Statistic Analysis
Although I dont have the answer to my question (at least, up to this precise moment
), I am now able to post some interesting results.
As mentioned before, the groups were not balanced, Group A and B are harder, than Group C and D.
So, I decided to do the following, inserted all the data in an excel file, and made a few calculations. How can we measure the difficulty of each group? I had an idea! Calculated the average % of wins (%W) and average % of wins + draws (%W+D) for all the tactics used on each group, achieving the following results:
View attachment 210735
From the table above, its possible to see that the tested tactics only won 44% of the total matches played, while on Group B, the tactics won 63%, 62% on Group C and 71% on Group D. The same happened with Wins + Draws, being Group B harder than Group C, on this parameter.
From the data inserted, I could see that the tactics never repeated the same group on 2 different rounds! In order to be able to evaluate the performance of each tactic, I decided to do something like this: If Tactic Proxy, for example, only has 1 participation, and it was on Group D, it was expected to have won 71% of the matches played and won or draw 86% of the matches played. Summing all the games made by Proxy, we have a total of 20 matches (18 from the league + 2 from the cups), from which they have won or draw 85% and won 65%, which is -1% than the expected %W+D (86%), and -6% than the expected %W(71%). In average, it has a performance of -2,5% when compared to the average!
If a tactic had 2 participations, it will consider as expected performance the average of each Group, and make the same calculations.
I've also included a % of Cup Achievements, that takes into consideration the maximum number of cups (2, i.e. playoff + world league cup), and winning 1 cup will be valued as 100% of the goal, and being in a final will be considered 50%. Koflok's Tactic, for example, has 1 participation, which means that the maximum number of cups he could win is 2. It won 1 and didn't managed to reach the final on the other, so it has a 50% achievement (1 x 100% + 0) / 2 = 50%.
With all this, we are able to see the following:
View attachment 210738
Koflok's tactic (with only 1 participation) is the one who did better, due to the fact that it did the best result in the harder Group (A!). It has an average performance of +16% compared to the average.
Xenon, was one of the best defenses, therefore had the highest %W+D and one of the best %W, and managed to reach 4 Finals (although its also the team with the biggest number of participations: 4, 1 in each Group).
JPWoody2 did well but didnt excelled in anything in particular (except maybe the defense, with 1,0 Goals Against per match).
Brobs has the highest number of Goals Scored per Match (3,2) and one of the best %Won.
Vodu has very similar stats to Brobs, and has the best %Won, and one of the highest number of Goals Scored per Match (3,1). Both Brobs and Vodu had nice defenses, but still above 1 goal per match (1,1).
Regarding the remaining tactics, considering there are 8 teams being tested each time, I've placed a Yellow Line marking the 8 best tactics, and also, marking the tactics that managed to get a rounded % of 0 or Positive, while the tactics below the yellow line (Biggest Loser lol) had performed worse.
In my opinion, FOR THE FINAL TEST, there should be a single League, with the best 8 tactics, and the best 8 teams of the world. All vs All, home and away. BUT MEANWHILE, we could keep using the current model, so we get more tactics tested and maybe other tactics will get above the yellow line!
I also think that the tactics that only had 1 chance, that are above the yellow line, or close to it, should have a 2nd chance! (Koflok, Knap, Raikan, Matti, Satorox, Proxy, BetterHalf and ZeroSea, 8 in total!)
Cheers,
Although I dont have the answer to my question (at least, up to this precise moment
), I am now able to post some interesting results.As mentioned before, the groups were not balanced, Group A and B are harder, than Group C and D.
So, I decided to do the following, inserted all the data in an excel file, and made a few calculations. How can we measure the difficulty of each group? I had an idea! Calculated the average % of wins (%W) and average % of wins + draws (%W+D) for all the tactics used on each group, achieving the following results:
View attachment 210735
From the table above, its possible to see that the tested tactics only won 44% of the total matches played, while on Group B, the tactics won 63%, 62% on Group C and 71% on Group D. The same happened with Wins + Draws, being Group B harder than Group C, on this parameter.
From the data inserted, I could see that the tactics never repeated the same group on 2 different rounds! In order to be able to evaluate the performance of each tactic, I decided to do something like this: If Tactic Proxy, for example, only has 1 participation, and it was on Group D, it was expected to have won 71% of the matches played and won or draw 86% of the matches played. Summing all the games made by Proxy, we have a total of 20 matches (18 from the league + 2 from the cups), from which they have won or draw 85% and won 65%, which is -1% than the expected %W+D (86%), and -6% than the expected %W(71%). In average, it has a performance of -2,5% when compared to the average!
If a tactic had 2 participations, it will consider as expected performance the average of each Group, and make the same calculations.
I've also included a % of Cup Achievements, that takes into consideration the maximum number of cups (2, i.e. playoff + world league cup), and winning 1 cup will be valued as 100% of the goal, and being in a final will be considered 50%. Koflok's Tactic, for example, has 1 participation, which means that the maximum number of cups he could win is 2. It won 1 and didn't managed to reach the final on the other, so it has a 50% achievement (1 x 100% + 0) / 2 = 50%.
With all this, we are able to see the following:
View attachment 210738
Koflok's tactic (with only 1 participation) is the one who did better, due to the fact that it did the best result in the harder Group (A!). It has an average performance of +16% compared to the average.
Xenon, was one of the best defenses, therefore had the highest %W+D and one of the best %W, and managed to reach 4 Finals (although its also the team with the biggest number of participations: 4, 1 in each Group).
JPWoody2 did well but didnt excelled in anything in particular (except maybe the defense, with 1,0 Goals Against per match).
Brobs has the highest number of Goals Scored per Match (3,2) and one of the best %Won.
Vodu has very similar stats to Brobs, and has the best %Won, and one of the highest number of Goals Scored per Match (3,1). Both Brobs and Vodu had nice defenses, but still above 1 goal per match (1,1).
Regarding the remaining tactics, considering there are 8 teams being tested each time, I've placed a Yellow Line marking the 8 best tactics, and also, marking the tactics that managed to get a rounded % of 0 or Positive, while the tactics below the yellow line (Biggest Loser lol) had performed worse.
In my opinion, FOR THE FINAL TEST, there should be a single League, with the best 8 tactics, and the best 8 teams of the world. All vs All, home and away. BUT MEANWHILE, we could keep using the current model, so we get more tactics tested and maybe other tactics will get above the yellow line!
I also think that the tactics that only had 1 chance, that are above the yellow line, or close to it, should have a 2nd chance! (Koflok, Knap, Raikan, Matti, Satorox, Proxy, BetterHalf and ZeroSea, 8 in total!)
Cheers,
Last edited:
