In today’s little chat about the Bundesliga ahead of matchday 12, the following points will be addressed: Firstly, the current table situation will be examined a little, the arrangement of which certainly surprises us all and which thus also provides plenty of excitement, but which could nevertheless gradually correct itself in the direction of “normal”. Then the positive and negative surprises are highlighted, but this is done intuitively at first.
Then the computer simulation is carried out and the current results are examined, especially in comparison with the results of a week ago. Who has improved their chances, who has worsened? By how much have they grown or fallen? This is particularly exciting if you look at it as a development, whereby there will certainly be a few curious shifts that still have to be analysed first.
The comparison of the chances determined by the computer via simulation in the form of relative frequencies with the betting market certainly also promises plenty of excitement. You can now either look at the differences and go in search of errors, to what extent the numbers determined on the betfair betting exchange by mass intelligence could be better than those calculated by the computer, or, alternatively, courageously place a bet, trusting the computer. But one can also enjoy how well intuitive, human insights and pure, therefore emotionless computer logic complement each other.
Finally, a few other interesting statistics should be offered, in which the yields achieved by the individual teams so far are compared with those expected by the computer. In relation to the intuitive observation above, this is, so to speak, a kind of objective yardstick: who has played the most above their expectation, who has underperformed it the most. Displaying this in a table could be quite interesting.
So let’s go:
1) The season so far.
A few words about the season as a whole: It is clear that all the teams are very optimistic. This refers to the fact that they all radiate that they feel capable of scoring a goal at any time and play courageously forward. There could very well be a connection between the performance of the German national team in South Africa, where they combined just as courageously and quickly forward and German football thus gained the recognition worldwide that several title wins could not bring it. This football, as they now presented it, was loved. This was expressed very well by many of the otherwise grudging nations – this adjective, however, has its origin in the fact that the successes were very often based on luck (at least as perceived abroad) (England, Italy as examples, where a very rapturous media reaction was experienced). It is obvious that this optimistic football with a lot of joy of combination was “copied”, so to speak, and directly transferred to everyday life in the league.
The pleasing consequences are, on the one hand, that there are more goals than before (about 3.2 goals per game at present, otherwise in recent years around 2.9), but also that one could experience surprising turns in many games, which is always a spectacle for every fan – except those of the team negatively affected by the turn. A game like Mainz’s 4:3 in Wolfsburg after 0:3 (surely the main reason for Mainz’s surge) will be talked about for years to come. Thanks to the fact that both teams – including Wolfsburg – were in a position to score a goal at any time and the leaders did not switch to managing the result.
Apart from that, the Bundesliga is simply fun. The matches always promise exciting football, varied games and great goals. As long as one disregards the proof that still has to be provided as to whether these strategies, which have been slightly modified compared to the past, will also prove themselves in the long term on the European stage, one can at any rate be happy about the variety of events for the time being. The upside-down table is nevertheless beginning to gradually free itself from this situation. But there are still enough positions where someone returning home after a long stay abroad and looking at it for the first time today might at least rub his eyes in amazement.
First of all, here is the pure table picture from Monday, 8 November 2010:
Sp | S | U | N | Punkte | Tore | Differenz | ||
1 | Borussia Dortmund | 11 | 9 | 1 | 1 | 28 | 27- 7 | +20 |
2 | FSV Mainz 05 | 11 | 8 | 0 | 3 | 24 | 19- 11 | +8 |
3 | Bayer Leverkusen | 11 | 6 | 3 | 2 | 21 | 22- 16 | +6 |
4 | Eintracht Frankfurt | 11 | 6 | 1 | 4 | 19 | 20- 11 | +9 |
5 | TSG Hoffenheim | 11 | 5 | 3 | 3 | 18 | 22- 15 | +7 |
6 | Hamburger SV | 11 | 5 | 3 | 3 | 18 | 17- 15 | +2 |
7 | 1.FC Nürnberg | 11 | 5 | 3 | 3 | 18 | 17- 15 | +2 |
8 | SC Freiburg | 11 | 6 | 0 | 5 | 18 | 17- 18 | -1 |
9 | FC Bayern München | 11 | 4 | 4 | 3 | 16 | 15- 13 | +2 |
10 | Hannover 96 | 11 | 5 | 1 | 5 | 16 | 13- 20 | -7 |
11 | Werder Bremen | 11 | 4 | 2 | 5 | 14 | 19- 27 | -8 |
12 | VfL Wolfsburg | 11 | 4 | 1 | 6 | 13 | 18- 19 | -1 |
13 | FC St. Pauli | 11 | 4 | 1 | 6 | 13 | 12- 18 | -6 |
14 | VfB Stuttgart | 11 | 3 | 1 | 7 | 10 | 22- 19 | +3 |
15 | 1.FC Kaiserslautern | 11 | 3 | 1 | 7 | 10 | 14- 21 | -7 |
16 | FC Schalke 04 | 11 | 2 | 3 | 6 | 9 | 13- 17 | -4 |
17 | 1.FC Köln | 11 | 2 | 2 | 7 | 8 | 13- 22 | -9 |
18 | Borussia Mönchengladbach | 11 | 1 | 4 | 6 | 7 | 17- 33 | -16 |
First of all, per visual impression, once again the positive surprises that the table shows: Borussia Dortmund with an outstanding haul of 28 points after 11 games, as coach Klopp also pointed out in an interview. Since, according to many years of observation, anyone who averages two points per game is a serious title contender, Dortmund’s 28/11 = 2.5454 is simply outstanding. After just under a third of the season, there is no way around their first mention.
