A new column has been added in the overall season doc displaying player consistency.
For those who don't care about the technical explanation, please skip to the start of the images!
If the formula looks Chinese to you, please allow me to explain the proces of it below. Keep in mind that what concerns consistency, the close to 0, the better. The higher, the worse.
To calculate consistency, we look at the standard deviation of a player's performances. The higher this number, the more inconsistency between their matchday averages. That kind of speaks for itself. However, there's already an issue with this narrowed formula. A player who continiously gets 4/10, will be labeled as consistent - regardless of how donkey ass they are. Likewise the opposite is true - a player (mostly strikers) who get some amazing games with many goals, with a lot of normal games on the side, will look rather inconsistent, despite them being one of the best players of the team based on their average. We want to think of consistency in terms of helpfulness to the team. So we make a correction by adjusting the standard deviation with the subtraction of "a player's personal season average and the total season average (of all players)". This is so that players overperforming relative to the team, will benefit from additional consistency points, whereas players who are terrible will see their score plummet (upwards that is), even if they're consistently bad. On top of this, 2 more adjustments are made. First and foremost, the same completely arbitrary "total games played by Inter, divided by 3 and rounded up" is used, just like we do for having player's total average count towards the season list. Simply because we don't want low number players to influence the outcome of our entire season too drastically. This means that currently a player like Kolarov who would be terribly inconsistent, is not on the list. I'm putting this up for debate, but a caesura is imperative. Secondly and lastly, we've also put a hard cap of the total team average on 6. This means that if the total team average is lower, like right now (5,84), the formula will apply 6,00 when subtracting from a player's individual average, and not 5,84. That is because we see a 6,00 as the absolute average game. If the total team average is higher than that, good. That means we are doing good and the team is competitive and it feels absolutely essential to be weighted against the strength of the team. But when a team's performance dips below 6, we're starting to fish for players that are looking consistent relative to the underperforming team. While in absolute terms, even they might not be performing that well.
I hope all of this makes sense and if you have any additional question, opinions, ... about this, please let me know. I'm watching Brehme and Ethor. :datass:
- - - - - -
- - - Additions - - -
The submissions for this game can be found here:
https://docs.google.com/spreadsheets/d/1l213kicQaKDqnLcjGwyyLxizccg1z5bUJVasUCXHt68/edit?usp=sharing
Overall Season Doc:
https://docs.google.com/spreadsheets/d/1u6kI-Pjc7Am-LwpUUxa9-t-5a9tbeGQeveWdZJWzbyE/edit?usp=sharing
- - - END - - -
Hope everyone enjoyed participating and I hope you like the result.
Feel free to discuss below.