This is my new little corner on the web. It is has been about three years since I started running, mostly for fun, some statistical models on football ("soccer") scores and championships -- nothing really fancy, but friends seem to like it and one of them asked me to start posting the simulations after each round. So, here you have it.
For those interested, what I am going to post are the results of a statistical model (technically, a Bayesian Monte Carlo method) that, given past results, the history of performances, and the scores in an on-going competition, returns the probability of each team achieving a given position. For the English Premier League, for example, given the bookmakers' odds for all teams, the scores of matches already played and the next fixtures, it informs the probability of each team finishing first, qualifying to the Champions League or Europa League, or being relegated.
Feel free to drop me a message if you find this interesting, want me to start simulating for other competitions, want to collaborate, or whatever! :)
Saturday 30 March 2019
Monday 15 August 2016
Brazilian Série A (Brasileirão) 2016
The English Premier League has just begun, but we can test our model in a championship with many rounds already played. Here is the data for the current Brazilian Série A.
Palmeiras is by far the favorite for winning the Brasileirão 2016, and only Santos and Grêmio have real chances of claiming the title (but we can still expect some surprises by Atlético-MG, Corinthians, and Flamengo). América-MG is almost relegated and there are few hopes for Santa Cruz, but all teams from Chapecoense to the bottom of the table should pay attention -- including Internacional, which for some rounds was a likely candidate for the title (but now has less than 1 chance in 250 of winning).
The columns are:
Palmeiras is by far the favorite for winning the Brasileirão 2016, and only Santos and Grêmio have real chances of claiming the title (but we can still expect some surprises by Atlético-MG, Corinthians, and Flamengo). América-MG is almost relegated and there are few hopes for Santa Cruz, but all teams from Chapecoense to the bottom of the table should pay attention -- including Internacional, which for some rounds was a likely candidate for the title (but now has less than 1 chance in 250 of winning).
The columns are:
- w, d, and l are, as expected, the number of matches for each team ending in a win, a draw or a loss; pld is the total number of matches played and pts the current number of points
- gf and ga are, respectively, the number of goals scored ('Goal For') and goals suffered ('Goal Against), with gd the goal difference (gf-ga)
- Champion is the current probably for the team to win the English Premier League, given the current situation and historical performance both of the team itself and of teams in its current position
- Libertadores Group is the current probability for the team to finish among the three best teams, qualifying for the group stage of the Libertadores
- Libertadores First is the current probability for the team to finish in the fourth position, qualifying for the first stage of the Libertadores
- Rel is the current probability for the team to finish among the three worse teams, being relegated
| Team | w | d | l | pld | pts | gf | ga | gd | Champion | Libertadores Group | Libertadores First | Relegation |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Palmeiras | 12 | 3 | 5 | 20 | 39 | 36 | 20 | 16 | 43.6 % | 79.7 % | 9.8 % | 0.0 % |
