What we need is some good math that explains the value of the SS Glyph relative to the duration of your diseases. The value of the glyph goes down as duration increases (e.g., if diseases last 10 minutes you wouldn't use the glyph at all, right?). Likewise, as disease duration gets longer the probability of them being refreshed by the glyph increases, so each incremental extension in duration is less valuable. So Epidemic is actually less valuable than it was a few days ago, for people who have the Glyph.
I agree that there is a point where extending the length of diseases would be devalued but I don't think even at 21 or 24 seconds we are close to it. Why over-complicate it? With the glyph your concern is simply IT+PS less and SS more. With an extra 3 seconds we will most definitely be SS'ing more. If anything the value of both the glyph and the talents have gone up.
Also, even if you aren't using the SS glyph you are guaranteed 100% uptime if you are playing properly and so you won't have to worry about EP dropping off and losing raid dps.
I agree that there is a point where extending the length of diseases would be devalued but I don't think even at 21 or 24 seconds we are close to it. Why over-complicate it? With the glyph your concern is simply IT+PS less and SS more. With an extra 3 seconds we will most definitely be SS'ing more. If anything the value of both the glyph and the talents have gone up.
Also, even if you aren't using the SS glyph you are guaranteed 100% uptime if you are playing properly and so you won't have to worry about EP dropping off and losing raid dps.
There isn't a threshold at which it's devalued; it's a diminishing gains function. Thus, extending diseases from 21 to 24 seconds is in fact less useful than extending them from 18 to 21 (for example). So the question is, what is the actual dps value of those additional 3 seconds, and thus what is the actual value of that point in Epidemic?
EDIT: I'm only talking about in the context of having the SS glyph.
I'm on the other side of the fence with your theory. As the length of diseases increases and gives you greater # chances to refresh it, then the value of the SS glyph increases. Until the % chance of the diseases not being refreshed is < 1-2%, then longer duration is better. With 5 SS (the number you can squeeze into 21 seconds), there's still a 23.7% chance that the diseases won't be refreshed. It wouldn't be until 8 SS (~30 seconds) that you get down to a 10% chance of the diseases not refreshing. So, until we have 40 second diseases, the SS glyph is gtg.
I'm on the other side of the fence with your theory. As the length of diseases increases and gives you greater # chances to refresh it, then the value of the SS glyph increases. Until the % chance of the diseases not being refreshed is < 1-2%, then longer duration is better. With 5 SS (the number you can squeeze into 21 seconds), there's still a 23.7% chance that the diseases won't be refreshed. It wouldn't be until 8 SS (~30 seconds) that you get down to a 10% chance of the diseases not refreshing. So, until we have 40 second diseases, the SS glyph is gtg.
Why 1-2%? Why not 3%?
Point being that there isn't a threshold...it's an asymptote. Yes, the longer the diseases the last the greater the chance that SS Glyph will proc, but there's an even greater chance that the extra disease time will be wasted. If you were to graph the distribution of "SS's until refresh" you would find that although the average value was a few SS's, the single largest category was 1 SS. Why? Because of all the SS's you make, only 75% make it to the 2nd SS. Only 75% of those make it to the 3rd SS. Etc.
So if you extend your disease duration out, say, to 10 SS's, the chance of actually getting to the 10th one is 0.75 ^ 10, or about 4%. So 96% of the time you get no benefit from extending your diseases past 9 SS's.
There isn't a threshold at which it's devalued; it's a diminishing gains function. Thus, extending diseases from 21 to 24 seconds is in fact less useful than extending them from 18 to 21 (for example). So the question is, what is the actual dps value of those additional 3 seconds, and thus what is the actual value of that point in Epidemic?
EDIT: I'm only talking about in the context of having the SS glyph.
I just ran some terribly basic simulations for 1040 SS's per scenario.
The assumptions were a SS every 4 seconds so 5 SS's every 20 seconds. 25% chance on each SS to proc the length of diseases.
I then got the average uptime for each possible situation:
TOTAL UPTIME:
15 seconds: 68.96%
18 seconds: 75.86%
21 seconds: 80.47%
and if they ever add that third point...
24 seconds: 84.21%
Therefore, for:
1 point in Epidemic you gain 10% more uptime
2 points = 6.08%
and the hypothetical
3 points = 4.65%
So you are absolutely right that each point depreciates in value at least in this basic simulation but how that translates into DPS I'm still not sure.
