Theck's Simcraft Thread - WoD/6.x

Seems like I'm a little late in this thread. So, this post will contain several disconnected comments.

theckhd wrote:Now that we have a fairly solid baseline APL, let's recall the things we still want to test. Most of these will probably be done by simple hand-comparisons of profiles we tweak manually.

• Avenger's Shield - with and without Focused Shield
• Sacred Shield - optimal position for refreshing w/ empowered seals
• Sacred Shield - optimal duration remaining at which to refresh
• Seraphim/SotR - try different lockout periods before Seraphim cooldown is up.
• Seraphim - with DP glyphed
• AoE rotations

As an addition to Avengers shield, are we certain on it's postion when considering avengers shield with and without Grand Crusader procs. In particular I think AS /w GC may potentially be prioritized over HW during sanctified wrath.

How does using SotR perform versus WoG spam?

Also, and this one of low priority and may involve quite a bit of work. Does the conclussions hold true in the future teirs with significantly better gear and harder hitting bosses? Since Theck have already done all the ground work, this would "simply" be rerunning all the sims again with a significantly better geared toon (custom stats) and a new boss.

theckhd wrote:

Code: Select all

actions+=/seal_of_insight,if=talent.empowered_seals.enabled&buff.uthers_insight.remains<cooldown.judgment.remains
actions+=/seal_of_righteousness,if=talent.empowered_seals.enabled&buff.uthers_insight.remains>cooldown.judgment.remains&buff.liadrins_righteousness.down
actions+=/seal_of_truth,if=talent.empowered_seals.enabled&buff.uthers_insight.remains>cooldown.judgment.remains&buff.liadrins_righteousness.remains>cooldown.judgment.remains&buff.maraads_truth.down

To explain that in words: use Seal of Insight if Uthers Insight will expire before our next Judgment, use SoR if Uther's Insight will last until after our next Judgment but Liadrin's Righteousness is already down, and use Seal of Truth if both Insight and Righteousness buffs will last until after our next Judgment but Maraad's Truth is down.

What I wasn't sure about here is the "down" part. It wasn't clear to me whether it would be optimal to wait to swap seals until after the respective buff has already expired, or whether we should try to refresh them before they expired. So in addition to his version of the conditionals (which I call the "down" or "d" version), I added a second version that looks like this (the "up" or "u" version):

I hope I can provide a bit of rationale to this part since they originate from this post. I used 4 different conditionals for each buff:

Code: Select all

buff.uthers_insight.down
buff.uthers_insight.remains<time                               -- I used any amount of time here ranging from 1-10
buff.uthers_insight.remains< cooldown.judgment.remains
buff.uthers_insight.remains<(cooldown.judgemen.remains + time) -- as above

Each of these conditionals could be applied to each buff seperately. Given the addition of variable timing to some of the conditions this resulted in a monumental problem space. However, these conditions can be ordered in how much priority they give to refreshing each seal where the first had gave the least priority and the latter gave the highest. Since we are already ordering the seals in priority in the APL it made sense to use the same ordering for the conditionals. The simple time conditional was removed early in the process and replaced with 2 last conditionals. What remained was just to test all remaining combinations iteratively - which I did. Now, this wasn't included in the post since the post wasn't about optimizing the EmpS conditionals but rather to test the relative benefit of the new haste buff (which was just changed from 15% atk speed to 10% haste). The tests immidiatly showed that it was best (TMI wise) to let truth and righteousness drop rather then refresh early. However, insight was worth refreshing early using the judgement cooldown conditional. Finally, I achieved inconclusive results with respect to adding additional time to the judgment cooldown conditional and therefore the simpler was choosen.

I never fully understood the rationalle behind letting righteousness and truth drop completely before switching seals before now. Even the cooldown.judgment conditional wont keep refresh the seal until we are 100% certain that the buff will drop before the next time we can refresh it. However, as the post showed, the empowered righteousness buff alone provides a significant TMI gain over just righteousness. However, as Theck concludes having near 100% uptime on SoI and UthersI is significantly better then 50% uptime on SoI, 50% uptime on SoR and 100% uptime on both UthersI and LiadinsR. That is 50% uptime on SoI > 100% uptime on LiadinsR with respect to TMI. Based on this it's no wonder that I saw a TMI gain for letting the empowered rightenousness and truth drop for prolonged periods of time in favor of higher SoI uptime. With this said I'm very happy that you improved the conditional to be a 100% greedy algorithm, that was my goal in the first place however I could never manage to get it right without being overly complex. Also sorry for the lack of !seal, I was under the impression that seals would never refresh themself like that - also I didn't know about ! :(.