Mainz also remain above this yield and are therefore worth a mention. In any case, the Frankfurt Eintracht still deserves a positive mention, as they have also achieved far more than just a small series with great football in the last few weeks. It’s simply fun to watch them play. Of course, Freiburg and Nuremberg should not be forgotten in the list of positive surprises, although Freiburg did not show such outstanding performances in all their games. In their case, one could already speak of a lucky course of the season so far. Of course, there were still enough good games. As small examples of luck, just the winning goal in injury time in Frankfurt, which was also declared “offside”, or the 2-1 victory against Kaiserslautern in a game in which Lauter actually had the greater number of excellent scoring opportunities and at least deserved the draw. Nuremberg, on the other hand, has a points tally commensurate with its performances.
There are also negative surprises and they are mentioned in this order: Schalke, Stuttgart, Werder Bremen, FC Bayern München and VfL Wolfsburg. Schalke and Stuttgart, in particular, have been winning recently, while Bayern in itself had a clear upward trend – also in terms of dominance on the pitch – but in the game against Gladbach suddenly gave a euphoric opponent, who had not been given the third goal in time, the lead and then almost had to speak of luck to at least be able to take the one point. An inspiring game, no question, especially as Bayern continues to divide the nation and there are always plenty of people who rub their hands when they are spat upon. By the way, Gladbach deserved a little reward after a long dry spell with not only bad games.
Werder Bremen really scored a kind of own goal with the conceded 0:6. Although the performances in the games before, against Twente in the Champions League and especially the very unfortunate cup exit in Munich were rather brilliant, so that in itself you could see enough elements of a class team, because they were the only ones this season who dominated Bayern – and even in Munich. While sobering for the moment, even Sky live commentator still acknowledged after the break in Stuttgart – for newcomers — that the interim result of 0:3 expressed a superiority that did not exist on the pitch. Just remember the penalty at 0:1, which Torsten Frings missed after 19 in a row (as was learned last time).
Of course, VfL Wolfsburg owes its mention in this category to the two home defeats against Mainz and Leverkusen, both of which came after a clear lead (3:0 and 2:0 respectively), and thus provided a spectacle, but only for all non-Wolfsburg players. One enjoys such surprises, but twice in such quick succession is really outstanding, perhaps even a little tragic from their point of view (in contrast to the “self-inflicted blame” that is often attributed in this country).
There is also some statistical evidence on this in figures, expressing the size of the surprises – positive or negative – which comes a little later. This part here was the rather intuitive – but first, before looking at the numbers — reasoning.
2) The results of the simulation
After the computer had given Dortmund the best chance last week, with 57.1%, compared to Bayern with the second best chance of second place with only 15.4%, the results on the weekend of 6 and 7 November naturally played into the computer’s hands in the sense that it had already recommended a bet on Dortmund last week and this must have gained in value.
At this point, it is worth mentioning that the increased value of the bet–analogous to an acquired and increased share–could of course be realised on the betting market. Those who followed the computer last week and placed a bet could realise their profit today by selling the bet.
For sure, the chances have improved considerably due to Dortmund’s convincing and, in the end, really outstanding (sure, after the sending-off Hannover collapsed despite defending the penalty) 4:0 victory compared to Bayern’s altogether unfortunate away 3:3 at Gladbach. This will be reflected both by the computer in its re-run simulation and by the betting market.