One possible way to figure this out would be this:
TOTAL DAMAGE =UPTIME%*SSdamage + DOWNTIME*(IT+PSdamage)
That would mean that with an average of 5300dmg for SS and an average of (1511+2105)3616 PS+IT Damage you would get the following average per FU pair...
0 points = 4777.32 or (*5 FU pairs /20 seconds = 1194.33dps)
1 point = 4893.411 or (1223.353)
2 points = 4971.134 or (1242.784)
3 points = 5034.095 or (1258.524)
... how that translates into REAL dps I'm not sure yet. I'm tired now.
15 seconds: 68.96%
18 seconds: 75.86%
21 seconds: 80.47%
and if they ever add that third point...
24 seconds: 84.21%
By "uptime" do you mean "the % of FU (or DD) runes that are used for SS instead of PS/IT", as opposed to the more typical meaning "% of fight that buffs/debuffs are active".
On a related note, another factor that makes this hard to model is that you don't (I think) actually want to SS until you actually lose the diseases, especially if there's going to be a long gap until they can be reapplied. Maybe the SS buffs in 3.1 have changed things, though.
By "uptime" do you mean "the % of FU (or DD) runes that are used for SS instead of PS/IT", as opposed to the more typical meaning "% of fight that buffs/debuffs are active".
On a related note, another factor that makes this hard to model is that you don't (I think) actually want to SS until you actually lose the diseases, especially if there's going to be a long gap until they can be reapplied. Maybe the SS buffs in 3.1 have changed things, though.
Like I said, it's a very basic simulation. "Uptime" is in fact the amount of time the debuffs are up on the mob. It's not very realistic because you would never 'let' less than 99% uptime occur if you are playing properly. However, it does help show how often your FU pairs would need to be shifted from SS's to PS+IT. Another inaccuracy is the spacing of FU pairs being 4 seconds exactly.
Until I can actually model the SS glyph on a spreadsheet properly, I'm fairly confident that this is the best I can contribute to the discussion. I'll give it some more thought though.
Thats odd, my necrosis accounts for more than 1% damage per point.
As a matter of fact in some fights my necrosis is insanely high. In yesterdays Loatheb kill it was 10% of my total dps where my white dps was 20%. Probably due to all the buffs or something, but it surprised me.
If your necrosis does that little that doesn't seem totally right.
Also dont forget that BCB is getting a big nerf. Some people have calculated that its potential will drop by 25-50%. That means that even if BCB is currently 1% per point, it will only be 0.5% per point after the patch.
But as far as that talent build goes, i wouldn't get a spec without at least 5/5 in necrosis. ATM i put some filler points in BCB, but that is because i don't know how well Ghoul Frenzy and Imp UP are gonna be.
For just personal speed the latter seems quite expensive, and the former gives so little dps increase that i doubt its worth the unholy rune.
I totally forgot about the BCB nerf. As for necrosis, recount doesn't show it's damage per say (it lists necrosis as ~0.1% of my damage), so I calculated 0.69% per talent point by looking at the damage done by melee attacks. So either I made a mistake or you have something funky going on with yours (I'm more inclined to go with the former).
As far as I see it, unless they change the cost of Ghoul Frenzy to either RP or to 1U+1F (and up the damage from it), it will end up being a worthless talent.
So now we need to wait and see what the jury decides on IUP. If it turns out that IUP isn't worth the point, then the filler points in BcB does seem like the best option. Something like:
I just ran some terribly basic simulations for 1040 SS's per scenario ...
0 points = 4777.32 or (*5 FU pairs /20 seconds = 1194.33dps)
1 point = 4893.411 or (1223.353)
2 points = 4971.134 or (1242.784)
3 points = 5034.095 or (1258.524)
... how that translates into REAL dps I'm not sure yet. I'm tired now.
I'll caveat the math below with my feeling that Epidemic has more value than simply increasing the probability SS will proc, such as keeping diseases ticking for longer when you have to run out or there's an air phase (which is going to be magnified if you're the only 13% magic debuffer), giving you more globals to play around with, and simplifying your rotation.
That said, purely on the gained dps from the increased chance SS will proc, the best method I can think of to calculate expected dps gain is to look at the probability we have to refresh diseases each span of disease time. Multiply the probability you have to refresh at least once in that span by the damage you lose when you're forced to refresh, and divide by the span of time to get the expected dps loss. You can then do the same thing with one step higher of disease length, look at the difference, and the decrease in expected dps loss is what that talent point gains you. I'm using the same numbers as Methods for consistency's sake; 5300 average SS, 3616 average PS+IT (aka 1684 less damage), an SS every 4 seconds. I'm also ignoring RP for the moment.
