[Math] Bloodlust/Heroism timing - Page 6

Wraithlin · 03/12/09, 11:53 AM

Double post but:
Along a similar vein and in reply to Kyths point on threat. In the early BL scenario wouldn't the optimum solution be for rogues and hunters to use their threat transfer abilities during the BL rather than before? (Not familiar with the mechanics of these abilities).

galzohar · 03/12/09, 1:24 PM

No, I was saying that if you add someone who doesn't execute then the executer is better off using a personal cooldown during execute. If that cooldown is raid-wide and they scale equally with it, though, then it doesn't matter when to use it.

MD is "next 3 attacks", so bloodlust/heroism by itself doesn't matter (although personal CDs stacked with it do matter). Tricks of the trade is duration-based, though, and thus benefits from bloodlust/heroism. Tricks of the trade though can be traded by rogues to increase their personal damage and trade threat so it effectively does nothing to their threat - whether or not that's actually a good idea depends on a lot of factors.

Jayde · 03/12/09, 1:46 PM

Originally Posted by Wraithlin

How many times must I repeat this?
There is (almost) no advantage to lining up a raid-wide bloodlust with execute. Infact, if you read Binks article there is no advantage to lining up trinkets with execute either, and all for the same reason (DPS increases are modeled as a general "dps increase" without being specific as to its source). Lining up DPS buffs with execute shortens the execute phase, so while you might get OMGDPS while everything is up, you lose a significant chunk of pain old execute-dps time.

That's not what I said in the slightest bit, so before you harm your forehead too much you should re-read what I questioned.

Looking at the Mage's personal buffs, if they have a trinket (or mana gem from set bonus) which is a static increase to spell power once per fight. We know they have this. This is a personal buff.

A static buff such as +300 spell power would be magnified both by Molten Fury and by Heroism in value. With Molten Fury due to the increasing returns on how much it adds per bolt during the uptime, and during Heroism by being able to fit more bolt in with the bonus during the uptime.

By splitting the two factors, the Mage will have to choose which of the two periods to use the static gain during--the execute phase, or the heroism phase. Either one will be better than just using using the cooldown randomly, and one will probably be better than the other. However, one could presume that for that personal trinket/buff/whatever specifically, it will be most amplified by being under the effects of Molten Fury and Heroism at the same time.

This is an issue of personal buffs, not raid-wide buffs. Which is different than what most of the posts in this thread have been dealing with. The generalistic raid-wide principle is quite straightforward and I'm not questioning that at all.

Mynak · 03/12/09, 4:13 PM

Jayde, yes you're correct that for a Mage it would be better personal DPS at that moment to use Mana Gem while Bloodlusted/MF. The thing is though that unless you're looking at a sub-2:00 kill that will only let you use that cooldown once, then that stacking that Mana Gem with everything won't matter.
Bloodlust is a 30% damage increase (1% haste = 1% damage).
Molten Fury is a 12% damage increase.

SITUATION A:
0:00 Fight starts
0:10 Mana Gem for extra Damage
2:10 (Boss is at 35%) Bloodlust/MF/Mana Gem for extra Damage
2:50 Boss dies

SITUATION B:
0:00 Fight starts
0:10 Bloodlust/Mana Gem for extra Damage
2:00 (Boss is at 35%) MF starts
2:10 MF/Mana Gem for extra Damage
2:50 Boss dies

At the end of the day, you either get in one Mana Gem with no extra benefit and a fully benefiting Mana Gem, or you get in two Mana Gems that are both benefited. Situation A = Situation B

galzohar · 03/12/09, 6:04 PM

We're kinda assuming without saying that you can use common cooldowns at least twice a fight, in which case Jayde's argument, which is valid for the "general" case (if you had such cooldowns) isn't valid. This is generally a good assumption considering fight durations and cooldowns excpet for exceptionally short fights or fights you overgear by a lot. Generall if theorycrafting for farm content becomes more work than theorycrafting for progression, I'd rather just theorycraft for progression and not spend so much time theorycrafting about stuff that may reduce maybe up to 1 sec off of fights that are on farm. But overall, yeah, if the fight is so short you only get one use of cooldowns you can factor in the inability to stack them on top of all the other considerations mentioned in this thread.

Wraithlin · 03/12/09, 7:01 PM

Originally Posted by Jayde

However, one could presume that for that personal trinket/buff/whatever specifically, it will be most amplified by being under the effects of Molten Fury and Heroism at the same time.

This is an issue of personal buffs, not raid-wide buffs. Which is different than what most of the posts in this thread have been dealing with. The generalistic raid-wide principle is quite straightforward and I'm not questioning that at all.

We have done the math for multiplicative and additive buffs, there is almost no change in fight duration when you start using additive buffs in execute range. Stacking execute with dps buffs, any dps buffs, trinkets, bloodlust, ANYTHING, does not noticeably shorten the total fight duration. Execute, in effect, does not matter when it comes to stacking buffs, ignore it, pretend it doesn't exist. It is not something you need to align with external cooldowns.

It is something we SPECIFICALLY modelled, and lining trinkets up with execute (in a 0 shaman raid) does not change the fight duration. The problem is you are obsessed with peak dps, when what matters is fight duration; stacking dps buffs with execute has no noticeable effect on fight duration, stacking buffs with each other does.

To give an example:
Super trinket gives 200dps mage + 100dps for 10s, MF is +10% dps, 20k boss health.

Late Tinket
65s @ 200 = 13k
10s @ (200+100)*1.1 = 330 = 3.3k
16.8 @ 200*1.1 = 220 = 3.7k

Total time = 91.8s

Early Trinket
10s @ 300 = 3k
50s @ 200 = 10k
31.8 @ 220 = 7k

Total time = 91.8s

Notice anything ?

