Tottenham Spur has been deemed to be outclassed than teams other than the top five teams in the league. Sometimes, they look so competitive than people may even believe the team can challenge the consistent domination of the top five teams over the league. However, defeats after defeats with big goal margin against the top tier teams, knock people away from the illusion. The team is sitting at the 7th position behind the 6th place Southampton by only goal margins. Since the team is behind 9 points from the European Championship zone, the team just need to secure the position for them to play in the Europa League. Tottenham Spur has a reputation that the team can always win against the weaker teams. In the last eight games, the team will play against Manchester City and Southampton, and except these two games, Tottenham Spur will not face any team arguably better than they are. However, is it true that the team can overcome these challenge and guarantee their position?

As usual, we decide to use Simple K-means clustering to start the analysis. Based on the Cattell scree test, we use five clusters in this case. Here is the clustering assignment,



and here is the partial result of the clustering,


We can see clearly that the team cannot perform in their away games, especially the games in cluster 0 and 1, 8 games in total. Also, most of the games in cluster 2 are also away game, and they just win by small margin.

First of all, we use ANOVA to prove that the clustering assignment strongly depends on whether they play at home field or not. We got the 99.99999% p-value that the clustering assignment strongly relates to the Home variable. We know that the team still have five away games awaiting for them in this season, and we also know that the team gets worse results in away games than in home games. The trust in this team is not so convincing anymore, since they have 5 away games even though 3 out of 5 are against teams sitting behind them.

Based on our statistic model, the team has a rather stable and consistent defensive pattern. Let us take a look at the defensive power changes along the cluster assignment,



we can see that when the team played in away field, they tend to be scored much more than usual, since cluster 0 and cluster 1 with rather large p-value. According to our statistical analysis, the opponent team tends to play much more aggressively at home than they did away; the numbers of shots, chance created, crossing, and Aerial duel winning percentage increase significantly.

We run the ANOVA between goals and the cluster assignment; as our expectation, the team perform consistently, rarely affected by playing at home or away,


Let us go back to the cluster assignment, we can see that in the last four games, 2 home games and 2 away games. There are 1 game in cluster 0, 2 games in cluster 3, and 1 game in cluster 4. Also, we can see that all the games in cluster 3 locates in the second half of the results. we can see that the team is improving their away game performance bit by bit. Therefore, we will underestimate their chance to maintain their position currently, but I will suggest our customers to get rid of this team.



Leave a Reply