4-1-3-2 Samurai // Over 4 goals scored per UCL play-off game

First, I would like to thank user Augustin79 for the idea of ​​this experiment (I hope you will do the same). After posting my past tactics based on the principles of machine learning (it can be found here), he brought in a commentary table with the tactics of highly respected Knap and wondered what would happen if you train a model on them? And I got down to business.

The first thing to note is that all of Knap's tactics (who are not yet in the know) are tested for three types of teams (leaders, mid-table teams and outsiders), therefore, looking ahead, I will say that this will be my first of three publications on machine learning based on of this table. This particular tactic is intended for the top teams in their league, as well as to make your attacking players feel like a fish in water in front of someone else's goal.

I probably will not fully describe the process of training the model within the framework of this particular article (if anyone is interested, welcome to the comments, or to my previous publication, the link to which I left above), since it can be quite long and tedious.

But let's get down to business. After a long learning process and equally long testing at Liverpool, here are the results the team has achieved:

1) English Premier League:
League win - we suffered only two defeats (all of a sudden from West Brom and Brighton), scored 98 points (14 points ahead of Manchester United runner-up), scored 111 goals (2.92 per game on average, closest pursuer still the same - Manchester United - hit only 71), conceded 27 (not the best defense in the league, but who cares when the attack breaks everyone and everything). On xG, we also tore the league apart - 85.12 versus 67.35 for our other pursuer, City. In addition, we scored the most shots, shots on target and converted the highest percentage of them. Offensive dominance in the league as it is.



As for personal achievements: here we can note Roberto Firmino with 32% realization of shots on goal (16 goals from 49 shots in 24 games - he terrified the opponent's defensive lines), the three best assistants in the league - Thiago, Milner and Trent, three of them scored 33 assists (and evenly!), and Thiago separately for 0.55 assists on average per game and 49% cross accuracy.

2) Champions League
To say that Liverpool with this tactic staged a mayhem in the main club tournament in Europe is to say nothing. In seven playoff matches (two first-round matches with Ajax, the quarterfinals against Bayern, the semi-final against Juventus and the final against PSG) we have scored 30 goals. Once again - 30 goals in seven matches with far from the worst teams in Europe! That is an average of 4.29 goals per game. Fantastic result. Additional comments are superfluous.



3) Home cups
The Carabao Cup was conquered by us without any problems, but the FA Cup went to the blue part of Manchester. We lost to Manchester United in the semifinals thanks to Henderson's mistakes. It's a shame, however, it is worth noting that in test saves, Manchester United very often appears as an almost invincible rival.

In general, it should be said that even without any transfers, this tactic is ready to meet the challenges of top teams. And in the next article I will lay out the same tactics for teams from the middle of the table.