Best Possible Feral DPS in SimulationCraft - Page 3

Allev · 11/16/09, 11:26 AM

Ah, you're right. Probably impossible.

My only reservation about the 4.8% versus 5% is that more than half of hellord's parses showed his hit rating over 5% (with 5k-10k hits). The number of actual swings necessary to narrow down whether the number is 4.8% or 5% is something around 60,000 swings. Assuming a constant miss rate, we still don't have enough data, unless we got "lucky" and the number is further away (i.e. 4.7% or 5.1%) from the mean. In which case we might be even less certain about our "guesses".

Sidoine · 11/21/09, 4:07 PM

Originally Posted by Elentor

Ovale has FBN in its code:

if ComboPoints(more 4) and BuffPresent(SAVAGEROAR 8) and TargetDebuffPresent(RIP 8 mine=1)
Spell(FEROCIOUSBITE)

Might not be working at all though, I'll have to take a look in-game. But yeah, implementing the SimCraft list in Ovale would be sweet.

It was a syntax error. I fixed it in 3.2.19, thank you. When I tested this script on my PTR character, I forgot to put the glyphs so I was not able to reach this condition.

I made some other changes, I hope it is in a better shape now.

Starfox · 11/23/09, 6:50 AM

r3939 - simulationcraft - Project Hosting on Google Code
Feral profiles updated.
Also made the default actions the ones in the first posting.

Thanks for your work on optimizing the profiles.

Allev · 11/27/09, 3:51 PM

I think next phase is working on T10 profiles. While we can't necessarily figure out the gear yet (especially with stat caps being so important), we might want to look at doing a few things with the t10 profile:

- With less energy consumed by Rip with 2t10, we might see a damage increase by not glyphing for rip, or possibly shred for mangle-bots.
- With rake being such high DPE now, optimizing its uptime might be more important. It went from "slightly better than shred" to "far better than FB", and so it might see increased value.

I'll be experimenting with this this afternoon. I'm expecting different results from the shred and mangle profiles.

Allev · 12/02/09, 11:46 AM

No luck in producing positive results on either changing glyphs or switching priorities. Re-glyphing seems to be having a significant impact on Rip uptime; to fix that, some closer inspection of CP usage is probably necessary (i.e. figuring out a good way of specifying a "goal CP" for SR). It seems like the current priority list is also good enough for Rake uptime.

But, next topic is SR usage-- something is incomplete. It appears there's a bug in the following line:

actions+=/savage_roar,cp>=1,savage_roar<=1

I wrote this so that anytime SR is less than 1 second (or not up at all) and we have a CP, we'll SR. In theory, this should lead to no more than 6.2 second gaps in SR usage. However, the PTR page dynamic buff section shows that the SR buff isn't being applied until 7 seconds into the fight. We're also casting it every 11.7 seconds, which is probably too often, and why we see so few FBs.

Starfox · 12/02/09, 12:01 PM

Originally Posted by Allev

No luck in producing positive results on either changing glyphs or switching priorities. Re-glyphing seems to be having a significant impact on Rip uptime; to fix that, some closer inspection of CP usage is probably necessary (i.e. figuring out a good way of specifying a "goal CP" for SR). It seems like the current priority list is also good enough for Rake uptime.

But, next topic is SR usage-- something is incomplete. It appears there's a bug in the following line:

actions+=/savage_roar,cp>=1,savage_roar<=1

I wrote this so that anytime SR is less than 1 second (or not up at all) and we have a CP, we'll SR. In theory, this should lead to no more than 6.2 second gaps in SR usage. However, the PTR page dynamic buff section shows that the SR buff isn't being applied until 7 seconds into the fight. We're also casting it every 11.7 seconds, which is probably too often, and why we see so few FBs.

Dynamic Buffs Start  Refresh   Interval Trigger  Up-Time   Benefit
savage_roar   7.0    19.1      43.7sec  11.5sec  96%      96%

I think you mean this. That 7.0 means that the buff gets started 7 times on average, not that it gets started 7.0s into the fight on average.
SR just happens to drop about 6 times during the fight with this actionlist

Allev · 12/02/09, 12:16 PM

Ah. That makes a lot more sense-- big brain fart on my part.

Still, the really low trigger time bothers me...

Zelspawn · 12/04/09, 11:40 AM

Anyone know how you can get the simulator to model the engineering tinker: "Hyperspeed Accelerators"? Removing this enchant did not alter my DPS result (Removing any other enchant results in a lower DPS) Without being able to model this engineering tinker properly, you can't compare engineering properly against other professions.

Edit: I did some more thinking, and added a static 68 (?) haste to gloves to simulate them somewhat (340 times 12 seconds uptime divided by 60) This allows for a somewhat flawed modelling (Stacking Hyperspeed with say, Tigers Fury should yield more DPS) However with JC + Engineering I came to 10141 DPS only, which is far from the optimal (10323.3) Its also worth noting that I am using a slightly different gear set than what proposed by this simulator (I used the BiS I made with Rawr some time ago, it's almost the same, but has some slight differences and based on my professions) I will play around with this some more to see if can find anything useful with other profession. My initial question of a way to proper simulate the Hyperspeed enchant still stands.

Mihir · 12/08/09, 4:43 AM

Did you try adding something like

,use=340haste_12dur_60cd

to the end of the gloves line?

psuman99 · 12/08/09, 5:53 PM

Is the feral dps module working correctly. I have noticed on my own sims I have run (mostly 245 bis gear) that Str is more valuable than agility and for your bis T9 list it is significantly more valuable than agility. This appears to go against most of the other theorycrafting I have seen for the class.

