Tank statistics - based on selected range of player's WN7/8 (or battle)
This post was made as an issue
I would like to ask about an additional "small" feature when browsing tank statistics:
I understand, that shown data is based on how all players perform?
For example, I would like to remove all statistics from players with WN8 <1000 (or bad battles) so we could throw out all inexperienced players that are in some way lowering stats.
Same thing for very good players, I understand that these are somewhat average data, but I think it would be a new interesting statistic.
Especially for premium tanks, that are bought by very inexperienced players and greatly decreasing "ratings".
I don't know if there is possibility to calculate "player rating" from a single battle - if yes, thanks to that, we could set a filter to remove all battles that players performed really bad (and maybe really good too).
Why to calculate from single battle?
We could generate direct "player-tank effectiveness" from a battle and have new comparable data.
Other way, add player rating at a time when he battled, with his battle score.
Because when I greatly improve my "rating"(play style) after 1 mnth, it would be necessary to link my "bad" battle with my previous "skill", so data could be most accurate and real.
We could know data spread between different groups of players, now it's a one big bag with random data(?).
Set ranges per 250 points, 500 or maybe even 1k (maybe to big step) so it could be less detailed, and easier to calculate and store.
It's just an idea, I don't know how data is gathered, I don't know what information are send by players and processed by any mechanics you've created, so I can't predict how complex it could be to implement.
Oh I've forgot, what about to additionally separate actual player rating from his single battle rating:
my recent WN8 1500
my last battle with a "put a tank name" - 600 (yea I've screwed this one)
I could be counted differently:
ppl could remove bad battles (below WN 1000), and my battle won't be counted
or totally bad players (< 1000) but in this case, my bad battle could be counted in
But I think that overall player rating change too slow to rely on it (maybe last 30 days?), or maybe only focus on calculating single battle rating, that maybe could be more accurate.
I understand that everyone can have a bad battle, and average data is somewhat OK, but we still don't know, "how good we can maximally perform" on particular a tank.
What do You think about that?
Thank you for reading