With the result of the EmpS sim, I'm quite happy that the 3 seal rotation is not the best rotation for everything. Especially the tank switch fights are significantly more interesting. I would imagine us using SoI while tanking, and then switching to SoT or SoR the instant the boss gets taunted off us. And as an added benefit we can probably time it so what we have all 3 buffs active and SoI when we taunt back. While it might not be as strong as seraphim being active on taunt, EmpS does not loose anything when the taunt does not line up exactly at the 30 sec interval.

As a last comment, there was a 60% self healing nerf to tanks in the form of a hidden debuff on beta back in september (source) which also used to reduce our lay on hands to 40% healing. The lay on hands part have since been fixed. However, I couldn't find any information about this nerf being removed since. So, I have to ask, is this still active and if so has this been taken into account in simcraft?

Zephyr wrote:As a last comment, there was a 60% self healing nerf to tanks in the form of a hidden debuff on beta back in september (source) which also used to reduce our lay on hands to 40% healing. The lay on hands part have since been fixed. However, I couldn't find any information about this nerf being removed since. So, I have to ask, is this still active and if so has this been taken into account in simcraft?

The -60% healing debuff is actually part of the Resolve spell data, and it's being applied in SimC. Neither part of Resolve applies to percent heals (in game or in SimC). I have LoH coded as a percent heal, so it should properly be ducking the -60%.

Also thanks for the clarification on where the conditionals for Empowered Seals came from.

I agree that checking for e.g. AS+GC vs AS is something to pursue. We didn't see any advantage in the sims we ran, but that was early on in the process. There may still be a slight gain to be had by doing e.g. AS+GC>HW+SW>AS.

Also note that it's perfectly fine for you guys to start testing these things. It may take me a while to get to all of them, so feel free to beat me to any of these tests.

theckhd wrote:Now that we have a fairly solid baseline APL, let's recall the things we still want to test. Most of these will probably be done by simple hand-comparisons of profiles we tweak manually.

• Seraphim/SotR - try different lockout periods before Seraphim cooldown is up.
• Seraphim - with DP glyphed

theckhd wrote:Also note that it's perfectly fine for you guys to start testing these things. It may take me a while to get to all of them, so feel free to beat me to any of these tests.

With that in mind I thought I'd try to test out the different lockouts on using SotR before Seraphim, as well as going back and re-testing the Divine Protection and Seraphim "chaining" vs "as available" with the changes Theck recommended (popping CDs on the pull). I actually ran short sims of each of those separately, but I'm posting both together because I noticed some odd behavior, which I think is more easily explainable seeing everything together.

Glyphs: alabaster_shield/focused_shield/divine_protection
Talents: 0032222
Throughout I'm using "Ch" to refer to chaining APLs and "NC" to refer to not chaining (using as available). The number at the end is the cooldown remaining on Seraphim that SotR should wait. The other seraphim sims Theck's been doing have been 5 for reference.

These are the 2 basic APLs I'm using, then just varying the wait time on the SotR line:

