would our shadow 4 set change our starting rotation?

[Weena2]

100% agree with the above post, its more or less all the infomation you will need to know on dropping MB from your rotation.

General rule incase you didn't read it is basically as your MS/Latency increases, the less dps you get from using MB every CD.

[Worshaka]

Interesting but im not sure I agree with a lot of what you say... while it may be true that DP > VT for DPEt, when you evaluate DP and VT as dps cycles im not sure delaying VT by a GCD results in higher DPS. MB was the highest in the priority in 3.2 even though VT and DP outstripped the DPEt by almost 3 times the amount of MB.

In any case your post has intrigued me enough to run some numbers and see if my impressions are actually correct or not.

[Griemak]

There was an error in both of my posts above where an addition was performed instead of a multiplication regarding haste. All has been edited to fix this error.

[Worshaka]

Interesting but im not sure I agree with a lot of what you say... while it may be true that DP > VT for DPEt, when you evaluate DP and VT as dps cycles im not sure delaying VT by a GCD results in higher DPS. MB was the highest in the priority in 3.2 even though VT and DP outstripped the DPEt by almost 3 times the amount of MB.

In any case your post has intrigued me enough to run some numbers and see if my impressions are actually correct or not.

I need to do this a bit more concisely but here is my analysis based off some numbers I was producing in 25man ICC last night.

VT average tick 2700
DP average tick 2200 (crypt fever present)

We can ignore haste and crit, reason being is that they both scale both spells identically... while haste and crit will change the nominal DPS figure it wont' change the % relationship between the 2.

As I suspected while DP is higher DPEt, this doesn't mean it's a higher priority spell. Spriest DPS is about maximising DPS cycles and an unhasted DP is a 24 second DPS cycle vs VT which is a 15 second cycle. Delaying DP by 1 GCD (1.5 seconds) is less of a proprotion of lost cycle DPS than it is to VT which is on the quicker cycle.

Using the numbers above VT provides 13500 DPEt vs DP 17600 + 30% for IDP = 22880. When you analyse these numbers as a DPS figure you come up with:

VT = 900 DPS
DP = 953.33 DPS

If you then analyse their DPS by delaying their cycles by 1 GCD, ie. VT goes from a 15 sec cycle to a 16.5 & DP goes from 24 to 25.5 you see that the DPS changes are:

VT = 818.18 DPS
DP = 897.25 DPS

That means that delaying VT is causing a DPS loss of 81.82 DPS vs DP's 56.08... it is therefore desirable to delay DP because the net loss in DPS is less than that of VT. The percentage difference between these numbers is 45.9% and that figure remains the same for whatever haste and crit numbers you decide to come up with. I ran the same anlaysis with 1050 haste rating (about what I have) and factored in the 3% & 5% haste raid buffs and the numbers come out that if you delay VT you lose out on 116.82 DPS vs DP's 80.07, as you can see the ratio remains the same.

So while DP has a far higher DPEt this doesn't make it higher in the priority in terms of your theoretical max DPS... having said that DP and VT rarely conflict in my experience and if they do conflict once per minute, prioritising DP over VT is hardly going to make a noticible impact on your overall damage. The largest factor will be if casting DP or VT first guarantees another DP/VT conflict in the near future and that will all depend on your level of haste, ping and reaction time. That's something individuals are going to have to identify and experiment with.

[Griemak]

Subtracting DP's total DPEt from the dot component = IDP's DPE is 5,280.

DPS cycles are great for mages and rogues, we don't deal in DPS cycles, we deal in DPEt cycles. Delaying DP will increase DPS at the moment you do it, but how does it impact total damage done as we are DoT class, not a nuke class. If you were to manage our spell priority by calculating instant DPS cycles based on delay, you would find that no other spell will beat Mind Flay, yet we know this is not true; Mind Flay is not the top priority spell to increase damage done.

For example, MF DPEt for the average ilvl258 Spriest is around 5300. Over 3 seconds this is a dps of 1766. If we delay it 1.5 seconds it drops to 1177, a loss of 589 dps. According to this, we should only Mind Flay, nothing else. This is true, when you examine DPS cycles, but you will result in less total damage done and less overall dps as a result, even though your action right now is superior DPS to cast mind flay above all else, it is not superior damage output over the total fight length to do so.