Probability of SS failing to proc at least once in Y casts of SS = (1-.25)^Y. On average you'll SS 3.75 times in a 15 second disease window, 4.5 times in a 18 second disease window, and 5.25 times in a 21 second disease window.
15 seconds = 34.0% = 573 expected damage loss = 38 expected dps loss
18 seconds = 27.4% = 461 expected damage loss = 25 expected dps loss
21 seconds = 22.1% = 372 expected damage loss = 17 expected dps loss
With this math, Epidemic point 1 appears to give you 13 dps and point 2 appears to give you 8 dps.
I'll caveat the math below with my feeling that Epidemic has more value than simply increasing the probability SS will proc, such as keeping diseases ticking for longer when you have to run out or there's an air phase (which is going to be magnified if you're the only 13% magic debuffer), giving you more globals to play around with, and simplifying your rotation.
That said, purely on the gained dps from the increased chance SS will proc, the best method I can think of to calculate expected dps gain is to look at the probability we have to refresh diseases each span of disease time. Multiply the probability you have to refresh at least once in that span by the damage you lose when you're forced to refresh, and divide by the span of time to get the expected dps loss. You can then do the same thing with one step higher of disease length, look at the difference, and the decrease in expected dps loss is what that talent point gains you. I'm using the same numbers as Methods for consistency's sake; 5300 average SS, 3616 average PS+IT (aka 1684 less damage), an SS every 4 seconds. I'm also ignoring RP for the moment.
Probability of SS failing to proc at least once in Y casts of SS = (1-.25)^Y. On average you'll SS 3.75 times in a 15 second disease window, 4.5 times in a 18 second disease window, and 5.25 times in a 21 second disease window.
15 seconds = 34.0% = 573 expected damage loss = 38 expected dps loss
18 seconds = 27.4% = 461 expected damage loss = 25 expected dps loss
21 seconds = 22.1% = 372 expected damage loss = 17 expected dps loss
With this math, Epidemic point 1 appears to give you 13 dps and point 2 appears to give you 8 dps.
Im not sure about this, as a Scourge strike is a discreet event. You cant have fractional scourge strikes. Nor is the spacing evenly distributed. It is at relatively predictable intervals, but in clumps. To do a proper probability analysis you have to break down the rotation case by case, figure out how many FU and DD pairs will refresh before diseases run out, and then calculate the % of events that covers.
Exceedingly complex unfortunately. My guess is that the 2 points in epidemic is a rock solid 5 refresh chances, 1 point is more consistently 4 (to make 5 it would have to last till the same pair, which is impossible), and 0 points is pretty close to the 3.75 you listed (possibly higher, possibly lower). You will drop an refresh opportunity if the refresh occurs right before the blood runes are converted to death runes, and possibly with the UF pair just before that as well, not quite reaching the DD pair.
Im the good old fashion 17/0/54 spec so i only get 5 expertise from my spec. I have also have not been very lucky when it comes to expertise item drops so i only have [Ruthlessness] equipped at the moment. The other day i was tired of only having 8 expertise so i dropped 6 [Bold Scarlet Ruby] and replaced them with 6 [Precise Scarlet Ruby].
When i consume a expertise elixer i am at soft cap (26 expertise). My question is....is it better to sit at 20 expertise and close to 3600ap or is it better to have 8 expertise have over 3800ap and a lot more strength for my base stats?
Im the good old fashion 17/0/54 spec so i only get 5 expertise from my spec. I have also have not been very lucky when it comes to expertise item drops so i only have [Ruthlessness] equipped at the moment. The other day i was tired of only having 8 expertise so i dropped 6 [Bold Scarlet Ruby] and replaced them with 6 [Precise Scarlet Ruby].
When i consume a expertise elixer i am at soft cap (26 expertise). My question is....is it better to sit at 20 expertise and close to 3600ap or is it better to have 8 expertise have over 3800ap and a lot more strength for my base stats?
Modeling expertise's worth is still not completely agreed upon. There are many ways to look at it and even more ways to model it. In some of my builds I've seen expertise worth as much as 3.5 AP where other times I've seen it as low as 1 ap in equivalent points. The main source of variation has been addressing how to model the delay caused by a miss/dodge. I'm confident that expertise is far less valuable for 17/0/54 than say 51/0/20. We have room to miss as unholy and therefor shouldn't be worried about DPS loss from a dodge or two unless they happen in succession of each other. Maxing Exp is a bit too costly IMO and so you should just be aiming for the smallest amount of dodges without sacrificing anything else generally accepted as 'rock solid'.
