Go to Format->Sheet->Show, Show the appropriate sheet. Then you can either
-Insert a new row(s) between the first and last item, it shouldn't break anything, or
-Enter your items on new rows on the bottom, then go to Insert->Names->Define, find the relevant names (handlist, ohlist), and expand the ranges to include the new rows you created items on.

This spreadsheet seems pretty sweet, but I don't understand how to input my cycles, and furthermore I want to try to create the best cycle fo rmy gear/talents.

Would that just be done by inputing different cycles and observing which come out on top in terms of dps?

Thanks.

The answer to your question is yes. There is a guide to using the sheet linked from the first post, it's hard to miss.

This spreadsheet allows you to input custom values for X and Y in the Xe/Yr (CttC) cycles. How good is the modelling for stupidly small values of X and Y?

I ask because I've recently been noticing that as crit rates rise, and once I hit the expertise cap, my cycles have been becoming more degenerate. The high crit rate for Mutilate and the high proc rate of Ruthlessness means that a majority of my finishers only require a single Mutilate to reach 4+ combo points. With this in mind, I examined the impact of completely ignoring combo points and shifting to the simplest possible "rotation" - i.e. a strict alternation of Rupture and Envenom as finishers, with a single Mutilate between each. In the spreadsheet, this is achieved by setting X and Y both to 2. I was shocked to find that despite completely gutting everything even remotely difficult about Mutilate play, it only dropped my estimated DPS from 5200 to 5066. That's a drop of slightly over 2.5%, well within the day-to-day margin of error even for a high end rogue. Conversely, setting both X and Y to 5 only dropped the predicted DPS to 5175, a drop of less than 0.5% DPS. This (very nearly) corresponds to the equally degenerate rotation of Mut/Mut/Envenom/Mut/Mut/Rupture, which again completely ignores combo points.

If these result proves to hold true, then I'm extremely discouraged. It seems like a return to BT days when you could achieve 95% of your DPS output by just keeping S'n'D up. Moreover, with HfB becoming a 1-minute buff in 3.1, with a lower energy cost, the cycle becomes even more degenerate since it requires refreshing less often and has less of a cycle impact to do so. Once the wind-up period is past, you could actually achieve >=97% of your total Mutilate damage output with a macro.

Aldriana has stated in various threads (including his magnificent thread on the official forums) that optimal Mutilate play is extremely challenging, requiring you to optimise Rupture uptime, watch DP ticks and watch for procs from trinkets / enchants etc. I don't doubt that these effects exist, but I think they're unlikely to be significant, especially given the above finding. Certainly despite repeated testing I have never found any benefit from using a DP tick timer, whether in real play or against a dummy.

Simply put: I think the rogue community at large has been deceiving itself about the complexity of the Mutilate playstyle, and has been making heroic efforts for minuscule gains. I am no longer sure that Mutilate rewards good play to any meaningful degree. Almost the only "skill" involved is that of not letting HfB or S'n'D drop, and the latter is much more easily achieved by using one of the two mindless "rotations" outlined above.

I can't comment specifically on the validity of the model for small Envenom sizes, but if I had to guess I'd say it's likely not that good. However, this brings up another issue that I've been thinking about lately (and may start a thread about soon) - I think our existing modeling - both here and in my spreadsheet - is somewhat flawed in that it looks primarily if not exclusively and what moves are being done, but pays very little attention to when they're being done. And given that our class is somewhat unique in that we have a lot of control of when moves are performed (rather than chaincasting stuff on cooldown like most classes), I think it's worth looking at - particularly since there's reason to believe that it matters.

How much it matters, of course, is a very good question - and one that, with current theorycraft, we haven't answered. But I think there's reason to believe that proper attack timing can gain up to a couple hundred DPS over just doing stuff as soon as you have energy (or, alternately, queuing everything without consideration as to why). So I think the next major modeling challenge is going to be to work out the importance and prioritization of the various factors than influence when you want to do moves.

Would it be accurate to to compare the damage gained from envenoming+envenom buff inside of a mirror or other attack power proc with the damage lost from the lost slice time?
How complex a job is it to calculate the value of that lost slice and dice time? Would this translate into a minimum energy queued or minimum time spent queueing value that you could apply to your decision making?
I suspect it is not nearly so simple but I look forward to further discussion on this topic.