| Santos | 11 | 3 | 6 | 20 | 36 | 35 | 17 | 18 | 14.2 % | 50.6 % | 15.0 % | 0.0 % |
| Grêmio | 10 | 5 | 4 | 19 | 35 | 29 | 19 | 10 | 16.1 % | 53.5 % | 14.2 % | 0.0 % |
| Atlético-MG | 10 | 5 | 5 | 20 | 35 | 33 | 28 | 5 | 10.2 % | 42.5 % | 16.0 % | 0.0 % |
| Corinthians | 10 | 4 | 6 | 20 | 34 | 28 | 18 | 10 | 8.4 % | 33.3 % | 17.1 % | 0.0 % |
| Flamengo | 10 | 4 | 6 | 20 | 34 | 23 | 20 | 3 | 6.7 % | 29.2 % | 13.9 % | 0.0 % |
| Atlético-PR | 9 | 3 | 8 | 20 | 30 | 20 | 19 | 1 | 0.4 % | 5.5 % | 6.4 % | 0.1 % |
| Ponte Preta | 9 | 3 | 8 | 20 | 30 | 25 | 31 | -6 | 0.1 % | 2.3 % | 2.9 % | 0.4 % |
| Fluminense | 7 | 7 | 5 | 19 | 28 | 21 | 18 | 3 | 0.2 % | 2.8 % | 3.4 % | 0.6 % |
| Sport Recife | 7 | 5 | 8 | 20 | 26 | 30 | 29 | 1 | 0.0 % | 0.2 % | 0.4 % | 6.4 % |
| São Paulo | 7 | 5 | 8 | 20 | 26 | 21 | 21 | 0 | 0.0 % | 0.3 % | 0.6 % | 4.1 % |
| Chapecoense | 5 | 9 | 5 | 19 | 24 | 27 | 33 | -6 | 0.0 % | 0.1 % | 0.0 % | 20.9 % |
| Botafogo | 6 | 5 | 8 | 19 | 23 | 24 | 29 | -5 | 0.0 % | 0.0 % | 0.3 % | 15.3 % |
| Vitória | 5 | 8 | 7 | 20 | 23 | 26 | 30 | -4 | 0.0 % | 0.0 % | 0.0 % | 33.6 % |
| Internacional | 6 | 4 | 9 | 19 | 22 | 21 | 23 | -2 | 0.0 % | 0.0 % | 0.2 % | 17.0 % |
| Coritiba | 5 | 7 | 8 | 20 | 22 | 23 | 28 | -5 | 0.0 % | 0.0 % | 0.0 % | 44.2 % |
| Figueirense | 4 | 9 | 6 | 19 | 21 | 16 | 22 | -6 | 0.0 % | 0.0 % | 0.0 % | 42.1 % |
| Cruzeiro | 5 | 5 | 10 | 20 | 20 | 26 | 34 | -8 | 0.0 % | 0.0 % | 0.0 % | 45.7 % |
| Santa Cruz | 5 | 4 | 11 | 20 | 19 | 22 | 28 | -6 | 0.0 % | 0.0 % | 0.0 % | 72.9 % |
| América-MG | 3 | 4 | 13 | 20 | 13 | 12 | 31 | -19 | 0.0 % | 0.0 % | 0.0 % | 96.6 % |
English Premier League 16/17
These are the probabilities for the English Premier League 16/17 after the first round (missing Chelsea vs. West Ham United).
There is not much to analyze yet, as there is only one round to modify the probabilities drawn from past results and bookmakers. Still, the current simulation is useful for the future analysis of the moving probabilities for the teams final positions, and, in particular, for checking the likelihood of new underdogs -- or, how likely is a new Leicester City? The model is showing that the probability for Leicester City repeating the last year position has not decreased significantly, but the Hull City has already decreased the probability of relegation -- way to go, Tigers! The more likely candidates for relegation are now Burnley, Sunderland and Bournemouth -- but these probabilities, as described, are still under the hevay influence of bookmakers' odds and past performances, whose importance for the model will decrease when it has more actual results to work with.
The columns are:
There is not much to analyze yet, as there is only one round to modify the probabilities drawn from past results and bookmakers. Still, the current simulation is useful for the future analysis of the moving probabilities for the teams final positions, and, in particular, for checking the likelihood of new underdogs -- or, how likely is a new Leicester City? The model is showing that the probability for Leicester City repeating the last year position has not decreased significantly, but the Hull City has already decreased the probability of relegation -- way to go, Tigers! The more likely candidates for relegation are now Burnley, Sunderland and Bournemouth -- but these probabilities, as described, are still under the hevay influence of bookmakers' odds and past performances, whose importance for the model will decrease when it has more actual results to work with.
