When analysing vehicle statistics, its best that sll players statistics should be bracketed for their average win rate/Kill Death ratio and then compared to their relative performance in a particular vehicle.
So if Player A has an average K/D of 2.1, but in the BMPT they get a kill/Death of 5.1, then the BMPT would be overperforming.
Its easier to do for win rate.
You check all the players win rates, and compare the players average win rate, and their win rate in the BMPT.
Average win rate should be 49%, so of all the players who have a win rate of 49%, what is their win rate in the BMPT.
Win rate however is not a good metric for a game that can allow multiple spawns of different vehicles. As the win rate is determined by the performance of the whole lineup, not just one vehicle.
This way of viewing vehicle performance also shows how good the vehicle is for less skilled players. Usually, a slower better armoured tank performs well for weak players, but not as a good for very good players, with some exceptions.
Only recent playtime should be used as well, because balance evolves dramatically over time. A particular vehicle might have been very OP and gotten a lot of playtime, but is now very balanced, such as the Harrier GR1 and Hunter F6. Both were at one point very broken, now both are very weak since the missiles were neefed, flares introduced, and better missiles added.
Its a great metric for Air RB, as you only get a single spawn in a single aircraft, and you can compare both average winrate and average kills of the player vs the vehicle.
