Post Your Augmented Items IV: NPC it.

[pchan]

Player A: 450 900 1350 1800 2250 2700 3150 3600 4050
Player B: 528 1056 1584 2112 2640 3168 3696 4096
What accuracy should I be assuming here? 80%? 95%?

Let's say both have ~80%ish accuracy, so they both miss 1 swing, mob dies at 3600. Now we give Player B 1% more haste. We can agree the swing that used to be at 3696 is now at ~3659, which is still after the mob dies. The other swings going earlier affect nothing, as the mob still dies at 3600.

You don't understand what's going on. As I said, in order to compare setup1={PlayerA: 450, PlayerB : 528} where Player A has 450 delay and player B has 528 delay to Setup2={Player A : 450 , Player B : 528*0.99=522}, you have to examine all possible sequences (of any length) of randomly generated numbers that can be 1 with probability 95% and 0 with probability 5%. Yes I'm taking 95% accuracy, the actual number doesn't really matter, the only thing is that it is not 100% or 0%. So let's take a sequence :

seq={1,0,0,0,1,0,1,1,1,1,0,0,0,0,1,0,1,1,1,0,1,1,0 ,0,0,1,1,1,0,1,1} for instance

This represents the sequence of hits and misses as they appear chronologically (in the log). Of course, any sequence can happen, even the sequence with 56 '0' or a sequence with 17 '1'.

For each sequence, you need to calculate the time of death, which, in our case, is the first instant where 12 '1' have appeared. Now, and I know from your earlier posts that you agree with this, for a given sequence, the first time to reach 12 '1' for setup1 (call it T1) is always greater or equal to the one of setup2 (=T2). Since any sequence can theoretically happen (hi2u non capped accuracy), all you have to do to prove that on average T1 > T2, which means probabilistically that the expectation value of T1 is > the one of T2, is that T1>T2 for just one sequence :

This is fairly easy and many sequences will happen. Of course a lot of those have low probability, depending on how close the 1st and 2nd delay are, but they still exist.

very trivial way to see it :
all the sequences where player A misses his first 20 hits and player B hits his first 20 hits will have T1>T2 obviously ! Of course they have low probability, but they exist. In your case however, all of the sequences with 3 misses will work :
Let's take seq={1,1,1,1,1,0,0,0,1,1,1,1,1,1,1} for instance. The probability that exactly 3 miss happen in the first 15 hits is 3% . So when does the mob die ?
setup1 : 450 [528] 900 [1056] 1350 [miss@1584] miss@1800 [miss@2112] 2250 [2640] 2700 3150 [3168] 3600 [3696] -> death
setup2 : 450 [522] 900 [1044] 1350 [miss@1566] miss@1800 [miss@2088] 2250 [2610] 2700 [3132] 3150 3600 [3654] -> death

As you see the mob dies @ 3696s for setup 1 and @ 3654 for setup 2, in this case setup2 is strictly better. In case you argue that this sequence is unlikely, well, we can always consider the sequence of only consecutive '1' :
seq={1,1,1,1,1,1,1,1,1,1,1,1}. Unlikely ? nope, this has probability 54% ! When does the mob die in this case ?
setup1 : 450 [528] 900 [1056] 1350 [1584] 1800 [2112] 2250 [2640] 2700 3150-> death
setup2 : 450 [522] 900 [1044] 1350 [1566] 1800 [2088] 2250 [2610] 2700 [3132] -> death

Ding the mob is dying at 3150 for setup1 and 3132 for setup2. Guess what ? Haste is faster again.

Not satisfied ? Let's examine the setups where only one miss happens :
seq={1,1,1,1,1,1,0,1,1,1,1,1,1} probability 35%
setup1 : 450 [528] 900 [1056] 1350 [1584] miss@1800 [2112] 2250 [2640] 2700 3150 [3168] -> death
setup2 : 450 [522] 900 [1044] 1350 [1566] miss@1800 [2088] 2250 [2610] 2700 [3132] 3150 -> death
zing setup2 is faster again...

Oh and yes the setup with 2 hits gives the same time, but all others don't, proving my point : on average the second setup is faster.

[FailureMidgard]

Post I didn't even read

I've never once said haste doesn't lower the kill time on average.

not to mention a higher base dmg has pretty much zero chance ( unless its stacks by huge amounts) of making you do more damage on a merit mob for example ( or mobs with similar kill time). your parser will say you do more, but its not the case: it only increases your "overkill" damage typically, unlike delay/haste/acc, that increase your kill speed.

No, every stat gives you a "chance" to make it die faster, and a a very likely scenario that it does nothing on any given mob. a 1% difference in anything isn't noticeable, but when you stack them all up from various sources on various stats...sure, some mobs will die same speed and get overkilled more, and some will die faster and save swings. Haste/delay/acc are in the same boat as damage and attack here...

No, haste reduces the killing time even if it doesn't save you a hit. Attack doesn't. What's better to PL a level 5 job (aka mobs that die in 2-4 hits) ? Double march or double minuet ? The answer is double march obviously.

There's the first 3 posts as to why I got into this discussion. I don't even know what you're arguing anymore, you're agreeing with me that on average haste is better now, but you and mdkuser originally stated that haste always reduces kill time, which is all I am arguing against. I agree haste "always lowers the average" kill time, so does attack though. 1 attack will "on average" lower the kill time.

Why do I even respond to you...stop arguing with me when you agree with me.

[Karbuncle]

You both must be extremely bored, today, and yesterday, for hours on end.

[Fievel]

You both must be extremely bored, today, and yesterday, for hours on end.

.

[pchan]

1 attack will "on average" lower the kill time.

sigh

[Aeowin]

stop arguing with me when you agree with me.

That would be too easy.

[Chopx2]

That would be too easy.

^

Pink IS Green.

[Snick]

Jesus christ, no wonder i lurk only.

[Kuuhalee]

http://www.moofah.com/temp/media/ima...arwolf-acc.jpg

Hurray! Something useful!

[noodles355]

What's this..? And Augment in the Augment thread?! My god.

Nice btw.

[Arximiro]

http://www.moofah.com/temp/media/ima...arwolf-acc.jpg

Hurray! Something useful!

Sexy :D Grats

[Despain]

Grats on that belt. I've been dying to get something even at some level of equivalency to that belt for quite sometime :/ but I gave up on augmenting.

[FailureMidgard]

http://www.moofah.com/temp/media/ima...arwolf-acc.jpg

Hurray! Something useful!

I dropped one identical to that the other day

[Despain]

I dropped one identical to that the other day

Oh you.

[Sanjurio]

Finally something better than shitty
http://img43.imageshack.us/img43/579...0117233213.png

[Despain]

NQ Earth Staff.

Spoiler: show

I'm so bad

[StevenW]

Finally something better than shitty
[img]http://img43.imageshack.us/img43/5790/img20100117233213.png[img]

I want to transfer servers and jump in that ToAU shout group. No one does AMs on Bismarck. That happens to be the mission I'm on too.

[Kilorilo]

is he doing it right
http://img31.imageshack.us/img31/9699/earp.jpg

[Zer]

Even my WAR would give that Ragnarok a better use

[noodles355]

You trying to say he's gimp? looks like pretty obvious town gear to me.