So here are the simulation results, which express the chances in probabilities. The first number indicates the result of the computer simulation in absolute occurrence of first places in 1000 runs:
Anzahl 1. Plätze | in Prozent | Faire Quoten | |
Borussia Dortmund | 726 | 72.60% | 1.38 |
Bayer Leverkusen | 95 | 9.50% | 10.53 |
FC Bayern München | 75 | 7.50% | 13.33 |
TSG Hoffenheim | 29 | 2.90% | 34.48 |
FSV Mainz 05 | 26 | 2.60% | 38.46 |
Hamburger SV | 18 | 1.80% | 55.56 |
VfL Wolfsburg | 10 | 1.00% | 100.00 |
Eintracht Frankfurt | 5 | 0.50% | 200.00 |
Werder Bremen | 5 | 0.50% | 200.00 |
VfB Stuttgart | 4 | 0.40% | 250.00 |
FC Schalke 04 | 3 | 0.30% | 333.33 |
1. FC Nürnberg | 2 | 0.20% | 500.00 |
SC Freiburg | 2 | 0.20% | 500.00 |
1000 | 100.00% |
Borussia Dortmund has increased its chances considerably to a whopping 72.6% (according to the computer). This is partly due to their own result, of course, which was a clear 4-0 away win, but of course also to the results of their rivals. Both Bayern, with only one point, and Mainz, who were the closest, with one defeat, lost ground. Leverkusen and Hoffenheim, of course, made up ground with their victories.
The fact that Schalke and Stuttgart are now in the list is of course thanks to their clear victories, but also indirectly shows the respect (of the computer) for their true performance (which is of course only stored in the form of parameters).
Here is the list of changes from the previous week, still from a pure computer point of view:
Gain/loss compared to the previous week
absolut | in per cent | |
Dortmund | 155 | 15.50% |
Leverkusen | 21 | 2.10% |
Bayern | -79 | -7.90% |
TSG Hoffenheim | -16 | -1.60% |
FSV Mainz 05 | -57 | -5.70% |
Hamburger SV | -2 | -0.20% |
VfL Wolfsburg | -11 | -1.10% |
Frankfurt | 4 | 0.40% |
Werder Bremen | -25 | -2.50% |
VfB Stuttgart | 4 | 0.40% |
FC Schalke 04 | 3 | 0.30% |
1. FC Nürnberg | 1 | 0.10% |
SC Freiburg | 2 | 0.20% |
0 | 0.00% |
Dortmund thus made the biggest jump in the positive sense, with a gain of 15.5%. They inherited these chances mainly from the main losers Bayern (-7.9%) and Mainz (-5.7%), which already adds up to 13.6%.
The fact that HSV, for example, loses chances despite a win may be due to the fact that their main rival achieved an even better result than they did, and this even in an away match, but their chances were already relatively good beforehand and therefore did not improve. However, it is also possible that in the last simulation they were lucky in the 1000 runs — despite the multiplicity intended to exclude luck — whereas this time they were somewhat unlucky. This is made possible by the fact that the odds are not that high and thus a simulation can also turn out favourably for them here or there by chance, and this repeatedly (one already senses that an absolute occurrence of 18 or 20 times is not an absolutely reliable measure; for sceptics gladly mentioned: Clearly, a simulation can also not provide definitive answers with a claim to absoluteness).
Obviously, the collapse of Werder Bremen’s chances, who not only weakened their goal difference with a 0:6, but also lowered their playing strength quite considerably with this result (the computer does indeed take the amount of results into account).
3) The betting market on the question "Who will be German Champion 2011?"
For comparison, here are the numbers on the betting market, as traded on the betting exchange on 8.11.2010:
Betfair figures | |||
Dortmund | 481 | 48.08% | 2.08 |
Leverkusen | 111 | 11.11% | 9 |
Bayern | 250 | 25.00% | 4 |
TSG Hoffenheim | 22 | 2.17% | 46 |
FSV Mainz 05 | 29 | 2.94% | 34 |
Hamburger SV | 69 | 6.90% | 14.5 |
VfL Wolfsburg | 12 | 1.18% | 85 |
Frankfurt | 8 | 0.77% | 130 |
Werder Bremen | 8 | 0.83% | 120 |
VfB Stuttgart | 5 | 0.50% | 200 |
FC Schalke 04 | 6 | 0.63% | 160 |
1. FC Nürnberg | 1 | 0.14% | 730 |
SC Freiburg | 1 | 0.10% | 1000 |
1003 | 100.34% |
Again, the backmost column with the odds offers was inserted first. The other figures were calculated from this to make them comparable with the simulation results.
