GG(x) is the engine that computes metrics such as Performance Rating, Team Luck, Quality Play and Feeding. These metrics are derived from machine learning and a comprehensive statistical algorithm. Results are in a normal distribution with a mean value and a standard deviation with respect to the player's tier and position.
For example, Silver ADCs average 5.6 CS per game with a standard deviation of 1.4. 68.2% of Silver ADCs are within one standard deviation, averaging between 4.2 and 7.0 CS per game.
GG(x) makes a variable out of every event in a game that correlates to a game being won or lost and computes the data in its native function. Results are compared in both absolute and in relative terms to the player’s team and to the opponent’s team for validity. A massive amount of data is continuously updated and computed for up to date analysis and precision.
Performance Rating is a numeric indicator of how well you play in a game. The scoring system is from 0 to 10, with 10 being the highest score.
|Rating||Expected Win Rate|
|0.6||6% Expected Win Rate|
|1.7||17% Expected Win Rate|
|2.8||28% Expected Win Rate|
|3.9||39% Expected Win Rate|
|5.0||50% Expected Win Rate|
|6.1||61% Expected Win Rate|
|7.2||72% Expected Win Rate|
|8.3||83% Expected Win Rate|
|9.4||94% Expected Win Rate|
|9.9||99% Expected Win Rate|
Team Luck measures how well your teammates perform in a game. If your teammates perform very well, your Team Luck score will be high.
Quality Play is when you perform fairly well in a game. Your performance is determined to be a Quality Play when it is above the mean in your respective tier and position.
Feeding is when you perform really bad in a game. Your performance is determined to be a Feeding when it is a certain percentage below the mean in your respective tier and position.
How accurate is the Performance Rating?
Extremely accurate. Metric is accuate enough to pass a blind test performed by a Grand Master player.
How is Performance Rating calculated?
Performance Rating is calculated by creating a variable out of every data and event that happens during a game that correlates to a player’s skill. We also include things like, how well you performed to a lvl3 gank, how much focus you attracted from the opponents jungler so your other teammates were risk free of a gank, how your ward placement revealed the opponents jungler and really almost anything you can think of. Stats such as kills, cs, damage dealt, level and gold are obviously included as well. It is also important to note that Performance Rating is calculated relative to a player’s position and tier.
Does it take factors that are not saved as data into consideration?
Everything is taken into account except voice and chat. A comprehensive evaluation of a player’s game could be accurately examined with the immense amount of game data that is available and we have fine tuned it to perfection.
Don’t you just get a high rating if your KDA is high?
Not at all. KDA is just one of the many variables we take into consideration.
How is Team Luck calculated?
Team Luck is calculated by taking the performance ratings of the player’s teammates and comparing it to the player and to the opponents team at the MMR at which the game was played.
Does Quality Play and Feeding take position into account?
Yes. Position is taken into consideration for all GG(x) metrics.
This stuff only applies to low elo since there are things high elo players do that are not shown in the stats sheet. How does this work in high elo games?
The model we have developed works for any and all tiers. We are able to forecast end of season challenger SoloQ rankings based on previous performance ratings with a very low margin of error. We can even predict what tier you will finish the season at with extreme precision.
Where did this idea come from?
We wanted players to know how well they performed during their games to help players improve. We wanted to solve the player's pain of not knowing if and how much they were responsible for the outcome of a game.