The Art of Statistics (not to be confused with stats) at our service
Yesterday we re-published an article on the academic approach to modeling soccer games. Some of the global leaders in Sports Analytics have developed a statistical model that has been well-established since 2003.
If you are seriously interested in Football Analytics, we cannot stress how important it is to look at the aforementioned article. As professors of Statistics and founding members of the AUEB Sports Analytics Group they have great contribution to the field.
On Leonardo Egidi’s blog there is a section with EURO 2020 model predictions. The model he uses is an application of Karlis and Ntzoufras 2003 paper “Analysis of sports data by using bivariate Poisson models“. The model has been quite successful so far! In the Round of 16 it correctly saw that the Czechs had good chances against Netherlands, that Croatia and Spain would go to overtime, and that England would finally beat Germany.
Probabilities & Odds
Our attempt here is to find where bookmakers’ overvalue or undervalue an outcome. To do that, we compare the probabilities and odds of the bookmakers to those of the bivariate Poisson model.
Our article on the first day of the EURO 2020 Quarter-Finals correctly suggested betting against a Spain win, with good value of a draw.
Note that all odds below exclude the usual bookmakers’ vig, i.e. the profit they make. All probabilities add to 100%.
To convert from a probability to decimal odds, divide the probability by one. For example, in the probability table below, in the “Model Switzerland-Spain” row we see that Spain’s probability is 40%. The odds therefore are 1/0.4=2.50.
Vice versa, to convert from decimal odds to implied probabilities, we divide 1 by the odds. For example, odds of 2.50 give probabilities of 1/2.50=0.4 or 40%.
The bookmaker odds we have in the tables below are derived from around 20 UK bookmakers.
Czech Republic vs Denmark
Interestingly enough there is not much difference between the odds the books give and the model predicts. This is contrary to both games played on day 1 of the EURO 2020 Quarter-Finals.
Czech Republic Win
- The books give the Czech Republic 22% chances to win. Average odds: 4.55.
- The model implies that the chances are 26%. In odds, this is 3.85.
- Our experience says that the Czech Republic has exceeded expectations without having any big names in their squad. The last time they reached a semi-final was 2004 and Pavel Nedved led a wonderful side.
- The Czechs play with confidence, they are capable of creating quality chances for Patrick Schick who has been exceptional.
- The Czech Republic will have to counter Denmark’s diverse tactical portfolio if they are to be successful. What is needed is a high level of preparedness by their coach and devotion to the cause by his players,
- We like what we have seen so far but we were expecting the bookmakers to name them the outsiders. Given the odds, we won’t fault anyone that bets on them. However, their victory still doesn’t represent a high value.
- Verdict: A bet on The Czech Republic holds small value- if you believe in them don’t be discouraged.
Czech Republic-Denmark Draw
- The books give 29% chances of a draw. Average odds: 3.45.
- The model implies that the chances are 28%. In odds, this is 3.57.
- Our experience says the odds fairly represent the value of the draw. Both teams play for the victory and we have seen them creating chances and focus on scoring goals. We don’t expect them to sit behind the ball and wait for the game to end but we also don’t expect them to foolishly attack, paying less attention to their defensive responsibilities. Therefore, we have a fair chance to see them draw and a decent odd to bet on.
- Verdict: The draw has good value and is quite likely to happen.
- The books give Denmark 49% chances to win. Average odds: 2.04.
- The model implies that the chances are 46%. In odds, this is 2.17.
- Our experience says Denmark has been amazing even without their best player. We have analyzed how well they played in their last game, and we couldn’t agree more with the model’s prediction.
- If you examine the way Denmark played in their last game, you will find that they were prepared for every scenario. The players fullfilled their coach’s expectation and performed admirably.
- Denmark has good players in every line as we have already analyzed. Players like Kjaer and Maehle rightly get all the attention they deserve. Not only they do a good job defending but they also contribute in possession and shot creation. The midfield looks fresh and will welcome another challenge again the Czechs.
- If you were expecting better odds for Denmark’s win then you clearly haven’t watched their last game. The value it represents is fair given that they seem to be getting better by every game.
- Verdict: A bet on Denmark is quite fair.
Ukraine vs England
- The books give Ukraine 10% chances to win. Average odds: 10.00.
- The model implies that the chances are 19%. In odds, this is 5.26.
- Our experience says Ukraine has been a big surprise. During the Group Stage, they had some good and some bad moments and their performance is quite unstable even when comparing first to second halfs of their games.
- Ukraine fought hard against Sweden to get the win but it was the red card that gave them the edge they needed to push forward.
- Yarmolenko has been absolutely sensational for them, making a difference with his quality touches and creative flair that leads to wonderful assists.
- Ukraine is a good team without a superstar, like their coach Andryi Shevchenko used to be. Even without a Shevchenko, they know what they want from a game, they clearly have a plan on how to move the ball in a direct manner to reach their opponent’s box and finish. They are their coach’s team after all.
- Verdict: The bookmakers clearly overvalue Ukraine’s win and we are obliged to note that there is a great value here.
- The books give 22% chances of a draw. Average odds: 4.55.
- The model implies that the chances are 25%. In odds, this is 4.00.
- Our experience says A draw is always on the cards when the stakes are high. Both teams have shown that they will play hard until the final whistle. Ukraine scored against Sweden in the last minute of overtime and England sealed their victory against Germany with a goal at the end. Although there is a 1/4 chance of a draw according to our model, we would have liked the draw in a better odd. Don’t get us wrong it is perfectly fine as it is, more than fair.
- Verdict: The odds are fair and both teams will work hard to not lose. If it is a close game, the draw will be a likely outcome.
- The books give England 68% chances to win. Average odds: 1.47.
- The model implies that the chances are 56%. In odds, this is 1.79.
- Our experience says England is a very good team that doesn’t feel comfortable being the favorite. They look much better when they aren’t in control of the ball and the rhythm of the game. Exactly what we have seen when they played against Germany.
- England’s counter-attacking playing style clearly suits them, as they have struggled against opponents that defend and give the majority of possession to them. Their worst games in the tournament came when they were labeled clear favorites and that is very indicative.
- Nevertheless, they are a very strong side highly capable of winning this year’s EURO. If they start being more confident in their ability to control a game, they might even have a better chance against Ukraine that what the model implies.
- As expected by the bookmakers, the odds are less representative of the tactical aspects of the game and more representative of the betters’ belief in the favorite. We do agree that there is a favorite in this game, England, but what odds would fairly represent their chances of a win?
- Verdict: A bet on England is what millions of betters will be pursuing. Although their win is the most likely outcome, we don’t think that the odds hold any good value.