What isn't hard to model and has been pretty consistent throughout all spreadsheet builds and specs has been Strength.
Strength, depending on gear level, is generally worth between 2.7 and 3 AP. As a rule of thumb for Unholy I would say that you should gem nothing but STR no matter what your gear is like. There are far too many factors in it's favor to say otherwise. This has also been discussed in great detail since the launch of WotLK.
Modeling expertise's worth is still not completely agreed upon. There are many ways to look at it and even more ways to model it. In some of my builds I've seen expertise worth as much as 3.5 AP where other times I've seen it as low as 1 ap in equivalent points. The main source of variation has been addressing how to model the delay caused by a miss/dodge. I'm confident that expertise is far less valuable for 17/0/54 than say 51/0/20. We have room to miss as unholy and therefor shouldn't be worried about DPS loss from a dodge or two unless they happen in succession of each other. Maxing Exp is a bit too costly IMO and so you should just be aiming for the smallest amount of dodges without sacrificing anything else generally accepted as 'rock solid'.
What isn't hard to model and has been pretty consistent throughout all spreadsheet builds and specs has been Strength.
Strength, depending on gear level, is generally worth between 2.7 and 3 AP. As a rule of thumb for Unholy I would say that you should gem nothing but STR no matter what your gear is like. There are far too many factors in it's favor to say otherwise. This has also been discussed in great detail since the launch of WotLK.
Thanks. I really needed a second opinion on this matter. I guess i just have to hope for some expertise drops.
I totally forgot about the BCB nerf. As for necrosis, recount doesn't show it's damage per say (it lists necrosis as ~0.1% of my damage), so I calculated 0.69% per talent point by looking at the damage done by melee attacks. So either I made a mistake or you have something funky going on with yours (I'm more inclined to go with the former).
Necrosis is a very funny ability because of the way it double dips from certain buffs.
Take for example a white hit for 1000.
Normal necrosis would be 200.
Raid buffs increase the white hit to 2000.
Necrosis increases to 400.
White hits benefits from Blood Presence and increases to 2300
Necrosis increases to 460.
Necrosis benefits from Ebon Plaguebringer (im not sure, but im inclined to say so)
Necrosis increases to 520.
Necrosis benefits from Blood Presence as well (double dip).
Necrosis increases to 600.
So a 2300 attack produces a 600 dmg necrosis which isnt 20% of its dmg but 26%
-------
That kind of weird stacking happens with multiple stuff.
For example if i take my last WMO report and check Loatheb fight (its the only one with some decent numbers, since i was DCing all the time.
Max white hit: 4715 dmg
Max Necrosis: 1288 dmg
Thats not 20% but 27.3%
In the same trend my white dps was 17.8% of total
My necrosis was 5% of total. (Usually its higher, but Loatheb is a bit odd due to the 50% increased crit)
Again thats 28% of my white dps as necrosis dps
So thats what im saying. Necrosis scales a bit differently with certain raid buffs.
I think this is right, and going further than Bladed Armor, out of Morbidity, Necrosis, BcB, and Dark Conviction, I think Dark Conviction is where you'd want to lose points. I ran some numbers on an all-25s boss parse I have (Sarth+3D, all of Naxx, Malygos) to get a broader picture and avoid any anti-Patchwerk comments, and it looks like I'm getting 1.04% per point of Necrosis, .96% per point of Morbidity, .71% per point of BcB, and .6% per point of Dark Conviction. Keep in mind the crit from that talent doesn't benefit gargoyle or ghoul, and D&D only gets a 1.5x multiplier, that's why it has the lower number.
Looking at only single-target fights, Morbidity and BcB may swap places, but Dark Conviction remains steady at around .6% per point. We'll have to see how badly the nerf hits BcB; it'll have to be more than a 10%-15% nerf for it to drop below Dark Conviction.
Taking this further, post 3.1 do you guys think it would be worth it to drop points from Dark Conviction to fill some into Desecration, even in a SS spam build? Or do we still feel that the actual uptime of Desecration will be so low that the crit from Dark Conviction would still be a stronger investment?
Taking this further, post 3.1 do you guys think it would be worth it to drop points from Dark Conviction to fill some into Desecration, even in a SS spam build? Or do we still feel that the actual uptime of Desecration will be so low that the crit from Dark Conviction would still be a stronger investment?
