I have been playing World of Tanks nearly every day since the start of this year. Unsurprisingly that has resulted me becoming better at it. I only have detailed statistics for the last 6 weeks, but during that time my WN8 rating went up every week, by an average of 4. As the WN8 is an average over all your battles, and I have 12k battles, the increase is slow. But the fact that it is going up steadily means that I am now pretty consistently playing better than I was before.

However I noticed that there is no correlation with my win percentage. I have been between 48% and 49% win rate since I started playing in 2011 (it's not 50:50, because a draw is basically a loss for both sides). In the World of Tanks community there is an everlasting discussion whether one should play to win, or whether it is better to play for your personal score, e.g. WN8. Quickybaby, probably the game's best known Twitch streamer, once took the time during a stream where people wanted to look at World of Tanks gameplay to show with an Excel table that the very best players have an average win rate of up to 70%, and the very worst players down to 30%; for him that proved that your skill had an impact on your winrate. Which at the extremes is certainly true.

But if you look at Quickybaby's own win rate of just over 60%, with him being quite good at the game, and playing to win, how many of his games did he actually influence? 60% win rate to me suggests that he didn't have enough influence to turn the other 40% of games into wins. And it would appear logical to me that this also means that of his wins, 40% would have been wins even without him trying. Only in 20% of his games, 1 in 5, did he have enough influence to affect the outcome.

Mathematically speaking, if the very best players have a positive influence on their win rate, and the worst players have a negative influence on their win rate, somewhere in between the curve must hit zero. And because skill is probably a Gaussian distribution, I would say that a vast majority of players are simply too mediocre to influence the win chance of their team in either direction. Me included. Playing to try to improve my personal score is the better option for me, because I have proven that I can do that. I am not convinced that I could play random battles without forming a platoon in a way where my win rate would be measurably higher.

Skill *does* have a marked effect on win probability, but mostly in the aggregate. There are mods that show you the sum of WN8 ratings of the players on each side. If one side has a significantly higher sum of WN8 ratings, it usually wins. But if the matchmaker puts you on the side with the lower skill, as a regular player you simply can't "carry the game" hard enough to make a difference.