Calculating a Meta Strength Score
Introduction
After reconsidering prior stances regarding how to properly gauge a champion's potential in the current meta, I think I've managed to come up with a way to in one number get an accurate description of that champion's power in-game. It would be calculated per champion per lane and would be filterable by tier. I make an assumption that to "main" a champion, you must have played at least 100 of your last 500 ranked games as that champion in that lane. Only players that fit this description shall be considered in the calculation for a given champion.
What It Is
Meta Strength Score is a number that would reflect how powerful a certain champion is to the current meta. Unlike win rate which has a tendency to normalize back to 50% over time, MSS will never auto-normalize and a champion will remain differentiated as long as they are in fact more or less powerful than the average. When nerfs happen then, you'll be able to see the uptick in MSS after the initial drop and know that it's actually reflecting how players are adapting to the changes. MSS is also going to reflect potency in counter-picking which alternative methods using teambuilder MMR would not.
The higher the MSS, the more potent a champion is. The lower the MSS, the less powerful that champion is. OP champions will have higher scores while UP will have lower scores. Overly niche champions will be assigned a lower score.
In other words, MSS calculations generate a single number that will tell you how the balance of a champion is. No more guessing about how the pick rate influences the win rate. No more guessing about the skill ceiling's relationship to the average player's skill. One number.
How to Calculate It
MSS = Meta Strength Score P = player count: number of players that main a given champion in a given lane R = match making rating: a player's MMR M = main factor: the percentage of the past 500 ranked games that have been played as a given champion by a given player A = average main factor: the averaged main factor for all players that main a given champion
MSS = (SUM( R x M )) / ( A x P ))
In other words, it's a weighted average of MMRs where the weight is the percent of games played as a champion by mains of that champion. This should accurately reflect a champion's power as players with higher MMR and higher play rates of the champion will be able to attribute higher volumes of their MMR to that champion.
How to Use It
After calculating each champion's MSS at each tier, you will generate the Average MSS across all champions that fill a given role. The AMSS in a given tier will be somewhere above the average MMR of that tier as we're dealing only with those that main champions. Once you have the AMSS, you simply perform MSS - AMSS which will give you how far above or below the average a champion is. The farther above, the more OP. The farther below, the more UP. In other words, MSS - AMSS tells you how much freelo you get from maining a champion.
Conclusion
Does anyone see any problems with a system such as this? If not, would it be reasonable to request that sites like lolking.net and champion.gg start displaying it? Alternatively, could it just be an API call that such sites can use to grab the data which would be calculated by Riot? The second option would be preferred as it would prevent needless parallel computation by external sources, prevent inaccuracies related to guessing a player's MMR, and shrink query count thus reducing load on Riot's servers.
TL;DR
MSS - AMSS gives you how much freelo maining a champion gives.
