Combat/Sub Spreadsheet for 4.0.1

Lightshadow · 09/30/10, 12:28 AM

I’ve created a very basic spreadsheet to model combat and DPS after the talent changes. It’s not intended to give an accurate overall DPS number (I’ve omitted bandit’s guile entirely, for instance), but rather to give at least a little guidance on stats, glyphs, and talents. It also doesn’t have the shiny front end you’ve come to expect from the other spreadsheets/sims – the shelf life for this is pretty limited.

DropBox Link: http://dl.dropbox.com/u/11257250/4.0.1RogueV0.12.xls

Note: Version 0.12 contains significant upgrades to both sheets

(Now with sp00n's UI changes!)

Usage: The character stats tab is the only directly useful one, and the boxes with the green backgrounds are the stats you should fill in from your unbuffed character sheet. The sheet gives you a projected DPS number for your gear for each spec you select. (Note that your character sheet agility will change significantly as you swap specs.)

Important Note: The sheet will allow you to say "Use RS when at 5CP but not when at 4CP", but this is never optimal. Despite the fact that the sheet will let you select those options, the result is nonsense because you'd never want to do it.

Current (10/14) Quick Guide to Combat in 4.0.1:

Typically agility is the best overall stat, then spell hit > expertise > haste > crit > white hit > Mastery. Mastery's worth less than 40% of Agility even in the most favorable gearsets. With hit being a blue socket it’s almost always worth matching socket colors (especially for agi socket bonuses). The sheet likely slightly undervalues spell hit and expertise, but you almost surely want to cap them anyway.
With the loss of poison talents in 4.0.1 wound poison/deadly poison and instant/deadly are much closer than before. In particular, for very low levels of mastery (say 0) or in fights with significant target switching wound poison/deadly may be higher DPS in addition to the debuff.
Double Mongoose is clearly the way to go for enchants. An offhand 1.4 speed dagger seems preferable to a 1.5 speed axe/fist/sword/mace, but not by enough to make it worth dropping down a tier.
I’ve assumed that the glyph of Sinister Strike is worthwhile, it’s hardcoded into the spreadsheet. From some extra messing around this appears to be a reasonable assumption.
With the change to adrenaline rush (haste + energy regen) and the short cooldown (~90 seconds), the glyph of adrenaline rush is solidly the second best prime glyph.
Rupture is much better than eviscerate, even without the glyph.
The rupture glyph is typically a bit better than the RS glyph.
Without the RS glyph it's typically a DPS gain to use RS at 4cp and not if you're at 5cp. For some gearsets it's not worth using RS with eviscerate at all, however. It never appears to be worthwhile to use RS if you're already at 5cp, even with the glyph.
3/3 relentless strikes and 2/3 lethality is clearly the best option for the 5 non-combat points.
The t10 2p is worth about 65 agility, while the 4p is only worth about 50.

Current (10/14) Quick Guide to Sub

Agility >> Expertise > Spell Hit = Haste > Crit >> White Hit > Mastery > AP. (See EP values at the end). Yes, mastery is that bad. (The sheet underestimates the value of spell hit a little bit, but 264+ gear will surely spell cap anyway).
The highest DPS build appears to be this with the point in enveloping shadows subject to the user's discretion. There's an argument to be made for 1/2 Lethality as a DPS gain, but given premeditation it's worth maybe 1-2cp per minute, which is a < 0.1% gain.
If you really wanted either enveloping shadows or cheat death, the worst talents per point are elusiveness (0.3% per point but other utility), the 3rd point of imp ambush (0.48%), and preparation (0.7% dps). If you happen to have so much hit you can't get down to 342 (spell hit cap with 2 precision) you can drop a point of precision for coup de gras.
The Prime glyphs appear to be backstab, shadow dance, and hemo. Hemo isn't worth using without the glyph unless you need the debuff, but I can't find a build that does better than using the glyph and thus using hemo.
Garrote can be a (very) minor DPS upgrade after vanish or as the first action of a shadow dance (i.e. when FW isn't already up). At any other time it's a DPS loss.
Do use a 5pt rupture before MoS wears off -- it keeps the 10% buff as long as you keep it rolling.
Do keep recuperate up as much as possible. Energetic recovery is a huge DPS gain.
Your weapon enchants should be double mongoose, no question. The MH absolutely needs to be a slow dagger, the OH should just be as fast as possible. (That's probably a dagger, but a 1.5 speed 277 axe is probably better than a 1.4 speed 264 dagger).
Instant/deadly looks to be ever so slightly better than wound/deadly, but only if you're on one target 100% of the time. If deadly is allowed to fall off (HLK, Sindy, and PP all come to mind) you should probably be using wound on the MH.
The subtlety "priority" list is more or less Keep SnD up > Evis to keep rupture up > keep recuperate up > eviscerate to burn CPs. Vanish and shadow dance should be used on cooldown (pooling to 80 before shadow dance) but always at least 10s apart to get as much find weakness uptime as possible. Shadowstep should be used before vanish but after the first ambush during shadow dance. 4pt eviscerates are never a good idea unless rupture will fall off before you can get 5. <5pt SnD and Recuperates aren't ideal but are better than letting them drop off.
The t10 2p is worth ~80 agility and the 4p is worth ~40 agility. The 4p might be worth somewhat more if you're having issues keeping the full rotation going. It won't fix the worst case scenario but it will help the average case.