You're asking the following question: "Is it ever worth it to time things for a slightly more powerful damaging finisher at the expense of a few seconds of lost S'n'D time?". By "damaging finisher" I mean Eviscerate, Envenom or Rupture - something that directly damages the target and benefits from AP procs etc.

That would be meaningful for Combat but not for Mutilate. As Mutilate, you cannot afford to let S'n'D drop, ever. The reason for that is that Mutilate rogues don't use S'n'D in their rotation. For a Mutilate rogue there is no such thing as an S'n'D drop, considered in isolation. If your S'n'D drops, you not only suffer the penalty the lost S'n'D time, you also lose a complete Envenom - because you have to restart S'n'D instead of just refreshing it with an Envenom.

Given that, we can rephrase the question as follows: "Is it ever worth it to time things for a slightly more powerful damaging finisher at the expense of a few seconds of lost S'n'D time and the complete loss of a damaging finisher?

The answer then is self-evidently No.

Maybe I am not reading his question correctly but I think he may be referring to the lost SnD time from refreshing SnD via envenom right when a trinket procs rather then waiting till the last few seconds of SnD or pooling. The reason I am speculating this is because while you have SnD up, there is no decision to get a envenom in at the expense of SnD uptime (with a mutilate spec).

If this is the case, first you don't really lose dps from the lost seconds of your SnD, you do lose potential dps that you would have gained from pooling before an envenom (such that you land a mutilate or two immediately after while the envenom buff is up) and from your DP stack increasing futher if it wasn't 5 already (probably less of a problem in 3.1 with a fast weapon and Master Poisoner but you would still want to wait till right after a DP tick). Second you would have to consider that if you continued to pool as normal before an envenom you DP stack will tick at a higher value as it is refreshed during the AP proc unlike with Rupture where it will tick for the same value throughout it duration regardless of changes to your AP in its duration.

For combat, as Songster has already explained it's just not worth it, the choice between maintaining 40% haste to 45-50% of your damage (not considering Combat Potency and Wound Poison procs) and small increase to a Rupture ~7% of your damage or Evis ~4% should be quite clear. Keep in mind your white attacks gain from AP procs as well.

Sorry if this is not what you were originally asking.

Or perhaps - what do I gain in dps by assiduously watching buff procs and trying to do envenoms more in line with these than with maximum pooling, versus the cost of less pooling and adding one(?) more thing to track, which will generally have a negative impact on your situational awareness etc.

It's pretty clear that any loss of slice uptime will be a net loss. Another thing to consider: is there a set of trinket/enchant/buff procs for which it is better to hit another envenom than to re-apply rupture?

I would really like to know why you have reason to believe this, because my feeling is the exact opposite. Let's make the point in a brutal manner. Yes, we can't model it accurately, but we can get a reasonable estimate with the current tools we have and some napkin math. The three most powerful AP procs we have are the Mirror of Truth, and MH/OH berserking. Let's take a look.

The Mirror of Truth is the most powerful, at 1000 AP.

Taking Vulajin's current sheet, with everything on default, we note the following:
Buffed AP: 6158
Mutilate DPS: 890.36
Rupture DPS: 475.80
Envenom DPS: 399.34

I'm only looking at the yellow damage skills since these are the only ones we can control the timing of. Now, the Mirror has around a 20% uptime (10s proc on 45s ICD), so in the default state with no special timing, it adds around 200 AP. We can pool energy for a maximum of 5-6 seconds waiting for a proc to happen. Similarly, we may be able to advance our moves by a couple of seconds at the other end of the proc and squeeze them in before it fades. How to account for this? Well, the question is how much extra energy we can "spend" within a single proc duration, either by pooling before the proc, or by squeezing skills in before it fades. Now, the maximum extra energy we can ever have is one full bars' worth. For the sake of argument, let's call that 10 seconds' worth of energy - in practice it's less because you also get energy from sources other than your natural regen.

That means the maximum benefit we can get from watching the proc times is to pool enough energy to get in 20 seconds' worth of energy-based attacks within each 10-second proc time: i.e. a doubling of the effective uptime of the trinket for our yellow skills.

