Amp curse only works on the base damage of CoD/CoA, not the damage added by your gear.
Right, which means you're getting an additional 2.1k damage to one out of three CoDs, meaning all are effectively increased by 700 damage or 16.67%.
Besides, deep affliction gets more dps out of CoA anyways.
Every modelling I've done has still shown CoD to eek out a little more DPS than CoA, even with the consideration of Improve CoA and Contagion in the mix.
What happens to us, however, is that affliction also has malediction -- so those folks end up doing CoE/CoS.
Depends on how deep you go I guess. With my present build of 42/0/19 I don't have Malediction as I went for Improved HoT for use in situations where it's more beneficial (like on Mags phase 1, in heroics, and general use while out farming).
Amp curse only works on the base damage of CoD/CoA, not the damage added by your gear.
Besides, deep affliction gets more dps out of CoA anyways.
What happens to us, however, is that affliction also has malediction -- so those folks end up doing CoE/CoS.
DPS from the curse alone, yes, total DPS, no.
With imp CoA a deep affliciton's CoA will deal more DPS than their CoD, however, CoA needs to be refreshed twice as often so the DPCT (damage per cast time) is actually lower than CoD. By replacing CoA with CoD you can spend the extra time you have from not casting CoA as often to spam SB's.
At some point in +shadow damage CoA will not only out DPS CoD but will out DPS it and the fraction of an SB extra you get from using it but this point in +shadow damage is over 1200 even for a deep affliction lock with every single CoA buffing affliction talent.
All the modelling I've done has said the opposite. But this isn't really the thread for it anyways . I could've made a math error, sure. For what it's worth Leulier's spreadsheet supports CoA > CoD for deep affliction, however.
I'd have to go back and re-do my own math to see if the SB point is valid. Although at that point I'm going to get even more skeptical, since there's always dot gaps anyways. But I guess the argument would be that CoA will have more gaps than CoD anyways.
Either way, I'm destruction and I use CoD :p.
Although interesting "side issue" with CoD is, say, if you're using it on Magtheridon.... and you're very close to the tank but trying to get as close as possible before hitting soulshatter because it's so early in the fight.... and your CoD goes off at *just* the wrong time....
Luckily all Mag did was cleave a rogue friend of mine when he turned to me, and I immediately shattered and no one else, including me, died. Vent was surprised though, since they hadn't thought I was over the tank. Well, I *hadn't* been, but..... .
*cough* will be more careful in the future, sir... *cough*
Leulier's spread sheet has CoD coming in ahead for total DPS vs CoA for all the attribute combinations I've been using in a heavy affliction build even with amp and imp CoA. The key word is, again, total, because CoA alone starts dealing more DPS than CoD alone long before it becomes the more efficient use of the global cool down. I'm using version 1.12 of the sheet sheet where amp only effects the base damage for CoD and SM and S&F have no effect on CoD damage too. It does have suc sack buffing CoD damage though, perhaps that is wrong?
Still, it's not a straight damage over duration comparison, you have to factor in the DPS loss from using more GCDs with CoA.
Leulier's spread sheet has CoD coming in ahead for total DPS vs CoA for all the attribute combinations I've been using in a heavy affliction build even with amp and imp CoA. The key word is, again, total, because CoA alone starts dealing more DPS than CoD alone long before it becomes the more efficient use of the global cool down. I'm using version 1.12 of the sheet sheet where amp only effects the base damage for CoD and SM and S&F have no effect on CoD damage too. It does have suc sack buffing CoD damage though, perhaps that is wrong?
Still, it's not a straight damage over duration comparison, you have to factor in the DPS loss from using more GCDs with CoA.
I stand corrected. As soon as I simulated the whole spell rotation CoD did come out ahead......
If you're 'fully aware' that the cap you are hitting is a function of your fixed inputs and not of game mechanics itself then why do you keep making statements like 'adding new locks after a certain number will not increase ISB uptime.' Adding locks with a crit rate higher than the average crit rate of the locks in the raid will always results in increase ISB uptime. The fact of the matter is the 'cap' is just an artificial result of your artificial inputs and the formula the program you have is using.
well, I hoped I dont have to explain in so many details , but maybe I did not explain well what i mean by 'cap'
If you have one affflic lock with about 17% crit, you have about 37% ISB uptime. If you add another similar afflic lock, ISB uptime jumps very noticeably to about 54% (around 17% more). Adding 3rd similar afflic lock would not noticeably change ISB uptime, it stays at about 54%. Adding 3rd demo lock with 20% crit instead will change ISB uptime, but only to 57%, which is increase of 3%. Adding another such 'above average crit) lock will increase ISB uptime to 58%....
So, let me repeat, and maybe it will be more clear what I think when i say 'cap':
- 1st lock (afflic with 17% crit) add on ISB uptime (from 0%) about 37%
- 2nd lock (afflic with 17% crit) add on ISB uptime about 17% more
- 3rd lock (afflic with 17% crit) does not add on ISB uptime, ie 0% more
- 3rd lock (demo with 20% crit) add to above ISB uptime about 3%
- 4th lock (demo with 20% crit) add to above ISB uptime about 1%
So we have +37%, +17%, +3%, +1% .... even when we were adding locks with more crit than average, it is very obvious that their effect on increasing ISB uptime is very limited (or "capped") after 3rd lock or so.