Changes for V.0.11

Both: Added More trinket procs to the front page
Both: Updated the weapons to calculate damage off DPS & Speed -- same outcome, but easier to switch between weapons.
Combat: Fixed a bug In the t10 4p calculation
Sub: Added a new option to choose how often HAT procs.
Sub: Fixed a minor bug in the hemo bleed debuff
Sub: The sheet now uses hemo only if you're using the glyph.

Known issues in v.0.12

The initiative talent isn't doing anything
Deathbringer's Will (both versions) is minorly bugged. See this post for a quick fix.

Feist-Mok · 09/30/10, 1:29 AM

What offhand attacks are normalized that dagger normalization would have any impact? Main Gauche procs? Has there actually been any testing of this? Historically, as far as I know, sword spec procs have not been normalized, as they've been considered additional auto attacks, so that would be a significant change.

Lightshadow · 09/30/10, 9:40 AM

Some MG testing is on my to do list, just as soon as I can get a character on the PTR that doesn't have the mastery bug. (both normalization and the ability of special attacks to proc MG). That said, without loads of mastery it just can't be the case that MG procs alone are driving what I'm seeing in the spreadsheet, so I think I need to take another look at that.

koaschten · 09/30/10, 9:52 AM

After about a thousand sinister strikes on a dummy I am pretty confident that Sinister Strike is not proccing MainGauche.

winst · 09/30/10, 11:10 AM

Originally Posted by Feist-Mok

What offhand attacks are normalized that dagger normalization would have any impact? Main Gauche procs? Has there actually been any testing of this? Historically, as far as I know, sword spec procs have not been normalized, as they've been considered additional auto attacks, so that would be a significant change.

Quite interested in any further information on this myself. I've been fishing for a Lungbreaker for 4.0.1 as the only special offhand attacks combat uses right now is Killing Spree. It feels at least that the damage loss from using KS with a dagger would be at least marginally offset by the poison/combat potency procs. What I am unsure of however, is that if without poison talents we are watching our deadly stack a little closer and considering shiv. In this situation it might be worthwhile sticking to the axes.

fierydemise · 09/30/10, 11:59 AM

I haven't been on the PTR in 2 builds so it could have changed but when I tested last I found energy regen rate to be energy/sec=12.5+.11*haste%.

Used FoK 10 times, recorded time till 10th FoK (how long it took to regen 400 energy)

Haste %   Time    Energy/Sec
0          32       12.5
1.04       31.66    12.63
5.55       30.4     13.157
28.03      25.7     15.56

ieatpaperbag · 09/30/10, 12:34 PM

Originally Posted by Lightshadow

I’m assuming that mastery procs are white hits instead of yellow (my rogue on the PTR has the mastery bug though, so I’d appreciate confirmation of this) and that said hits can proc combat potency (as per the GC post).

On beta, my MG procs are currently showing up as yellow, tonight or tomorrow I will do some testing to see if its damage is normalized or not.

Lightshadow · 09/30/10, 4:06 PM

Originally Posted by fierydemise

I haven't been on the PTR in 2 builds so it could have changed but when I tested last I found energy regen rate to be energy/sec=12.5+.11*haste%.

Used FoK 10 times, recorded time till 10th FoK (how long it took to regen 400 energy)

Haste %   Time    Energy/Sec
0          32       12.5
1.04       31.66    12.63
5.55       30.4     13.157
28.03      25.7     15.56

Sadly both of the hypotheses are pretty good fits for the data, especially given that lag between the client and the server can introduce non-trivial noise to the process. I'll try to do a 5 minute test tonight with as much haste as I can stack to see if I can't get something more definitive.

ieatpaperbag · 09/30/10, 8:33 PM

It does appear that MG is a normalized attack; I do not often make these calculations by hand so please look it over and point out mistakes. Also, if these samples are too small, let me know and I’ll redo them when the servers are back up.