Modelling this by adding 200 AP in the sheet shows the following:
Buffed AP: 6358
Mutilate DPS: 908.01
Rupture DPS: 483.76
Envenom DPS: 408.41

Total gain is 34.68 DPS


Now let's look at Beserking. To make a similar guesstimate, we have to work out how much we can increase the effective uptime of the Berserking buff by pooling and timing skills. This is more complicated since Berserking can self-refresh, and so I don't have a good handle on the expected proc uptime, and (more importantly) the length of each uptime session. As a very preliminary guess, since the proc itself is 15 seconds, let's assume that the average uptime session is 20 seconds' long. Once again, the maximum we can pool is a full energy bar. So in this case the best we can do is to cram in 30 seconds' worth of attacks within a 20 second uptime period: a 50% increase in effective uptime for our yellow attacks.

Once again looking in Vulajin's sheet, we see:

With Beserking on both weapons:
Mutilate DPS: 905.80
Rupture DPS: 482.77
Envenom DPS: 407.27


With both weapons unenchanted:
Mutilate DPS 880.65
Rupture DPS 471.42
Envenom DPS 394.35

This means that the total yellow DPS effect of Berserking on both weapons combined is 49.42 DPS.

Since we guesstimated above that we might be able to gain 50% effective uptime, this means that with extreme pooling and timing our yellow attacks for Berserking procs on both weapons, we might gain 49.42 / 2 = 24.71 DPS.

So, if we optimise for all of the procs from the Mirror, MH Beserking and OH Berserking, we might gain a total of 24.71+34.68 = 59.39 DPS.

But of course we can't optimise for them all. If you wait for a Mirror proc, your Berserking may fall off, or vice versa. So what we have here is an extreme overestimate of the effects of timing our skills. I've made very charitable assumptions for how much we can optimise Mirror procs and Berserking procs, and an even more charitable assumption about the combined effect of optimising them all - and yet it still comes out under 60 DPS benefit.


From your own sheet, the effect of timing Envenoms for DP ticks is less than 50 DPS. I hope I need not point out to you that timing Envenoms for DP ticks pretty much prevent you from timing them for AP procs. However, being charitable, let us add this to the 60 DPS overestimate. The total impact of watching DP ticks, Mirror procs, MH berserking procs and OH berserking procs is no more than ~100 DPS. And that is assuming you are able to pool energy for 10 seconds before every Mirror proc, every MH Berserking proc and every OH berserking proc, while still timing every single Envenom to within half a second of a DP tick

Even if I'm wrong, and your "a couple hundred DPS" is nearer the mark, I must point out that that is at best a few percent when set against the backdrop of a baseline ~5000 DPS output. That's not at all incompatible with my statement that you could maintain the vast majority of your DPS with a trivial macro.

I don' think it's that surprising that attack timing based on trinkets/enchants is only in the range of maybe ~100 DPS at the most. FWIW you missed the DMC proc which would add a little bit to that number. But regardless, that sort of optimization is fairly marginal and it obviously would have to be, because otherwise the trinket would be too strong.

I think the more interesting question to ask is out of the following cycles, can we empirically prove the DPS difference assuming we don't optimize for the cases above:
a.) Standard cycle where if we hit 4 CP we do a finisher (which is determined by the current state), else mutilate.
b.) Mutilate, Envenom, Mutilate, Rupture. Short cycles guaranteed, but a ~45% chance you will do a finisher at 3 CP, and the negatives that has associated with it. Also one thing I'd wonder about is what do you do in the case where it won't let you overwrite your Rupture?
c.) Mutilate, Mutilate, Envenom, Mutilate, Mutilate, Rupture. Long cycle guaranteed, and I would be concerned about whether or not you would occasionally actually drop SnD.

This is something that cannot be demonstrated completely accurately with the current spreadsheets to my knowledge. It's not something you can napkin math because there's simply too many factors (energy regen, poison procs, and so on). So I think first the baseline DPS difference should be proven before we start optimizing on the next levels. Because it's not like the person doing cycle B or cycle C couldn't also optimize it to take into account trinkets.

When it's all said and done though, I wouldn't be that surprised if the difference between case A and options B/C is less than 10%. It's intended that unskilled players can still do damage.