Originally Posted by tetracycloide
3% difference is not something I even remotely suggested between affliction locks. Considering that demonology and destruction not only get more crit% from talents but would also be stacking more crit on gear because 2 crit rating > 1 dmg for those builds where as in affliction 2 crit rating < 1 dmg then it is not even remotely valid to assume all three builds are running the same average crit rate. In the real world demonology and destruction are going to be running with at least 5% and 3% more crit respectively, anything less would dictate different gear levels.
affliction (or any other build) locks can and do have 3% or more difference in crit% between locks in same tree. Imagine that one of them decided to use Battlecast, other Spellstrike items - and similar choices on other items...it is very realistic to have 3% diff in crit% of two afflic locks with comparable level of gear
Originally Posted by tetracycloide
Yes there are. There is no such thing as a 'program simulation' that doesn't use formulas to generate results. Unless it's just a random number generator that is spitting out random results for you the program is taking the stats you give it, plugging them into a formula, and then evaluating it and returning a result. If you wrote the program then you should know exactly what algorithm it is using.
I think that yourself explained my point well with :
In all fairness your program doesn't use this equation explicitly in the sense that you never actually write it out that way in the code because it runs an interated loop which randomly generates ISB durations based on the crit rates and a random number generator.
and that is exactly what simulations are used for - they helps you to get close to real results without knowing or having full math model/formulas of situation, and later helps you to check your current math model and see why it gets different results, and improve your math model ... as you noticed and took into consideration that not all ISB end up with 4th SB.
All my modeling has CoA > CoD ever since patch 2.0.3.
Although...... Can CoD crit? My Warlock buddies never mentioned the possibility.
for this one, I agree with tetracycloide and others who posted that CoD > CoA for total warlock DPS. I used for that my own XLS calculation, where amplify curse affect only base dmg, CoA/CoD cant crit, and all +shadow% like SM or succ sacrifice influence CoA/CoD. I tested that demonology +dmg% like MD does affect CoD.
Originally Posted by Tahapenes
Unless the Affliction Warlock is way undergeared
as demo lock i'm most often 20% or so above 2nd lock in damage, and sometimes 50% or more above last. While my being 20% above 2nd lock is due to pet damage (ie my CoD is not better than his CoD, as would be expected), comparing to last afflic locks on majority of raids, my CoD does more damage .... therefore, sometimes you have afflic locks in raid with way less +dmg than demo/destro locks (not only on gear, but afflic lack +dmg from pet, talents etc, that also influence CoD).
In fact, I just checked on my XLS (Warlock DPS XLS) and if afflic lock would have ~1000-1100dmg, demo lock would need only 120 more +dmg to have better CoD - and its not hard to find such difference in gear.
So, while I agree that in majority of cases afflic should cast CoD, what i said originally stands:
lock with least total/average damage on CoD should cast CoS/CoE, and sometimes (although rarely) that lock can be afflic even with demo/destro in raid
stuff about 'cap' and diminishing returns to additional locks
We're on the same page, I think, in that we agree that additional locks suffer diminishing returns to the ISB uptime. I was a little thrown off by your use of the term 'cap,' however, because there is only a cap with a fixed average crit rate. Without a fixed average crit rate there is no true cap, simply diminishing returns to additional locks.
Originally Posted by nenad
is exactly what simulations are used for - they helps you to get close to real results without knowing or having full math model/formulas of situation, and later helps you to check your current math model and see why it gets different results, and improve your math model ... as you noticed and took into consideration that not all ISB end up with 4th SB.
For that I thank you. It's always nice to have someone else's work around so you can see how they tackled a problem. The single equation model and the simulation are built on the same math, one just includes variance and the other doesn't.
Originally Posted by nenad
well, I hoped I dont have to explain in so many details
Just for future reference I, at least, would really prefer it if you started with details and explained them afterwards as necessary, instead of the other way around. I'm perfectly capable of understanding the math in play here and I'm sure I'm not the only one.
Last edited by tetracycloide : 06/09/07 at 7:35 PM.
for this one, I agree with tetracycloide and others who posted that CoD > CoA for total warlock DPS. I used for that my own XLS calculation, where amplify curse affect only base dmg, CoA/CoD cant crit, and all +shadow% like SM or succ sacrifice influence CoA/CoD. I tested that demonology +dmg% like MD does affect CoD.
SM does not affect CoD. I'm affliction and my naked CoD does 4200 dmg (not 4620), so demo should always cast CoD.
well, I can not check if SM influence CoD since i'm not afflic, but that is very interesting information - although it does look like bug. Anyway, i fixed that in my XLS, and now for afflic CoA comes only few DPS below CoD in total damage .... while difference between CoD and CoA is more obvious for demo lock.
BTW, I tested once more if demo talents influence CoD, and confirmed again that all 4 demo lock +dmg talents influence CoD: Master Demonologist, Soul Link, pet dmg bonus, sacrifice dmg bonus
I removed all my gear and tested with different pets , and i got numbers like these:
I tried to compare those numebrs with my XLS when i put +dmg and +snadow dmg at 0 (and removed 130 from "added spell dmg" since i didnt use Fel Armor), to simulate "no gear", and i noticed small difference in numbers, that I tracked to difference in how my model in XLS calculate bonus pet damage ... for example, FG was +114 dmg, but it is not correct without gear.
Therefore i revised pet bonus damage model to include in calculation total lock HP and mana, which influence pet INT/STA, which influence that bonus +dmg - and now i have model where changes in lock HP/mana will accurately change pet bonus.
As a result, my new Warlock DPS XLS came within 0.1% of above numbers for simulated "no gear" situation.
But most important thing about this change (due to SM not affecting CoD) is what Bolche already posted above: if afflic and demo lock have similar gear, demo lock will do more damage with CoD
06/06/07, 2:37 PM
. I could've made a math error, sure. For what it's worth Leulier's spreadsheet supports CoA > CoD for deep affliction, however.
Similar Threads
