Sith Assassin Deception analysis

[Alratan]

It does 3 swings dealing 100% weapon damage each or 100% weapon damage split over the swings? And yes, I'd assume that all multiple-swings attacks like Thrash and VS will have separate hit/miss chances for each swings. Three swings for SS changes things - a lot!

[Grim]

For example, lets say the tooltip says betwen 165-180 over 3 swings (i'll get the exact verbiage when I get home). Each hit landed for about 45-50 damage (yay armor).

It's a tricky thing to watch in the floating combat text because it doesn't shoot out a new number each hit, instead it updates the number as each hit lands so I was having to follow it and noticed that most times it was finishing between 135-150 damage, and sometimes it was stopping in the 90's.

Same for double strike, I would get between 140-150 on the first strike, and then it would adjust the number on its ark to 280-300... but sometimes it stayed at the 140-150 figure which tells me one of the 2 strikes missed.

And then of course every now and then you could see the bigger crit number.

[Gorodetski]

aah cool, so it's a nice spammable filler attack I'll stick that somewhere easy to spam

also i like the sound of the updating numbers, would be cool if they could add a seperate 'miss' arc somewhere else though,

[Alratan]

Hrm, it's also worth bearing in mind that Cloud Mind should be used every cooldown, which can be substituted for a SS.

[Piree]

(09-16-2011 06:39 AM)Alratan Wrote:  Hrm, it's also worth bearing in mind that Cloud Mind should be used every cooldown, which can be substituted for a SS.

Not necessarily, if your tank can keep aggro, I don't see the use of Cloud Mind.

[Alratan]

While true, if tanks threat is done in a reasonable way (e.g. WoW TBC not WoW WotLK), some form of proactive thread reduction from each player will be required. Given that we lack threat meters and Cloud Mind has a cooldown of 45s, using it will not damage the overall DPS too much but will grant significant comfort zones. Testing will probably be required on a tank-by-tank basis, but I imagine that it will play an important part of the rotation.

[Kaedis]

Quote:lets take a hypothetical situation. 5min Boss fight, you use Discharge every 15 seconds, that's 20 discharges overall, even if you delay each of those discharges by ONE second, you end up with one Discharge every 16 seconds, for 18 overall in 5 mins, so that's 2 'lost' discharges, and that's with the tiny delay of 1 second, the longer the delay, the more 'lost' discharges you get.

That would only apply if you delayed by 1 seconds on *every single discharge*. Delaying by 1 seconds on a single discharge loses you 1/15th the average damage of a Discharge (or 1/12th, with Recirculation). That said, you also have to take into account the dps lost by delaying the entire rotation.

Incidentally, I ran the proc chance numbers just for the sake of having them:

Since VS only has a single chance to proc, a 12-second cycle nets you 5 proc chances (4 VS at 37.5% chance each, 1 Maul at 25% chance), which gives you:

Incorrect calculations, refer to my later post here
39.2% chance of exactly 1 proc
27.3% chance of exactly 2
9.44% chance of exactly 3
1.62% chance of exactly 4
0.111% chance of exactly 5

With an average of 1.291 charges (+7.746% additional damage to Discharge) per Discharge.

For these calculations, I'm using the following assumptions (most derived from Draeveus's post):
Shock - 944 base damage, +30% damage modifier, force attack
Voltaic Slash - 808 base damage, physical attack
Maul - 1404 base damage, physical attack
Discharge - 906 base damage, force attack
20% crit chance
200% crit damage on all 4 attacks
50% uptime on Exploitive Strikes (see below)
Base rotation 4xVS 2xShock 1xMaul 1xDischarge per 12 seconds

Now, it's remarkably simple to calculate the value of delaying. Using the above assumptions, with an average of 1.291 charges per Discharge, base damage per cycle is 9888.515. This is 824.04 DPS, or 1236.06 damage per GCD (DPG). If you delay Discharge by a single GCD, and assume that GCD is filled with the 2:1 weighted average of the damage done by Voltaic Slash and Shock, that GCD will be filled with 1229.231 damage (2/3rds of a VS + 1/3rd of a Shock + 1.2% of the damage of a Discharge from the additional average charges), at the expense of 1236.06 damage. This is a loss of 6.8 DPG, or roughly 0.55% dps. Delaying by a second GCD sees the same effect, reducing DPG by a further 5.5 (total 12.3) and dps by another 0.44% (total 1.0%). In other words, delaying is a dps loss, but not a catastrophic one.

However, this is assuming VS does the same damage as Thrash. In order for delaying to be a perfect wash, VS needs to do 1104.38 damage, 36.7% more damage than Thrash. If it does more than that, it's a slight dps increase to delay.

Edit 2: Redid the above numbers. On retrospection, my math was waaaay off. Sooo much simpler to calculate by DPG.