Delaying a dot pushes its uptime back allowing the fight to end before a tick could occur, which means the delay affects overall ticks. This is not reflected in the DPS calculation at the moment in conflict, but rather at the end of the fight. If we delay one of these dots by 1.5 seconds, it is exactly 1/2 of a dot tick. Haste and crit do not matter as both DP and VT are affected equally. We have a 50% chance of loosing that tick, over the long haul, this will be a 50% loss of a tick. If we delay DP cast by 16 times within a fight (384 seconds without haste is rather unrealistic this would occur, however with 900+ haste and 40 seconds of the fight under bloodlust/heroism, we wind up with a minimum fight length to delay it 16 times equal to ~250 seconds) we also loose one IDP execution.

Probable loss in total damage done by delaying DP: 1/16*5280 + 1/2*2200 = 1430 per conflict
Probable loss in total damage done by delaying VT: 1/2*2700 = 1350 per conflict

I agree your DPS numbers are correct, but they are a snapshot of the DPS at the moment in conflict and not reflective of the overall loss in damage done / entire fight length. Where it gets nasty is when IDP isn't involved. If there is no chance that 16 conflicts could occur, VT results in more damage lost by delaying it and the priority should be VT-DP. Where a chance that 16 conflicts could occur, DP-VT. Interesting, never noticed it that way before. There will be certain haste levels where VT and DP's lenghts cause more conflict than others.... for those certain haste levels, DP-VT is superior. For haste levels where VT and DP conflicts are fewer than 16, VT-DP is superior. This just got really ugly, as conflicts is not always VT and DP expiring at the same moment, it also includes target swaps such as Blood Princes and reapplication after fall-offs/buffs such as tear gas events on Putricide, any bloodlust/heroism where you reapply your dots, or special buff mechanics such as Blood Queen. There may not be a generalized "correct" answer here....

[Worshaka]

Subtracting DP's total DPEt from the dot component = IDP's DPE is 5,280.

DPS cycles are great for mages and rogues, we don't deal in DPS cycles, we deal in DPEt cycles. Delaying DP will increase DPS at the moment you do it, but how does it impact total damage done as we are DoT class, not a nuke class. If you were to manage our spell priority by calculating instant DPS cycles based on delay, you would find that no other spell will beat Mind Flay, yet we know this is not true; Mind Flay is not the top priority spell to increase damage done.

For example, MF DPEt for the average ilvl258 Spriest is around 5300. Over 3 seconds this is a dps of 1766. If we delay it 1.5 seconds it drops to 1177, a loss of 589 dps. According to this, we should only Mind Flay, nothing else. This is true, when you examine DPS cycles, but you will result in less total damage done and less overall dps as a result, even though your action right now is superior DPS to cast mind flay above all else, it is not superior damage output over the total fight length to do so.

Delaying a dot pushes its uptime back allowing the fight to end before a tick could occur, which means the delay affects overall ticks. This is not reflected in the DPS calculation at the moment in conflict, but rather at the end of the fight. If we delay one of these dots by 1.5 seconds, it is exactly 1/2 of a dot tick. Haste and crit do not matter as both DP and VT are affected equally. We have a 50% chance of loosing that tick, over the long haul, this will be a 50% loss of a tick. If we delay DP cast by 16 times within a fight (384 seconds without haste is rather unrealistic this would occur, however with 900+ haste and 40 seconds of the fight under bloodlust/heroism, we wind up with a minimum fight length to delay it 16 times equal to ~250 seconds) we also loose one IDP execution.