[ee]
Lets add in Ronny the Rogue, doing 100dps, no execute buff, no trinket, 30k boss health:
Late Trinket
65s @ 300 = 19.5k
10s @ 100 + 300*1.1 = 430 = 4.3k
19.4 @ 320 = 6.2k

Total time = 94.4s

Early Trinket
10s @ 400 = 4k
51.6s @ 300 = 15.5k
32.8 @ 320 = 10.5k

Total time = 94.4s

galzohar · 03/13/09, 9:37 AM

Yes but if the rogue and mage are teaming up on the same boss, if the mage uses his cooldown late rather than early the kill would get sped up, and if the rogue uses his cooldown early rather than late the kill will also be sped up. Fastest kill would therefore be if the rogue used his cooldown early and mage used his cooldown late. This result is different because both are fighting the same boss and can help/hurt eachother by changing the % of the time spent in execute range, while if they fight separate bosses they can't do that and then the timing of the cooldowns doesn't matter. If they're fighting the same boss and have to use their cooldowns together, then what matters is how big is each one's cooldown - if the rogue's is bigger it should be used early, and if the mage's is bigger it should be used late.

Handyhoof · 03/13/09, 10:33 AM

That's not what we've shown here. Using Wraithlin's numbers above,

Rouge Early, Mage Late

10s @ 600 = 6k
8s @ 500 = 4k
10s @ 650 = 6.5k
6.48 @ 540 = 3.5k

34.48 s

Rogue Late, Mage Late

20s @ 500 = 10k
10s @ 760 = 7.6k
4.44 @ 540 = 2.4k

34.44 s

No appreciable difference, and we've already shown that Mage Early/Late doesn't matter. It's algebraic, but I thought the numbers might help.

Edit: I read his labels wrong. The math is correct, but this actually shows 2 identical mages and a rogue. Leaving original post as is so MatsT's makes sense.

galzohar · 03/13/09, 11:40 AM

With each doing 100 DPS and trinket adding 200 DPS, and MF at +10%, I don't see how you're getting those DPS values. Heck 760 or 650 DPS is plain impossible with his numbers, even with both using trinkets at molten fury would result in 630 DPS.

I don't see much point in posting a simulation but if you run a proper one you would see that rogue early and mage late provides the fastest kill, and it makes sense too. Maybe I will post an excel file later on.

MatsT · 03/13/09, 1:02 PM

Originally Posted by Handyhoof

That's not what we've shown here. Using Wraithlin's numbers above,
.. Wall of random numbers ...

Not sure how you can claim you are using Wraithlin's numbers since your numbers are all different. Actually using his numbers but with the rogue also receiving a trinket of +100 dps would yield.

Rogue trinket early, mage trinket late:
10s @ (100+100) + 200 = 400 = 4k total
51.67s @ 100 + 200 = 300 = 15.5k total
10s @ 100 + (200+100)*1.1 = 430 = 4.3k total
19.4 @ 100 + 200*1.1 = 320 = 6.2k total
Total time = 91.07s

Both trinket late
65s @ 100 + 200 = 300 = 19.5k total
10s @ (100+100) + (200+100)*1.1 = 530 = 5.3k total
16.25s @ 100 + 200*1.1 = 320 = 5.2k total
Total time = 91.25s

Results are as expected. Please make sure to double check your calculations before posting them.

Wraithlin · 03/13/09, 5:19 PM

Originally Posted by MatsT

Not sure how you can claim you are using Wraithlin's numbers since your numbers are all different. Actually using his numbers but with the rogue also receiving a trinket of +100 dps would yield.

Rogue trinket early, mage trinket late:
10s @ (100+100) + 200 = 400 = 4k total
51.67s @ 100 + 200 = 300 = 15.5k total
10s @ 100 + (200+100)*1.1 = 430 = 4.3k total
19.4 @ 100 + 200*1.1 = 320 = 6.2k total
Total time = 91.07s

Both trinket late
65s @ 100 + 200 = 300 = 19.5k total
10s @ (100+100) + (200+100)*1.1 = 530 = 5.3k total
16.25s @ 100 + 200*1.1 = 320 = 5.2k total
Total time = 91.25s

Results are as expected. Please make sure to double check your calculations before posting them.

Two points.
First, both your scenarios are inferior ones in which both mage and rogue trinket under BL. So while there is a benefit we can discount the early-trinket+late-trinket scenario as being inferior to double trinket + BL. If anything were to be concluded it would be that BL early is better as the trinket will likely be back off CD by the time you hit execute.

You have also assumed a fight length shorter than trinket CD, otherwise we could double trinket early and late. And we already know that simultaneous trinkets cause no change in fight duration.

Secondly:
The difference in those two is 0.19%, or 0.18seconds. I can live with calling that "a wash". Particularly because almost all practical considerations (people being alive, trying to align cooldowns, etc) will early bloodlust and likely more than compensate for the 0.19% difference.

I will add further that the difference is 0.19% in a scenario in which trinkets give a 100%/50% increase in dps. As trinkets usually give a much smaller boost we can confidently predict that the difference would be much smaller, perhaps as small as 0.02%. In fact it is so small it is even less important than the relative haste scaling brought up by songster.

Wraithlin · 03/13/09, 5:22 PM

Double post:
I will be turning a lot of this thread into a TTT article, I will try and acknowledge the posts of everyone where they are used, but please do make sure I don't forget you. I will put another post here in a few days when I have written everything up.

Wraithlin · 03/30/09, 1:14 AM

For those who followed this thread, the TTT article, along with SimCraft results are now posted.