The columns are:
- w, d, and l are, as expected, the number of matches for each team ending in a win, a draw or a loss; pld is the total number of matches played and pts the current number of points
- gf and ga are, respectively, the number of goals scored ('Goal For') and goals suffered ('Goal Against), with gd the goal difference (gf-ga)
- Champion is the current probably for the team to win the English Premier League, given the current situation and historical performance both of the team itself and of teams in its current position
- CL Groups is the current probability for the team to finish among the three best teams, qualifying for the group stage of the Champions League
- CL Playoffs is the current probability for the team to finish in the fourth position, qualifying for the playoff stage of the Champions League
- EL is the current probability for the team to finish in the fifth position, qualifying for the Europa League
- Rel is the current probability for the team to finish among the three worse teams, being relegated
| Team | w | d | l | pld | pts | gf | ga | gd | Champion | CL Groups | CL Playoffs | EL | Rel |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Manchester United | 1 | 0 | 0 | 1 | 3 | 3 | 1 | 2 | 21.9 % | 53.0 % | 11.0 % | 9.6 % | 0.4 % |
| Liverpool FC | 1 | 0 | 0 | 1 | 3 | 4 | 3 | 1 | 18.2 % | 45.5 % | 9.4 % | 8.5 % | 0.6 % |
| Manchester City | 1 | 0 | 0 | 1 | 3 | 2 | 1 | 1 | 21.8 % | 54.5 % | 10.9 % | 8.5 % | 0.2 % |
| Hull City | 1 | 0 | 0 | 1 | 3 | 2 | 1 | 1 | 0.3 % | 2.0 % | 1.9 % | 2.1 % | 21.3 % |
| Swansea City | 1 | 0 | 0 | 1 | 3 | 1 | 0 | 1 | 0.5 % | 2.8 % | 3.0 % | 2.8 % | 18.4 % |
| West Bromwich Albion | 1 | 0 | 0 | 1 | 3 | 1 | 0 | 1 | 0.4 % | 2.3 % | 1.4 % | 2.2 % | 21.5 % |
| Middlesbrough FC | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 0 | 0.2 % | 1.4 % | 1.2 % | 1.9 % | 28.9 % |
| Tottenham Hotspur | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 0 | 7.0 % | 25.8 % | 9.0 % | 9.2 % | 1.6 % |
| Stoke City | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 0 | 0.3 % | 2.5 % | 2.4 % | 2.4 % | 17.9 % |
| Watford FC | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 0 | 0.3 % | 1.9 % | 1.6 % | 1.7 % | 26.6 % |
| Southampton FC | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 0 | 1.7 % | 8.5 % | 4.2 % | 5.4 % | 6.7 % |
| Everton FC | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 0 | 2.5 % | 11.3 % | 5.6 % | 7.2 % | 5.1 % |
| West Ham United | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 3.1 % | 12.9 % | 6.7 % | 7.0 % | 4.5 % |
| Chelsea FC | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 12.5 % | 37.1 % | 10.8 % | 9.6 % | 1.2 % |
| Arsenal FC | 0 | 0 | 1 | 1 | 0 | 3 | 4 | -1 | 7.3 % | 26.8 % | 12.1 % | 8.8 % | 1.4 % |
| Sunderland AFC | 0 | 0 | 1 | 1 | 0 | 1 | 2 | -1 | 0.0 % | 0.6 % | 1.2 % | 1.5 % | 35.3 % |
| Leicester City | 0 | 0 | 1 | 1 | 0 | 1 | 2 | -1 | 1.7 % | 8.2 % | 4.8 % | 7.2 % | 6.4 % |
| Burnley FC | 0 | 0 | 1 | 1 | 0 | 0 | 1 | -1 | 0.1 % | 0.8 % | 0.6 % | 1.3 % | 38.2 % |
| Crystal Palace | 0 | 0 | 1 | 1 | 0 | 0 | 1 | -1 | 0.1 % | 1.0 % | 1.1 % | 1.4 % | 31.2 % |
| AFC Bournemouth | 0 | 0 | 1 | 1 | 0 | 1 | 3 | -2 | 0.2 % | 1.0 % | 1.0 % | 1.4 % | 32.7 % |
Subscribe to:
Posts (Atom)