Unless PS is so heavily buffed that we turn completely away from SS builds and don't even use the SS glyph, which I highly doubt given that SS is also getting buffed, Desecration will continue to be terrible. I generally have to PS 4 times in a 3m fight. That's 48/180 = 26.67% uptime, or a 1.33% dps boost for 5 points, or .26% dps boost per point. It's awful.
That definitely makes sense Stoical- along the same lines- I'm pretty sure I will use a movement buff of some sort, ether from boot enchant, or Imp UP if I can find spare points for it- have you guys been able to model just how much worth per point Imp-UP would be strictly from a 15% speed boost standpoint? i.e. would it be worth it to take points, say from Dark Conviction (since it's my whipping boy for moving points around from for this scenario at least), to put in the 15% movement buff, to gain a real offensive enchant for boots? Or would it be wiser to take the slower boot enchant (7% less movement and pretty crappy offensive stat) for an offensive boost through other talents?
I guess I'd like to see a percentage worth for 15% movement, so I could compare it to the other options available, especially since it's not raid wide now (there was no contest before imo).
Has your runic power intake been lowered so much in 3.1 that you can't get a 2nd DC off in your first rotation.
My rotation at the moment is
PS-IT-BS-BS-SS-DC-UB
SS-SS-SS-DC-DC-REPEAT(if diseases are about to fall off)
That depends if your specced into Dirge, but normally after my full rotation I can do a second DC sometimes. Also I haven't been able to test the new DC or UB glyphs which might enable me to add more into my rotations.
That depends if your specced into Dirge, but normally after my full rotation I can do a second DC sometimes. Also I haven't been able to test the new DC or UB glyphs which might enable me to add more into my rotations.
I'm not the greatest mathematician, so feel free to correct, but I worked on the value of glyphs; trying to decide on what would be the best third glyph (SS and ghoul as 1 and 2).
assume:
1 minute interval
3 20sec UB=2 30sec UB (not true, would actually be 63 ticks vs 62 ticks)
a easily achieved 4-SS rotation as seen below:
IT-PS-SS-BS-BS
SS-SS-SS
SS-IT-PS-BS-BS
SS-SS-SS
SS-IT-PS-BS-BS
SS-SS-SS
Total RP generation:
60 seconds in combat=12 RP from talents
3xIT=60 RP (w/ IT glyph)
3xPS=30 RP
6xBS=60 RP
12xSS=360 RP (w/4t7); 240 (w/o4t7)
total= 522 RP (w/4t7); 402 (w/o4t7)
UB glyph:
2 UB to cover the 60 seconds, using 80 RP. You save 40 RP and are able to cast 11.05 DCs (w/) or 8.05 (w/o) at 40RP each.
In other words, UB glyph saves you 40 RP per minute; this is what it will always save you.
DC glyph:
3 UB to cover the 60 seconds, using 120 RP. You are able to cast 12.5 DCs (w/4t7) or 8.81 (w/o4t7) at 32 RP each.
In other words, DC glyph saves you 100 RP (w/4t7) or 70 RP (w/o4t7) per minute. The more RP you generate, the better it gets.
Conclusion:
IF GCDs are not a problem, DC glyph would be better damage. Note, GCDs are generally not a problem for 2h unholy since they use so many 2-rune abilities.
On an interesting note, you see that the more RP you generate, the farther ahead DC glyph pulls. If you take a second to visualize it graphically, UB represents a fixed cost and DC cost represents the slope. You end up with two lines, y=80+40x and y=120+32x; they intersect at 280. Meaning if you generate more than 280 RP a minute, DC glyph is better. If you can't generate 280 RP over a minute...well you probably don't need to be worrying about min-maxing anyways (technically, an entire minute of SS spam, without dirge and 4t7 would only generate 270).
As for IT glyph, its worth 10 RP per cast. In the rotation I listed, it only works out to +30 RP, making it inferior to UB glyph. However, it depends on how often you are forced out of melee and casting heavier IT rotations.
One caveat about your analysis Ashelianna: you didn't include the SS glyph in your rotation. Realistically, SS and Ghoul glyphs are not negotiable for 2H unholy, meaning (a) IT glyph has very low value as you only cast IT when SS glyph goes through a string of non-procs, and (b) you're really only looking for one glyph out of the new ones. That said, your analysis does show pretty clearly that DC > UB and really highlights the value of DC glyph (which frost DKs already knew well with their FS glyph).
I figured everyone realized SS and Ghoul were mandatory, the point is deciding DC/UB/IT for the third. I'll edit it in though since it may not be clear.
02/26/09, 4:23 PM
Similar Threads
