In a follow-up to my previous post describing building a model to select a team of the year, here is the same model’s team of the 2010s. The model uses the formula: (Runs Scored — Runs Conceded) x (Wickets Taken — Wickets Lost) to select the 11 players that give this formula the highest value. Catches are considered 10% of a wicket and stumpings count as a wicket for the keeper and for the bowler. The only added constraint for the team of the decade being that players with fewer than 20 tests were not up for selection.
The data used was purely from test matches played from 2010–2019 and doesn’t include statistics from test matches outside that decade. With that in mind, here is the team selected by the model:
The team is clearly very good and shows how the formula values the amount of runs and wickets scored as well as the batting or bowling average. It doesn’t simply pick players with the best batting or bowling averages or with the most runs or wickets, but it chooses the combination that gives the best value for the team as a whole.
The fact that the side only includes teams from four nations reflects the dominance of the ‘big four’ test countries — not only in performance but in number of tests played.
De Villiers is the only player selected who could keep wickets and so must be the keeper in the side — this reflects both the model’s lack of constraints on having to pick certain types of players but also the fact that no specialist keeper dominated throughout the decade.
Selecting a ‘best team’ for any time period is always contentious and everyone has their opinions on players to include or remove from the team. However, this team shows that even a very simple model can be used to pick a team based on the simple principles of test cricket.
Comments