Of these, I think the only one that's likely to matter in practice is the Envenom buff. As you say, timing for procs is a crapshoot - all I was doing is proving that even if you could reliably time for it, it adds very little. Aldriana's modelled the effect of timing Envenoms for DP ticks, and the total effect is also minimal - less than 1%. My intent is to challenge the preconception that "Mutilate is hard to play", because I just don't think it is. You can make it hard to play by trying to let too many factors influence your choice of ability timing, but the gains you derive from that are not in practice sufficient to show any real difference between a good rogue and a bad one. Yes, there is a difference between a bad Mutilate rogue and a terribad one, in terms of letting HfB and S'n'D drop, but once you get past that hurdle, the spec in fact becomes easier to play than Combat.

Why am I doing this? Well, I think one of the reasons Mutilate doesn't have many DPS cooldowns to use is that the base spec is itself supposed to be more challenging, such that adding in extra buttons to push is something the developers may be avoiding. I believe this to be misguided.

Moreover, while the impact of HfB is too high, I am in fact in favour of keeping it (or some other button) as a regular part of the rotation. Moving it to a 1 minute buff is in my view exactly the wrong direction to take the spec. We already have too few buttons to care about, given the lack of cooldowns: if we only hit HfB a couple of times per fight, we become a three-button spec, with nothing but Mutilate, Rupture and Envenom. I dislike that intensely.

Finally, as I said in the first post, I'm in fact beginning to worry that with capped expertise and a high crit rate, we lose the RNG element of our combo point generation, such that shifting to a completely fixed cycle is in fact very little change in DPS. If this is true, then we can take the three-button spec and further simplify it to a single macro, which is frankly abhorrent.

In regard to the latter concern, I would like to try and work out the DPS impact of the following:

a) "Normal" 4+r/4+n/(4+n) cycling - i.e. always use 4+ finishers, and Envenom as many times as possible between ruptures.

b) "Tight" degenerate cycles: Mutilate / Rupture / Mutilate / Envenom, paying no attention to combo points. This cycle never drops S'n'D and never clips Envenom buffs, but may have an issue overwriting Ruptures and wasting energy on low-point finishers.

c) "Middle" degenerate cycles: Mutilatex2 / Rupture / Mutilate / Envenom or Mutilate / Rupture / Mutilatex2 / Envenom, once again paying no attention to combo points. This cycle will sometimes waste a combo point before the Rupture, and sometimes waste energy on a low-point Envenom. It is very unlikely to drop S'n'D, but may have an issue overwriting Ruptures.

d) "Extended" degenerate cycles: Mutilatex2 / Rupture / Mutilatex2 / Envenom, paying no attention to combo points. This cycle will never waste energy on low poin finishers, but will waste combo points from good SF/Ruthlessness procs. It also runs a risk of dropping S'n'D.


Based on my preliminary investigation with Vulajin's sheet, unless the modelling fails catastrophically for tight cycles, I suspect all of (a) to (c) fall within 3-5% of each other. (d) is almost certainly out due to wasting too many combo points and dropping S'n'D.

The major problem, as ever with rogues, is that our white damage and poison damage are collectively high enough that you can screw up the yellow portion of your damage almost totally and yet have little effect on your overall output. There's no real way round this except substantially nerfing white damage + poisons and substantially buffing yellow damage. However, this would be unacceptable in PvP because of the massive yellow damage burst that results when you empty the starting 100 energy into your target.

As I said, it's already a known that the current setup in the spreadsheet does NOT model even case A extremely accurately. When you set r or n to 4, all it's saying it's 4+r/4+n, and as far as I know it's NOT modelling the additional Envenom that you can tack on to the end. And we also know that cases B-D can't really be modeled particularly well either in the sheet. So yes, all your testing in the sheet is invalid, because even Vulajin will admit it's not entirely all there yet (but something he seems to be working on for 0.5).

Also, it's not really true that for Mutilate yellow damage is meaningless. It's usually in the range of 35-40% of your overall damage, and keep in mind that proper Envenom buff placement will have some effect on the percent of damage from poisons.

Again though, you probably won't see more than a 10% difference between the best and the worst of a-c, but I'd suspect it's higher than 3%.

...snipped the rest...