Armor
My rogue is lvl 83 and I did the test on a lvl 80 test dummy, I am not familiar with how much damage reduction from armor I should expect, so I did a test for this specifically using only Poniard in my MH. With no poisons, no enchant procs, trinkets, or the ICC ring, I had a constant 4790 AP. I expected to see and average of (it's 8.8 dps on the beta contrary to what wowhead shows).

1.3 * 8.8 + 4790 * 1.3 / 14 = 456.23

I instead had an average of 316 damage which gives approximately 30.74% damage reduction from armor.

1 - 316 / 456.23 = 0.3074

Slow Sword
Using the same dagger in the MH and Bloodvenom Blade in my OH, I had 4973 AP.

Expected Normalized Damage:
(2.6 * 250.6 + 4973 * 2.4 / 14) * 0.875 * (1 - 0.3074) = 912

Expected Non-normalized Damage:
(2.6 * 250.6 + 4973 * 2.6 / 14) * 0.875 * (1 - 0.3074) = 955

Fast Axe
Using the same dagger in the MH and Scourgeborne Waraxe in my OH, I had 4989 AP.

Expected Normalized Damage:
(1.5 * 250.7 + 4989 * 2.4 / 14) * 0.875 * (1 - 0.3074) = 746

Expected Non-normalized Damage:
(1.5 * 250.7 + 4989 * 1.5 / 14) * 0.875 * (1 - 0.3074) = 552

Fast Dagger
Using the same dagger in the MH and Lungbreaker in my OH, I had 4960 AP (it shows as 250.6 dps on beta). This test was quite a bit shorter because the server shutdown on me.

Expected Normalized Damage:
(1.4 * 250.6 + 4960 * 1.7 / 14) * 0.875 * (1 - 0.3074) = 578

Expected Non-normalized Damage:
(1.4 * 250.6 + 4960 * 1.4 / 14) * 0.875 * (1 - 0.3074) = 513

Lightshadow · 10/01/10, 1:24 PM

A new version is up, I've updated the first post with new conclusions.

Updates in 0.2

MG is now a yellow/normalized attack as per research in the cata mechanics thread.
AR energy now scales with haste (see cata mechanics thread)
RS updated for the announced 125% weapon damage buff, as well as no longer normalized.
SS updated for announced buff to imp SS
AP scaling updated for announced vitality buff
Eviscerate damage updated for announced buff. (The patch notes only mention the base damage and upper end AP scaling, both of which are 30%, so I've assumed the lower end AP scaling is also 30% higher than before).

Known bugs in v0.2

The spreadsheet cycle is using RS far too aggressively when the RS glyph is off.

• Aldriana · 10/01/10, 2:08 PM

On live right now low end Evis scaling and high end Evis scaling are the same - .07 AP per combo point (i.e. Wowhead is wrong). Hence I would assume that the scaling they listed (.091 AP per combo point) is for both low and high ends. We should test that (which is why its listed in the mechanics thread) but for now its probably as good a guess as any.

Lightshadow · 10/01/10, 2:43 PM

Oh, it's not just wowhead that's wrong then, it's also the tooltip? That's actually what I used to calculate the formula for eviscerate, didn't occur to me to check that it might be lying. The PTR patch with these changes won't hit till tonight, but I'll take a look afterwards.

• Aldriana · 10/01/10, 3:10 PM

Well, do half a dozen eviscerates and tell me: is the damage varying by the inherent 254 damage spread of the ability, or is it varying by the 1500+ damage you'd expect from having min and max damage separated by 20% of your AP as well?

envy860 · 10/01/10, 6:30 PM

Any plans for adding the ruby sanctum trinkets?

Edit: Scratch that, wasn't thinking. Figured out how to do this.

ninjawhite · 10/02/10, 1:05 AM

I don't know if you've looked at it or not, but maybe mongoose might be viable again. I know right now on live (according to Aldriana's combat spreadsheet) my agi is rated at 2.05 and haste at 1.96, for the sound of things agi should increase by around 1 ep atleast, and haste should go up as well. Mongoose at 120 agi and 30 haste rating might just come out ahead of zerking assuming they have the same proc rate.