name=SS_SW_Sera_Ch_5

Code: Select all

# Executed before combat begins. Accepts non-harmful actions only.
actions.precombat=flask,type=greater_draenic_stamina_flask
actions.precombat+=/food,type=talador_surf_and_turf
actions.precombat+=/blessing_of_kings,if=(!aura.str_agi_int.up)&(aura.mastery.up)
actions.precombat+=/blessing_of_might,if=!aura.mastery.up
actions.precombat+=/seal_of_insight
actions.precombat+=/sacred_shield
# Snapshot raid buffed stats before combat begins and pre-potting is done.
actions.precombat+=/snapshot_stats
actions.precombat+=/potion,name=draenic_armor
# Executed every time the actor is available.
actions=/auto_attack
actions+=/speed_of_light,if=movement.remains>1
actions+=/blood_fury
actions+=/berserking
actions+=/arcane_torrent
actions+=/potion,name=draenic_armor,if=buff.shield_of_the_righteous.down&buff.seraphim.down&buff.divine_protection.down&buff.guardian_of_ancient_kings.down&buff.ardent_defender.down
# This section covers off-GCD spells.
actions+=/holy_avenger
actions+=/divine_protection,if=buff.seraphim.down
actions+=/seraphim,if=buff.divine_protection.down&cooldown.divine_protection.remains>0
actions+=/guardian_of_ancient_kings,if=time<5|(buff.holy_avenger.down&buff.shield_of_the_righteous.down&buff.divine_protection.down)
actions+=/ardent_defender,if=time<5|(buff.holy_avenger.down&buff.shield_of_the_righteous.down&buff.divine_protection.down&buff.guardian_of_ancient_kings.down)
actions+=/eternal_flame,if=buff.eternal_flame.remains<2&buff.bastion_of_glory.react>2&(holy_power>=3|buff.divine_purpose.react|buff.bastion_of_power.react)
actions+=/eternal_flame,if=buff.bastion_of_power.react&buff.bastion_of_glory.react>=5
actions+=/shield_of_the_righteous,if=(holy_power>=5|buff.divine_purpose.react|incoming_damage_1500ms>=health.max*0.3)&(!talent.seraphim.enabled|cooldown.seraphim.remains>5)
# GCD-bound spells start here
actions+=/crusader_strike
actions+=/judgment
actions+=/holy_wrath,if=talent.sanctified_wrath.enabled
actions+=/avengers_shield
actions+=/execution_sentence
actions+=/hammer_of_wrath
actions+=/lights_hammer
actions+=/holy_wrath,if=glyph.final_wrath.enabled&target.health.pct<=20
actions+=/sacred_shield,if=target.dot.sacred_shield.remains<2
actions+=/consecration,if=target.debuff.flying.down&!ticking
actions+=/holy_wrath
actions+=/holy_prism
actions+=/sacred_shield
actions+=/flash_of_light,if=talent.selfless_healer.enabled&buff.selfless_healer.stack>=3

name=SS_SW_Sera_NC_5

Code: Select all

# Executed before combat begins. Accepts non-harmful actions only.
actions.precombat=flask,type=greater_draenic_stamina_flask
actions.precombat+=/food,type=talador_surf_and_turf
actions.precombat+=/blessing_of_kings,if=(!aura.str_agi_int.up)&(aura.mastery.up)
actions.precombat+=/blessing_of_might,if=!aura.mastery.up
actions.precombat+=/seal_of_insight
actions.precombat+=/sacred_shield
# Snapshot raid buffed stats before combat begins and pre-potting is done.
actions.precombat+=/snapshot_stats
actions.precombat+=/potion,name=draenic_armor
# Executed every time the actor is available.
actions=/auto_attack
actions+=/speed_of_light,if=movement.remains>1
actions+=/blood_fury
actions+=/berserking
actions+=/arcane_torrent
actions+=/potion,name=draenic_armor,if=buff.shield_of_the_righteous.down&buff.seraphim.down&buff.divine_protection.down&buff.guardian_of_ancient_kings.down&buff.ardent_defender.down
# This section covers off-GCD spells.
actions+=/holy_avenger
actions+=/divine_protection
actions+=/seraphim
actions+=/guardian_of_ancient_kings,if=time<5|(buff.holy_avenger.down&buff.shield_of_the_righteous.down&buff.divine_protection.down)
actions+=/ardent_defender,if=time<5|(buff.holy_avenger.down&buff.shield_of_the_righteous.down&buff.divine_protection.down&buff.guardian_of_ancient_kings.down)
actions+=/eternal_flame,if=buff.eternal_flame.remains<2&buff.bastion_of_glory.react>2&(holy_power>=3|buff.divine_purpose.react|buff.bastion_of_power.react)
actions+=/eternal_flame,if=buff.bastion_of_power.react&buff.bastion_of_glory.react>=5
actions+=/shield_of_the_righteous,if=(holy_power>=5|buff.divine_purpose.react|incoming_damage_1500ms>=health.max*0.3)&(!talent.seraphim.enabled|cooldown.seraphim.remains>5)
# GCD-bound spells start here
actions+=/crusader_strike
actions+=/judgment
actions+=/holy_wrath,if=talent.sanctified_wrath.enabled
actions+=/avengers_shield
actions+=/execution_sentence
actions+=/hammer_of_wrath
actions+=/lights_hammer
actions+=/holy_wrath,if=glyph.final_wrath.enabled&target.health.pct<=20
actions+=/sacred_shield,if=target.dot.sacred_shield.remains<2
actions+=/consecration,if=target.debuff.flying.down&!ticking
actions+=/holy_wrath
actions+=/holy_prism
actions+=/sacred_shield
actions+=/flash_of_light,if=talent.selfless_healer.enabled&buff.selfless_healer.stack>=3