Would it be of interest to test/model a 5th cycle as well? A "degenerate macro" cycle? Since there is (in my opinion, a valid) concern that the yellow damage output could be reduced to a macro, it seems like it might be worthwhile to model a 2-button cycle - Mutilate and Envenom/Rupture with a reset timer (16 or 20 seconds, depending on glyphed or not).

In other words (first Rupture included for starting the cycle after SnD):

e) Mutilate / Rupture / Mutilate / if over 16/20 seconds since Rupture, then Rupture, else Envenom

This is more simple than a state-based "is Rupture up" cycle and is likely relatively easy to macro. Even if *I* keep forgetting that castsequence reset doesn't actually work the way it does in my head.

If it's not of interest, that's fine, but it seems like it might produce dps approaching the "normal" rotation, while avoiding some of the pitfalls of the degenerate rotations. It also might more clearly highlight the concerns about the ease of Mutilate play.

Edit: Reduced the one-line sentences a little.

"Extended" cycle is very hard to maintain without Improved SnD (250 energy, 25 seconds in the really worst case scenario, 21 seconds on average) and has very poor uptime for Rupture and Envenom. So this cycle is rather bad idea.

And even assuming no Rupture glyph, "tight" cycle will take 130-180 energy (= ~10.8-18 seconds) and will give you 10-14s of Rupture. So it will for sure fail on AP buffs when it comes to refresh Rupture. So it can't be one-hit /castsequence macro, but if merged with "normal" cycling (Envenoming as many times as possible between Ruptures), it can be pretty powerful, yes. Although spreadsheet doesn't seem to model this at the moment.

But I'd suggest another cycle to your list: 1R/1E/1E instead of your "middle" cycle. It will most definitively need at least another 70 energy (= +~6s), so Ruptures won't overwrite. And with a little play on pooling energy, Envenoms won't overwrite either and will have quite a nice uptime. And it is degenerated and macro-spammable.

Bolded for emphasis. What that means is that in order to see a 10% drop in total DPS from scenarios (a) to (b) or (c), you actually need to lose 25-33% of your yellow damage. I think that's unlikely. Remember you're still using all your energy (not capping out) and keeping a similar (though not identical) balance of builders / finishers. Moreover, we're not actually looking at all yellow damage here. The bulk of the variation from (a) to (c) will be down to finisher damage. Remember that in (b) and (c) at least you are guaranteed never to clip your Envenom procs. We're also not looking at the impact of dropping HfB and S'n'D, since those shouldn't be an issue for any rotation except (d).

Let's lay out the effects of moving to scenario (a).

1) The damage from each individual Mutilate will be unchanged
2) The damage done by each Rupture / Envenom will go down since you're using lower combo point finishers. This is indeed a damage loss.
3) You will be doing proportionally more finishers and proportionally fewer Mutilates. Since even a 3-point finisher is better DPE than a builder, this is actually an increase in DPS, mitigating factor (2).


Now, some napkin math on factor (2), since that's where the damage drop comes in. With the standard 4+ rotation, most finishers are 4-point finishers, with a small proportion of 5-point finishers. Being generous, let's say that breaks out around 50:50 so the average is a 4.5-point finisher. With a Mut/Rup/Mut/Env rotation, most finishers are still 4 point finishers. ~45% of finishers come out at 3 points, and very rarely (~5%) you'll get a 2-point finisher. The average is approximately a 3.5-point finisher.

That means the net effect of scenario (a) is to shift from 4.5-point finishers to 3.5-point finishers. To a first approximation, that lowers the damage of those finishers by a factor of 3.5 / 4.5 = 0.7778

Rupture and Envenom are jointly responsible for about 20% of your total damage, so the effect of going to scenario (a) is about a 4.5% drop in overall damage. Set against this, we have factor (3). Thus I would expect the total damage lost from moving to a strict Mutilate/Rupture/Mutilate/Envenom rota to be less than 4.5%.

Approaching it from the other end: what would it take to actually see a 10% drop in our damage, as you suggest? Well, we would have to drop our Rupture / Envenom damage by more than half - i.e. be using ~2.25 point finishers. With Ruthlessness and Seal Fate, that's not actually possible. This is what I mean when I say that high crit rates (virtually ensuring 3 combo points from a Mutilate) and expertise capping (eliminating dodges) almost completely remove the RNG aspect of the Mutilate rotation.

