Forum

Pretty sure those stats from high include pauses on the dpm and hpm, reducing those numbers on games that have a pause in them. Also doing these stats for high when some teams play against much worse teams than others (unlike prem), means they are even less representative

Quoted from Spycy

[…]

Indeed they are. There’s nothing to prove from them and as I said that’s one of the reasons I didn’t publish them in the first place. The only real reason I made this is because I liked Job’s work tbh.

Also about the pauses I don’t think there is a way to fix it, but if there is please tell me.

Don’t get me wrong, I really like the work you’ve done and I love looking at stats. Not sure how we’d better be able to make them representative, are the prem stats including pauses? If not then it’d be worth asking ceejaey to see how he got rid of that problem. Would be interesting to see which of the drops on medics are uber vs kritz/quickfix too and what heal percentage each person receives on average so you can see how much work they are doing in relation to how many heals they get. Idk if this is all possible, but would be interesting, and well done so far :)

Quoted from Chronos

[…]

Don’t get me wrong, I really like the work you’ve done and I love looking at stats. Not sure how we’d better be able to make them representative, are the prem stats including pauses? If not then it’d be worth asking ceejaey to see how he got rid of that problem. Would be interesting to see which of the drops on medics are uber vs kritz/quickfix too and what heal percentage each person receives on average so you can see how much work they are doing in relation to how many heals they get. Idk if this is all possible, but would be interesting, and well done so far :)

Once I’m done ‘reformating’ the stats for every class I’m thinking of adding more class specific logs such as sniper duels zon, combo picks for spy, dmg dealt/heals received for heavy and demo etc

feed me ideas~

also I didnt take your message badly lol you dont have to apology, the points you’re making are very true.

All of the good stuff like the prem stats have would be nice to see aswell. Like games played, hours played, raw kill and death number and so on.

Quoted from MARS^

[…]

pyro matters even less than these stats do anyway

Thanks for the helpful input. (no, nobody gives a shit)

Damage per kill, damage taken per death
Also, maybe introduce relative k+a percentage (how many kills were you participating in), and try to compare these?

Srry chronos, but detoed will be playing med for us in playoffs since he is clearly better med than u. Stats dont lie m8 4Head

Quoted from Steve!

The assist system in tf2 is completely dumb, please don’t think of it as a meaningful stat. 3 damage right before a kill > 90 damage someone else does earlier — pretty silly if you ask me

how’s that different from getting a kill off a 3 damage last hit, or getting said kill stolen when you did 200 damage to the guy? ka/d may not be great but I don’t see why it’s worse.

If we go further in pigeon-holing classes’ roles then I’d say classes like scout, soldier, sniper and spy should still be based off k/d and dpm because they are supposed to be getting most of the picks and cleanup hits, heavy and demo be based off ka/d, damage and damage per heals but heavily biased towards the damage stats, since their role is more DOT than picks and as the most healed classes it’s important to distinguish who does more with less, and the remaining classes with their own specific formulas as support classes.

(btw I think it could be interesting to use damage per heals based on heal % rather than total heals for a match since a player shouldn’t be seen as “sucking up heals” when in fact their medic got out more healing overall.)