Blue Vlag kill rate is lower than explained by formulas

[yarite]

In addition to joeylamcy's thought, I would guess there is possibly some time lost during retargeting... I suggest testing with an enemy with much, much higher hp so you can test the same enemy over the entire period... I'm interested in what you find

(Keep in mind, this is my best guess off the top of my head...not an official answer LoL)

Thanks! That's interesting factor.
Ok, I'll re-distribute STR/DEX and do statistical measurements on Cyclops today.
That will minimize re-targeting and last-hit influence.

[blitzcraig]

Just remember I'm only guessing there
Post what you find... favorite thread in awhile LoL

[yarite]

Well, additional measurements on Blue Vlag done, and now everything looks more clear.

Experiment 2 _______________________________
Swordsman, STR:219 DEX:48, no special abilities

1. 4 kills per 64 sec
2. 4 kills per 62 sec
3. 4 kills per 63 sec

Sword speed is 7, so it should take 88-91 strikes to kill 4 Blue Vlags of 12000 total HP
21 full strikes and 1 final strike for each Vlag

expected DMG = ( 219 - 80 * 9/11 ) = 153.5
but with such damage it should take only 19 full and 1 final strike

so looks like 8 possible strikes are wasted here, maybe that is "retargeting time"


Experiment 3 _______________________________
Archer, STR:371 DEX:48, no special abilities

1. 7 kills per 58 sec
2. 7 kills per 58 sec
3. 7 kills per 58 sec

Bow speed is 7, so strikes number is 82 to kill 7 Blue Vlags of 21000 total HP
11 full strikes and 1 final strike for each Vlag

expected DMG = ( 371 - 80 * 9/11 ) = 305.5
again, that damage should kill Blue Vlag in 9 full and 1 final strike

and again, looks like 14 total strikes are wasted, 2 per Vlag.

So, it looks that retargeting "eats" 2 strikes (if DMG formula works correct).

[yarite]

So far using following formula for estimating XP per Minute for each monster:

XPperMIN = (STR_MAX - DEX - OPP_DEX*9/11) * OPP_XPperHP * (HITperMIN - 2 * TRUNC((STR_MAX - DEX - OPP_DEX*9/11) * HITperMIN / OPP_HP ))

[blitzcraig]

I'm interested to see this same experiment with monster with higher hp to test "retarget" idea...it's a theory I had when I was doing basically the same tests you are now so I'd know best places to grind, but never had time to fully test after I started working on maps...I had forgotten the whole idea until this thread

[yarite]

I've reconfigured wrong char for cyclops - 75lvl swordsman, too weak for testing : )

So I've tried on Sapphire Dragons today

Experiment 4 ______________________________
Archer, STR:385 DEX:48, no special abilities used
Dragon, ARM:140 DEX:30 HP:6000

1. 4 dragons per 67sec
2. 4 dragons per 70sec
3. 4 dragons per 68sec
4. 6 dragons per 104sec
5. 6 dragons per 105sec

expected Archer average DMG = (385 - 140 * 9 / 11) = 270
expected full strikes per dragon = TRUNC(6000 / 270) = 22
expected total full strikes per 6 dragons = 22 * 6 = 132
expected total strikes including final strike = 132 + 6 = 138

with bow speed 7, 105sec is time for 150 strikes, but only 138 strikes required to kill 6 dragons completely

so again we have "lost strikes" = 150 - 138 = 12 (again 2 for each retargeting?)
OR
we have wrong working formula with less real damage

[yarite]

Well, kind of bad news for "theory of retargeting" : )
I followed recommendations and tested DMG on Cyclops

Experiment 5 _______________________________

Archer, STR:359 DEX:76, no special ability used
Cyclop, ARM:150 DEX 70

1. 1 cyclop killed per 97sec
2. 1 cyclop killed per 96sec
3. 1 cyclop killed per 97sec

expected Archer average DMG = (359 - 150 * 9 / 11) = 236
expected full strikes per Red Cyclop = TRUNC(30000 / 236) = 126 , or 127 total strikes to kill

with bow speed 7, 96sec is time for 137 strikes, so we have clear inconsistency of 10 strikes (or 2360HP DMG)

OR, if we calculate real average DMG for 137 strikes, it's 216 - that's 20HP less then expected per strike. Like if STR is 20 points lower!

Once I LVL up few times and increase STR, I'll observe the influence on Expected/Real average DMG, and catch the dependency. But even now it's clearly visible that real DMG is much lower than expected according to formula.

[blitzcraig]

Hmmm.... interesting... just to be clear, how are you timing these tests?

[profa]

oh wait, in your excel, u mean the skill is always active ?

If you select skill it will be used as in the game, so calculator will use it for 20 seconds, and then cool down for 20 seconds, etc.

[yarite]

Hmmm.... interesting... just to be clear, how are you timing these tests?

I use precise stopwatch, and press Start/Stop button when begin/finish attacking Cyclop.
Human reaction delay is about 0.75 sec, and it happens both times so shouldn't cause major error, maybe 1-2 strikes only, but not 20.

I just spawn single Red Cyclop, wait until it comes to close range, then press Attack and stopwatch button at the same time. I press stopwatch Stop button when Red Cyclop HP bar disappears and he's start falling down the floor.

I suspect that if I had Archer with 30-40HP for experiment with few Cyclops in a row - it would also show lower average DMG, but can't proove that yet. All I can do now is just try how it works with higher STR, to see if the dependency is linear or not.