I know I'm oversimplifying, which is why I said it was an estimate. That's also why I tried to give a rough bound on how badly you'd have to screw over your cycle to get to the 10% difference you initially suggested. :-)

Yes, there are things I left out - slightly bad form to only quote the ones that favour your point of view though! For example, another thing I didn't include is the increase in DP damage from doing smaller Envenoms and thus not clearing the stack. As far as the accurate solving by spreadsheets goes, I suspect that's only because everyone has looked at Ruthlessness and Seal Fate and assumed that a variable cycle is the only option. A fixed cycle is actually considerably easier to model.

I'm going to start a new thread with this discussion, since we've kind of gotten away from talking about the spreadsheet in particular (other than it would be nice if the next version has more comprehensive cycle modelling!)

I'm in the process of working with SimulationCraft to try to get some gauge on the numbers, the results thus far are quite interesting.

Unrelated to the above, I found time to do some additional testing. All testing on the Heroic dummy in Exodar.

As I had posted in this thread earlier, I had discovered that the spell crit reduction aginst bosses was less than 3%. New testing I have gotten this down further. Relevant data listed below.

New data, got a crit at 101 crit rating...weeding out mostly irrelevant data

101 crit rating (2.2001%) - 1 crits in 1060 successful wound poison applications
99 crit rating (2.1566%) - 0 crits in 840 successful wound poison applications

I'm now thinking either 2% or 2.1% (being divisible by 3 which is the level difference)

-----------

Additionally, I've been parsing additional information out of my various tests. Last I knew, the resisted damage loss for spells against a boss was estimated at 6%. I compiled the resists on my various wound poison tests and have the following.

Edit: Nevermind looking at the data more closely, it appears resists are now in 1/10th increments, not 1/8ths as I was led to believe.


Against the Grandmaster dummy, had about 600 successful wound poison applications with no resists at all.

New totals
8080 - Successful poison applications
4415 - Full damage
2513 - 9/10 damage
1152 - 4/5 damage

Combined, I'm seeing a now combined 5.9616% damage loss from resists.

In my previous testing of PPM I did have Wound Poison applied to both weapons for a duration of 2x60 minutes.
I didn't come around to include poisons in my parser, but if you already have one, feel free to run it through.

That should be around 3400 poison procs with ~6800 landed attacks.
Combatlog

I am not sure how to model this, so I guess I will just ask. I have been using the spreadsheet. And with my current gear I have found that my best rotation is 1s/5r. The issue is when I am raiding occasionally I will have a "more powerful" rupture so I cannot overwrite it and my energy is about to cap out.

So I am not sure if I should SS, or if I should change my whole rotation because it happens kinda often (not the energy capping part but the inability to overwrite at the end of rupture)

THanks

Out of interest I tested on the PTR what the partial resist rate on 2 Shadow Priest spells were on the Heroic Training Dummy.

I had 3063 Mind Flay ticks and 1020 Shadow Word: Pain ticks.

Being sure to adjust the partial resist amount on a tick that crit by multiplying by 2.09 (I had CSD meta) I came out with the following:
Mind Flay had a 5.9% Partial Resist reduction from 3063 ticks.
Shadow Word: Pain had a 5.2% Partial Resist reduction from 1020 ticks.

Note: If saw I had an entry with 2202 in dmg done, 104 in resisted and a 1 in the crit field then it's partial resist % would be:
(104*2.09)/(2202 + 104*2.09) = 0.0898 == 9.0%.

The various partial resist bands I saw were 0.0%, 9.0%, 10.0%, 18.1% and 20.0%

For Shadow Word: Pain (1020 ticks) the bands broke down like this:
Partial Resist Band% Number Percentage
0.0% 593 58%
9.0% 135 13%
10.0% 169 17%
18.1% 53 5%
20.0% 70 7%
===============================
Total: 1020 100%

For Mind Flay's 3063 ticks:
Partial Resist Band% Number Percentage
0.0% 1636 53%
9.0% 426 14%
10.0% 537 18%
18.1% 199 6%
20.0% 265 9%
===============================
Total: 3063 100%


I definitely wasn't seeing a 7% or so partial resist rate.
I was curious about the disparity between Mind Flay's and SW:P's resist rates. I'd have thought 1020 ticks would be enough to shake things out a bit.