Please note: On a betting exchange, the odds are traded just like on the real exchange. You can buy a team (the stock exchange uses English technical terms, of course. “Buy” is therefore the same as “back”), as well as sell it (English: “lay”). You “back” them by betting on their success, i.e. on them winning the title in these betting offers. When you make a “lay” bet, you always win it if the team you are betting against does not win the title. Both sides are represented. The prices also develop according to supply and demand. So if Dortmund were bet on more now (as the computer “continues” to “recommend”, so to speak), the price would gradually fall. The prices on the “back” side are shown here. Those of the “lay” side would only differ marginally. One buys and sells, “backed” or “laid”, for almost the same price. Whenever a trade, a deal or a bet is made, two people have sought each other out – and found each other at the same price.
Here you can see a huge difference to the computer. This could make one worry about pure computer logic and its validity. However, human intuition is possibly far removed from any computer logic at such points – and still wrong.
However, one always has to consider smaller factors that might nevertheless prove the market, the so-called “mass intelligence”, somewhat right. For example, one effect worth mentioning is this: Bayern are still seen and perceived as the strongest team. The return of the saviours Robben and Ribery will always have to be kept in mind, but this is not the intended effect at this point. It is the following effect, which has also been observed elsewhere: The favourite has a small additional chance to steer fortunes in his favour by taking advantage of following the steps of the competitor. This means that the Bavarians know exactly how many points they have to make up and what they have to do.
If you now think your way into computer logic, you first realise that this influence cannot really be mapped. So it is also impossible to express this form of “superiority” in playing strength, to make it measurable. The reason is this: In other seasons, when the Bavarians (as perennial favourites) control the field from the top, they may fail to score many a point that they could force if they were fully concentrated — provided, that is, that their rivals are either right behind them or in front by the same margin. But since they don’t, since it wasn’t necessary in that season, the computer has little possibility of factoring in this “imagined victory”, of recognising this superiority in an appropriate form. Intuitively, it is there, it could be there. For computer logic, the effect is not really accessible, one would have to programme it in just as intuitively.
The practical example in which this effect was first observed by the author, in the early days of the computer programme: Halmstad BK had won a European Cup first leg against the then very strong AC Parma 3:0. Parma won the return match 4:0. Most likely they would not have won the home match 4:0 if it had been played first. There is an extra little chance when you know exactly what you have to achieve. And that speaks in favour of the favourite.
So Bayern and Dortmund are the huge divergences computer – betting market. It is difficult to “prove” which of them is right and by how much. You could just bet on this or that side, relying on this or that assessment.
If one is looking for the best bets, the biggest “recommendations” of the computer, then one should only go there in search. On the one hand, one could decide to place a “lay” bet against Bayern. This would mean that you automatically have all the other teams and henceforth take a clear bias against Bayern.
But one could also make a “back” bet on Dortmund and henceforth identify with their fate. This is to a large extent a question of character, whether one prefers to be happy about bad results of a single team or to be happy about good results of another. Of course, there is always a relationship, because whoever bets on Dortmund is at the same time betting a little against Bayern. So if Dortmund lose, but Bayern also lose, all could be right with the world again.
The deviations of the other teams will not be discussed in detail today. The reader can see this from the figures shown. With smaller chances (small probabilities), random simulation results, as explained for HSV, can always have a greater influence. This effect increases the smaller the chance. However, the deviations are not necessarily noticeable.
4) Deviations from expectations
This observation is a little different from the one we are used to seeing in the media. Over smaller periods of time, at least, it can be quite significant against whom one has to play and with which place right. So if, for example, a team has to play Bayern at home, then Dortmund, then Leverkusen at home and finally Werder Bremen in a row, but “only” gets one point from these games, then the media often already come up with crisis talk, even though the yield achieved may (from the computer’s point of view, but perhaps also objectively) be in line with their expectations, provided it is perhaps a relegation candidate.
Over 11 matchdays, of course, the opponents already play a somewhat smaller role, since it most likely resulted in a reasonably colourful mix for everyone. Nevertheless, the number of home games, for example, but also the quality of the opponents you met at home in relation to those you met away, can already have an influence.
But even otherwise, this kind of numbers game is something different. If you compare it with other leagues, you could even make the “crazy play” visible by adding up the deviations in absolute terms. Enough talk, now here are the figures, first in the form of the most relevant point yields.
