Re: EvenMatchOmni
Posted: Fri Sep 02, 2022 8:28 am
Haha.That's some pretty code
switch (int(PPH == -1) + 2 * int(PastPPH == -1) + 4 * int(PastPPHMap == -1))
Let's break it down,
PPH is current PPH, which is your "current score" set after so many points/times
So lets say you're having a good day 400 pph.
Normally, if your 300 pph eg. pastPPH
So no map specific it rates you 350 pph
Lets say on a map you're a mino whore and typically get 450pph.
0.4 * (400 + 450+ 0.5 * 300)
0.4 * (850 + 150)
0.4 * 1000 = 400
Or say you're having a bad day at 250, but you're excited that you've got the mino headed to the dognuts spot on minus where you can kill JKT repeatedly for big points 450 PPH
0.4 * (250 + 450 + 0.5 * 300) = 340 PPH which looks reasonable.
Or say you're having a good day at 350, but you're excited that you've got the mino headed to the dognuts spot on minus where you can kill JKT repeatedly for big points 450 PPH
0.4 * (350 + 450 + 0.5 * 300) = 380 PPH which looks reasonable, but not much different than above.
Or say its a bad map for you, you only get 200 pph, because dammit I just don't want to play MassD EVERY FREAKING NIGHT, and it hasn't started
0.8 * (PastPPHMap + 0.25 * PastPPH) = 0.8 * ( 200 + 0.25 * 300) = 220. I think this isn't right. A bad map score really drags you down, maybe its my numbers and I think this is where we said lets just use overall PPH.
Good players should be good on most maps - Enyo I think said it
.... and I think its the camper stains that skew on maps (eg. ion spamming on TAL) and give them too many points.
For the most part players will regress toward the mean. Meaning, a great map/match will usually be followed by one more toward mean/average, conversely a really bad match likely means play will get better the next time.
So overall, I think either way though will work.
I also think yesterday the matches were pretty balanced, Enyo's MB rant aside most were competitive, unless they did something stupid... I know there were a few (eg. Valarna, Randomizer where my team WASTED the power vehicles... and that hurts badly on randomizers).
switch (int(PPH == -1) + 2 * int(PastPPH == -1) + 4 * int(PastPPHMap == -1))
Let's break it down,
PPH is current PPH, which is your "current score" set after so many points/times
So lets say you're having a good day 400 pph.
Normally, if your 300 pph eg. pastPPH
So no map specific it rates you 350 pph
Lets say on a map you're a mino whore and typically get 450pph.
0.4 * (400 + 450+ 0.5 * 300)
0.4 * (850 + 150)
0.4 * 1000 = 400
Or say you're having a bad day at 250, but you're excited that you've got the mino headed to the dognuts spot on minus where you can kill JKT repeatedly for big points 450 PPH
0.4 * (250 + 450 + 0.5 * 300) = 340 PPH which looks reasonable.
Or say you're having a good day at 350, but you're excited that you've got the mino headed to the dognuts spot on minus where you can kill JKT repeatedly for big points 450 PPH
0.4 * (350 + 450 + 0.5 * 300) = 380 PPH which looks reasonable, but not much different than above.
Or say its a bad map for you, you only get 200 pph, because dammit I just don't want to play MassD EVERY FREAKING NIGHT, and it hasn't started
0.8 * (PastPPHMap + 0.25 * PastPPH) = 0.8 * ( 200 + 0.25 * 300) = 220. I think this isn't right. A bad map score really drags you down, maybe its my numbers and I think this is where we said lets just use overall PPH.
Good players should be good on most maps - Enyo I think said it
.... and I think its the camper stains that skew on maps (eg. ion spamming on TAL) and give them too many points.
For the most part players will regress toward the mean. Meaning, a great map/match will usually be followed by one more toward mean/average, conversely a really bad match likely means play will get better the next time.
So overall, I think either way though will work.
I also think yesterday the matches were pretty balanced, Enyo's MB rant aside most were competitive, unless they did something stupid... I know there were a few (eg. Valarna, Randomizer where my team WASTED the power vehicles... and that hurts badly on randomizers).