Starfox · 12/08/09, 8:12 PM

Originally Posted by Zelspawn

Anyone know how you can get the simulator to model the engineering tinker: "Hyperspeed Accelerators"? Removing this enchant did not alter my DPS result (Removing any other enchant results in a lower DPS) Without being able to model this engineering tinker properly, you can't compare engineering properly against other professions.

Edit: I did some more thinking, and added a static 68 (?) haste to gloves to simulate them somewhat (340 times 12 seconds uptime divided by 60) This allows for a somewhat flawed modelling (Stacking Hyperspeed with say, Tigers Fury should yield more DPS) However with JC + Engineering I came to 10141 DPS only, which is far from the optimal (10323.3) Its also worth noting that I am using a slightly different gear set than what proposed by this simulator (I used the BiS I made with Rawr some time ago, it's almost the same, but has some slight differences and based on my professions) I will play around with this some more to see if can find anything useful with other profession. My initial question of a way to proper simulate the Hyperspeed enchant still stands.

Hyperspeed Accelerators works for me, it gets parsed and added to the profiles action lists if I download my druid:

...
actions+=/use_item,name=runetotems_gloves_of_triumph
...
hands=runetotems_gloves_of_triumph,stats=389armor_67haste_109sp_77sta_77int_59spi,gems=23sp_5sp,enchant=hyperspeed_accelerators
...

Odas · 12/09/09, 2:04 AM

I recently acquired the Deathbringer's Will trinket. It seems to have a 105 second internal cooldown and procs randomly one of +600 strength, 600 agility or 600 armor pen. I haven't had time yet to get a measure on the proc chance. Does anyone know of any good addons that track trinket stats (internal cooldown, uptime etc)? I used to use proculas, but it doesnt seem to be working anymore.

tagosa · 12/09/09, 2:15 AM

Originally Posted by Odas

I recently acquired the Deathbringer's Will trinket. It seems to have a 90 second internal cooldown and procs randomly one of +600 strength, 600 agility or 600 armor pen. I haven't had time yet to get a measure on the proc chance. Does anyone know of any good addons that track trinket stats (internal cooldown, uptime etc)? I used to use proculas, but it doesnt seem to be working anymore.

Procodile - Addons - Curse

ithecho84 · 12/09/09, 4:01 PM

That trinket is pretty interesting. I think without a doubt that the static ArP value would be beneficial to helping us reach the static ArP cap of ~1401, but we should try to determine its value without having any use for the ArP proc and see if perhaps another trinket may be a better choice. Taking into consideration that the procs have a chance of being rather streaky and at a 90 sec ICD, that could mean going an entire fight sitting at half the ArP cap. Therefore, I think we can agree that the ArP proc from the trinket should not be taken into consideration when gearing towards a soft/hard ArP cap. With that in mind, assuming the wearer is actually at such a cap, we can effectively say that the ArP proc from the trinket would be wasted. The last known stat values taken from the FbN simulations (slightly dated) can be used in this rough approximation.

We can boil down the proc to being a static 200 stat buff, seeing as how it is a 30 second duration with a 90 second cooldown with a 600 stat buff proc. Assuming stat values of:
Agi = 1.42
Str = 1.31
ArP, which is usually 1.49, will effectively be 0 for this proc, since in an ideal gearset, it should all be over hard cap.
Since we are looking to average out the value of the procs, we will average out the value of these stats as well. This comes out to ((1.42 + 1.31 + 0)/3=.91). Assuming it is roughly equivalent to a static 200 stat buff, we can multiply that by 200 and come up with 182. Now that we have the value of the proc, it's pretty simple to add the value of the static ArP value to it, and then compare that value to the value of other trinkets, derived in a similar fashion.
So I guess the rankings would be

Heroic Deathbringer's Will = 461.16
Deathbringer's Will = 412.95
Heroic Death's Verdict/Choice = 399.8
Whispering Fanged Skull = 356.24
Death's Verdict/Choice = 353.8
Herkuml War Token = 352.24

So I guess there are the new trinkets. Bit surprising about Whispering Fanged Skull. It's impossible to plug in values for crit and haste, seeing as how the value of crit declines the more you get it, and the value of haste increases with gear level. I used the simulated stat values with early t9 gear. Personally, I would say that the Herkuml War Token would be better than Whispering Fanged Skull in T10 gear levels, so keep that in mind. Death's Verdict will also probably edge out over Fanged Skull. Obviously simulations will provide a more accurate picture and RNG will make the actual DPS contribution of Deathbringer's Will drastically better or worse.

Odas · 12/09/09, 5:22 PM

Procodile is outdated and doesn't seem to be tracking anything. Its possible it hasn't been updated since the combat log change a while back.

Ithecho84

Averaging over a proc duration is not the most efficient method of estimating its value, especially for stats like armor pen. 600 armor pen thats up 10% of the time is better than 60 armor pen up all the time because armor pen scales faster than linearly.

I tried modeling the trinket as having 3 different procs each with a 270 second cooldown in Rawr and it still was coming pretty far ahead of other trinkets. It also had some pretty strange effects on the value of armor pen. It decreased the value of armor pen when over the soft cap, but less so than Mjolnir does. I wish I could edit this trinket into simcraft but that is beyond my knowledge at the moment.