The table shows the expected points, the points actually scored and the deviations. The table is sorted according to deviations, from positive to negative: (Home and Away is only listed in terms of games played)
Home | Away | Total | Real | Deviation | |
Borussia Dortmund | 8.40 | 9.16 | 17.57 | 28 | 10.43 |
FSV Mainz 05 | 9.17 | 4.52 | 13.70 | 24 | 10.30 |
1. FC Nürnberg | 6.96 | 5.06 | 12.03 | 18 | 5.97 |
SC Freiburg | 8.69 | 4.18 | 12.87 | 18 | 5.13 |
Eintracht Frankfurt | 7.29 | 6.68 | 13.96 | 19 | 5.04 |
Hannover 96 | 8.09 | 3.51 | 11.60 | 16 | 4.40 |
Bayer Leverkusen | 11.87 | 6.53 | 18.39 | 21 | 2.61 |
TSG Hoffenheim | 8.13 | 7.71 | 15.84 | 18 | 2.16 |
FC St. Pauli | 5.99 | 6.11 | 12.10 | 13 | 0.90 |
Hamburger SV | 10.28 | 7.50 | 17.78 | 18 | 0.22 |
1. FC Kaiserslautern | 6.76 | 5.34 | 12.10 | 10 | -2.10 |
1. FC Köln | 6.55 | 5.05 | 11.60 | 8 | -3.60 |
VfL Wolfsburg | 9.38 | 7.54 | 16.93 | 13 | -3.93 |
Werder Bremen | 10.22 | 7.86 | 18.08 | 14 | -4.08 |
Borussia Mönchengladbach | 7.98 | 4.13 | 12.11 | 7 | -5.11 |
FC Bayern München | 13.30 | 8.43 | 21.74 | 16 | -5.74 |
VfB Stuttgart | 10.19 | 6.93 | 17.12 | 10 | -7.12 |
FC Schalke 04 | 10.86 | 6.89 | 17.74 | 9 | -8.74 |
Sum | 6.74 |
It is important to mention here: The computer calculates the chances for 1, X and 2 for each match. An expected value for a score from a match for a team is calculated as the probability that it wins this match * 3, since there are 3 points for a win, plus 1* the chance that it ends in a draw, since there is one point for it. The probabilities are not objective. Nevertheless, the calculated odds are usually in line with the market.
The sum of the deviations does not have to be 0. This is because the draw probability does not have to be met. There have been too few draws so far this season, so the value of the deviation is positive. So too many points have been translated. Reason: there have been too few draws (in relation to the expected number; you can even read off how many draws are “missing”: namely 6.74. Logical? Yes: each draw that occurs costs exactly one point, since a total of 2 points are distributed for draws, while 3 points are awarded for a decided game).
The amazement at 1st and 2nd place is certainly limited. They are the big surprises of the season, even if Mainz have now been downgraded a little to normal with two defeats in a row. So Dortmund has “taken the cake” with 10.43 points above expectation, but Mainz also remains houseful above all other “surprise unteams” at over 10 points above expectation.
Nuremberg in 3rd place is not necessarily a surprise, considering that they were ranked rather low to begin with. In relation to Frankfurt, for example, you can see that the computer “trusted” Nuremberg with only about 12 points so far, whereas Frankfurt already had 14. But one reason can always be: The programme so far and the number of home games, as people usually have higher expectations in those. You can see that Hannover, for example, was intuitively not mentioned in the above section. Surely it’s because they haven’t stood out with good results in the last few matchdays. Their great series was at the beginning – as a human being you have probably already forgotten and something like that could not happen to a computer.
The negative surprises are relatively clear. Schalke with a deficit of 8.7 points, Stuttgart still with 7.1 and Bayern with an impressive 5.7. Of course, intuitively you can see that. Bayern is missing a few things, especially since they are often the best at recognising what one would actually have expected. 16 points is not so bad. Unless your name is Bayern Munich…
The fact that Gladbach is even behind Werder (i.e. has a bigger deficit) probably also only amazes you as a person who intuitively assesses a 0:6 very negatively and admires a 3:3 against Bayern. Since this occurred recently, one might feel “Werder was even worse, though.” The fact that Werder are “only” four points behind makes the crisis talk even more unbelievable, since they would only have to “turn around” one result (for example, the one against Nuremberg a fortnight ago, when they were even leading) and everything would be just fine.
Here is the absolute total of the deviation, perhaps so that you can recognise the crazy game a little, but no more international comparisons will be made here:
The absolute deviation is 87.6 points, which corresponds to an average deviation per team of almost 5 points. This already seems quite a lot, but would certainly only become clear in comparison with another, better-sorted league. However, you always have to take into account the number of match days for something like this. Over the full distance of 34 matchdays, the deviation would be normal again, presumably.
Well then, in any case, we can be curious about the further development.
Will Bayern move up again? Will Schalke and Stuttgart be able to gain some breathing space at the bottom? Can Mainz regain their composure and inspire with the great, carefree attacking football they so often show? Will we see another outsider as a high-flyer? We shall see. In any case, there is plenty of reason to be curious.
Here’s to a new one next week!