ES Uptime: ES has no internal cooldown, and we have either 2 or 3 procable attacks every 12 seconds for it. The formula for uptime for a no-ICD effect is 1 - (1 - [ProcChance]) ^ ( [BuffDuration] / [TimeBetweenAttacks] ). Assuming a 20% crit chance, ES will have a 56.7% uptime if Discharge can proc it, 42.8% if it cannot.

Edit: Additional calculations:

Directly comparing the damage bonus provided by 2/2 Thrashing Blades and 2/3 Chain Shock. Assuming Draeveus's numbers damage numbers, 20% crit, 50% uptime on ES (about right with 3 procing attacks per 12 seconds, 20% proc (crit) rate). At that rate, 2/3 Chain Shock will do 1699 DPM (3 procs per minute, 566.7 average damage per proc with 20% crit) and Thrashing Blades, applied to Thrash instead of VS, will do 1207 DPM (20 Thrashes, 48.5 additional damage each with 24.5% effective crit rate). Give this, Thrashing Blades would only be worth taking over 2/3 Chain Shock if Voltaic Slash does a minimum of 41% more damage than Thrash. If Crackling Blasts does not apply to the Chain Shock procs, VS must do 29% more damage than Thrash for Thrashing Blades to be equal.

[Kaedis]

Oh, more data on Exploitive Strikes. I've graphed the uptime and effective crit chance as a function of base crit chance, along with base crit chance for comparison. Assumpion: base force crit chance = base melee crit chance. Probably an incorrect one, but lacking data of how they relate, there's no better way to handle it as far as I can tell. The uptime graph will be accurate regardless, but the effective and base crit chance may change intercept and slope based on the interaction between force crit and melee crit.

[Alratan]

Nice work Kore, on both. The Chain Shock information is also good to know for Tanking Assassins, as well, as a similar principle applies.

Regarding VS and Thrash, it's remotely possible that this one will change depending on gear, as Voltaic Slash may have a different coefficient for Thrash, but I don't think it's very likely. I'll have a look tonight (~10-12 hours) to see if I can find some reliable tooltip descriptions for Thrash and compare to Voltaic Slash, to see if there are any insights.

Regarding Exploitative Strikes, everything we know so far indicates that we can safely assume that melee crit chance = force crit chance for Assassins, except in the case that Exploitative Strikes is up, as they're both derived from the same stats. Additionally, screenshots/videos of high level players seems to indicate that a base crit chance of about 30% is reasonable, so if your graph is correct we know that we can generally assume an uptime of ~73% for ES, which might come in handy to know.

[Kaedis]

Ok, so I badly mangled the math on the charges in my post. I've redone them as a poisson binomial distribution (the proper way to calculate them), and we have the following tables for the probabilities of having each level of charge when Discharge is cast, along with the average number, based on the rotation used

4xVS 2xShock 1xMaul 1xDischarge
0 charges: 7.50847%
1 charge : 25.8625%
2 charges: 35.0395%
3 charges: 23.2155%
4 charges: 7.4581%
5 charges: 0.915909%
Averaged : 2.00000 (+12.0000% damage)

4xVS 2xShock 1xMaul 1xDischarge, +1 GCD (replaced 2charges:1 with VS or Shock)
0 charges: 5.63135%
1 charge : 20.5031%
2 charges: 32.3598%
3 charges: 26.6681%
4 charges: 12.0593%
5 charges: 3.09368%
Averaged : 2.28933 (+13.7360% damage)

4xVS 2xShock 1xMaul 1xDischarge, +2 GCDs (replaced 2charges:1 with VS or Shock)
0 charges: 4.22351%
1 charge : 17.3633%
2 charges: 29.8774%
3 charges: 27.8149%
4 charges: 15.0910%
5 charges: 5.62980%
Averaged : 2.49076 (+14.9446% damage)

Throwing these numbers into the calculations for delaying for charges doesn't really budge them at all. The only alteration that's above the level of uncertainty in these calculations is the necessary damage on VS, dropping from 1104.38 to 1104.15. Even using a 73% uptime on ES instead of the 50% I'd assumed doesn't change much, dropping the necessary VS damage to 1103.31, or 136.6% of Thrash's damage.

Numbers change is mainly just useful for determining the benefit of the charges, which average to a 12% increase in Discharge damage. Based on the numbers in Draeveus's post, this is equivalent to a 1.1% increase in damage done by the cycle. This is excluding damage done by Surging Charge's proc, since I don't have numbers for that effect at the same level and gear that Draeveus's numbers were. This is also not taking into account the fact that the target will likely have a differing mitigation level between physical and force attacks, and that Maul bypasses much of that resistance. I won't be able to calculate it's true overall dps benefit until we have solid math on the base and coefficient for each ability, the likely level of stats while in a level 50 operation, and the average mitigation for a boss-type mob at 50.