost always achieves a majority on the correct answer to the question intended for them. Accordingly, a betting exchange would have to provide a reliable indication of how likely a home win actually is in the Bundesliga and to what extent the current trend of declining home advantage is reflected.
And there one can only observe that the effect is quite small. The masses therefore tend to believe in a “random swing” and not a “systematic development”, which is how one would have to interpret it. By the way, here are the absolute figures that the computer offers for the current Bundesliga season (2009/2010):
Matches Home wins Draws Away wins Home goals Away goals
153 58 47 48 222 197
These are the numbers that reality has produced. Here now are the computer’s expectations:
Matches Home wins Draws Away wins Home goals Away goals
153 71.41 35.84 45.73 249.9 186.9
So the computer would have expected 13.4 more home wins over the period than occurred. This may sound like a lot, based on the figures, but it sounds a little less dramatic when you realise that almost all of these missing home victories have gone to the draws (and not to the away victories). This means that only the one goal, the winning goal, was missing. As indicated above, possibly related to injury time.
On the other hand, these 13 missing home wins are easily made up for, by a statistical swing in the other direction.
It was noted by the author a long time ago that an intuitively observed phenomenon can often be confirmed by the numbers (in the short term). However, it is often the case that this corrects itself again over time.
There is no final judgement according to the opinion represented here. The computer reacts very “reasonably”, so that one can be satisfied with the adjustment and there are no worries for the future in this sense. On the other hand, of course, current developments must be constantly monitored and attention must be paid to whether there is a change in the fundamentals.
In any case, the conclusion is that the article neglects very important phenomena that could be much more responsible for the possibly rather insignificant, random development.