Probable loss in total damage done by delaying DP: 1/16*5280 + 1/2*2200 = 1430 per conflict
Probable loss in total damage done by delaying VT: 1/2*2700 = 1350 per conflict

I agree your DPS numbers are correct, but they are a snapshot of the DPS at the moment in conflict and not reflective of the overall loss in damage done / entire fight length. Where it gets nasty is when IDP isn't involved. If there is no chance that 16 conflicts could occur, VT results in more damage lost by delaying it and the priority should be VT-DP. Where a chance that 16 conflicts could occur, DP-VT. Interesting, never noticed it that way before. There will be certain haste levels where VT and DP's lenghts cause more conflict than others.... for those certain haste levels, DP-VT is superior. For haste levels where VT and DP conflicts are fewer than 16, VT-DP is superior. This just got really ugly, as conflicts is not always VT and DP expiring at the same moment, it also includes target swaps such as Blood Princes and reapplication after fall-offs/buffs such as tear gas events on Putricide, any bloodlust/heroism where you reapply your dots, or special buff mechanics such as Blood Queen. There may not be a generalized "correct" answer here....

That's an interesting perspective but I think we can both agree that unless VT/DP conflict regularly that choosing either to be first in the priority isn't going to make a difference practically speaking.

Just taking a look at my latest log for blood queen (a wipe where I didn't get bitten), Imp DP average hit was 3940 so I think your estimate above is far too high. It may be that you are taking crit into account however the VT and DP numbers do not take crit into account. So if we normalise the numbers (ie just leave crit out of it) you need to be looking at

Imp DP = 4000
VT = 2650 (this was the figure I got in my latest parse)
DP = 2050 (this was the figure I got in my latest parse)

Reason i'm changing the VT and DP numbers is because they directly correlate to the imp DP figure i've pulled out.

So using your perspective

Probable loss in total damage done by delaying DP: 1/16*4000 + 1/2*2200 = 1350 per conflict
Probable loss in total damage done by delaying VT: 1/2*2700 = 1325 per conflict

Comes up with the same conclusion but the numbers are far closer together. I think what we can really conclude is if you prioritise VT or DP during a conflict the DPS effect is negligble at best. It's certainly something I won't be stressing over and for me it's more important about noticing whether VT > DP creates more conflicts than DP > VT and vice versa. At my level of haste I can't say that VT and DP ever conflict... perhaps once every 5mins and if im losing out on 25 damage every 5mins, so be it.

[Nezoia]

Who cares about the starting rotation? More essencial is that Mind Blast seems to be a dps loss. Eoy did some testing, and gained a few hundred dps by only using Mind Blast to keep up Replenishment.

That you seriously care about your first 5 spells is... incredible stupid. They have extremely little meaning over a 5 minute fight.

(And your 2 minute testing on a non-heroic target dummy is invalid.)

[Weena2]

As I said previously, taking both latency/lag as well as spellpower and coefficients into account when using the 4set, we should defiantly be dropping mind blast totally from our rotation; or at the very least using it to keep up replenishment.

[Cinosanap]

with my priest and (4) Set bonus ,
i do
Vampiric Touch -> Mindblast ( to trigger Replenishment ASAP ) -> Devouring Plague -> Shadow Word: Pain -> Mindflay -> Mindflay . to get as much benefit as possible. and i end up at almost 7k dps on lord marrowgar

in TOTC25 i can do like 8k with the right buffs

[Nezoia]

Anyone calculated how using SW as a opener, to proc Black Magic, would be compared to not?

I somehow could imagine using SW/Mind Blast as non-dot part of the opener to get a Black Magic proc would be pretty win?

[Myzrael]

Anyone calculated how using SW as a opener, to proc Black Magic, would be compared to not?

I somehow could imagine using SW/Mind Blast as non-dot part of the opener to get a Black Magic proc would be pretty win?

I usually open with MF then MB and in most cases Black Magic also procs, I guess you can add SW if it doesn't.
Then again, I'm not that much into theorycrafting and maths, and I have no idea if delaying dots to apply them few secs later with 250more haste is dps gain or loss.

[Nezoia]

I usually open with MF then MB and in most cases Black Magic also procs, I guess you can add SW if it doesn't.
Then again, I'm not that much into theorycrafting and maths, and I have no idea if delaying dots to apply them few secs later with 250more haste is dps gain or loss.

Opening with MF is only possible on a fight where you aren't in movement at the pull. ICC been rather special in regard to this, most raid fights in the game tend to be running pulls, hence SW or DP as starters.