| Betting tips today strongly relay on computer-based | | | | advantage factor was added to the mean of home |
| methods for match predictions. Such methods use | | | | team distribution. Assuming x and y to be the number |
| different mathematical models to estimate possible | | | | of goals scored by the home and away teams |
| match outcomes. The most popular mathematical | | | | respectively, then according to this model x and y |
| models used for soccer predictions are the Poisson | | | | come from independent Poisson distributions with |
| methods. | | | | means HM and AM respectively. In this case, the |
| This article summarizes the Poisson methods for | | | | expected number of goals is proportional to |
| soccer predictions that were developed in the early | | | | HA-AD+HG and AA-HD for the home and away |
| 90s, their advantages and disadvantages. | | | | teams respectively. HA and AA are home and away |
| The pioneering work of Moroney appeared in 1956 | | | | attack strengths, HD and AD are home and away |
| and defined the Poisson model to be the best suitable | | | | defense strengths and HG is the home ground |
| model for soccer score predictions. This model had a | | | | advantage. If the parameters above are calculated |
| number of drawbacks. Firstly, it didn't take into | | | | correctly, the match outcome can be successfully |
| consideration the effect of the opponent's strengths on | | | | predicted. |
| the final match score. Moreover, the model didn't | | | | The main improvement of this model compared to the |
| distinguish between attack and defense skills of the | | | | older one is in distinguishing between the attack and |
| teams; didn't consider the changes in team strengths | | | | defense skills of the opponents and, of course, in |
| over time; and didn't refer to the influence of home | | | | taking into account the influence of home field |
| ground advantage on the final score. | | | | advantage. |
| Only in the early 90s the new developments appeared, | | | | The model has two main drawbacks: it doesn't |
| improving significantly the Moroney's method. The most | | | | consider the time-dependent changes in team skills and |
| known method that used the Poisson model in this | | | | doesn't refer to the statistical dependency between |
| period was the method of Maher published in 1982. | | | | the Poisson distributions of the playing teams. |
| According to this method, the goals scored by the | | | | Later on, Maher's Poisson method underwent further |
| opponents during the game are drawn from the | | | | developments that were meant to tackle those |
| Poisson distribution. The mean of this distribution is | | | | problems. These developments will be presented in our |
| defined by the difference between the attack and | | | | next articles discussing the evolution of statistical |
| defense skills of the sides. Moreover, the home field | | | | soccer predictions. |