Here's the full report
And the full imported profile

Again, sorry I don't have a host to embed the results image (PM me a free host that let's me embed images?):

For those not wanting to download it the TMI results are:
SS_SW_Sera_NC_6 : 109.6k TMI
SS_SW_Sera_NC_7 : 109.9k TMI
SS_SW_Sera_NC_5 : 110.0k TMI
SS_SW_Sera_NC_1 : 110.6k TMI
SS_SW_Sera_Ch_2 : 110.9k TMI
SS_SW_Sera_NC_0 : 111.0k TMI
SS_SW_Sera_NC_4 : 111.0k TMI
SS_SW_Sera_NC_2 : 111.0k TMI
SS_SW_Sera_Ch_1 : 111.3k TMI
SS_SW_Sera_Ch_3 : 111.4k TMI
SS_SW_Sera_NC_3 : 111.5k TMI
SS_SW_Sera_Ch_4 : 111.8k TMI
SS_SW_Sera_Ch_0 : 111.9k TMI
SS_SW_Sera_Ch_5 : 112.2k TMI
SS_SW_Sera_Ch_6 : 112.7k TMI
SS_SW_Sera_Ch_7 : 113.2k TMI

So while not HUGE differences, the "not chaining" ques are winning again, sometimes by over 3K TMI compared to the same delay of "chaining" ques. That seemed really odd, chaining should be at least as good at reducing spikes right? Then I look at the ordering of the times and noticed that higher times do well for NC ques but not well for Ch ques. I started thinking about the rotation and remembered this:

theckhd wrote:

In the default gear set we're using we have just over 20% haste raid buffed (before procs). So if I recall right, that's 1.5 / (1 + 0.2) or a 1.25 sec GCD. That means that based on the APL we're popping our 8 sec Divine Protection at the start of the fight, but only 6.25 seconds later we have 5 HP. In the "chaining" APL, Seraphim can't cast at this point because DP is still up, and SotR can't cast because it's waiting on Seraphim. Our next cast is CS which gives us a 6th HP, which is wasted. This means the chaining ques should show an overrun of HP, and looking at the graphs of HP gain on those sims I think that's exactly what I see. Furthermore, since DP and Seraphim are both 30 sec CDs, this pattern may to be happening more than once throughout each fight, though I'm not as certain on that. In order to test this hypothesis I re-ran the sim changing the Seraphim line on the chaining ques to:

Code: Select all

actions+=/seraphim,if=(buff.divine_protection.down&cooldown.divine_protection.remains>0)|(holy_power>=5&(cooldown.crusader_strike.remains<=1.5|cooldown.judgment.remains<=1.5|buff.grand_crusader.react|(cooldown.holy_wrath.remains<=1.5&talent.sanctified_wrath.enabled)))

Basically what I'm trying to do is have Seraphim wait on DP to run out, unless the next ability will generate HP. There may be a cleaner way to do that, but let's look at the results using this method.

Here's the full report
And the full imported profile

And image of the results:

And TMI in text form:
SS_SW_Sera_NC_6 : 109.2k TMI
SS_SW_Sera_Ch_6 : 109.4k TMI
SS_SW_Sera_NC_7 : 109.5k TMI
SS_SW_Sera_NC_5 : 109.6k TMI
SS_SW_Sera_Ch_5 : 109.7k TMI
SS_SW_Sera_Ch_7 : 109.8k TMI
SS_SW_Sera_Ch_4 : 110.4k TMI
SS_SW_Sera_NC_4 : 110.7k TMI

Much better! The "not chaining" is still winning slightly at most of the times, but only by 0.2k to 0.3k TMI, pretty negligible. That would seem to further confirm that the "chaining" ques were wasting HP waiting on Seraphim. I'm not sure what's still causing the slight advantage to not chaining the CDs, but it could be as simple as the very, very slight reduction in uptime for Seraphim and DP in the chaining sims?

I didn't sim every wait time with the modified Seraphim line (ran out of time), but so far a 6 second wait is looking like the winner. I also wanted to get some other eyes on this first to make sure I wasn't making an error somewhere and to decide if the Seraphim line I'm using above is the best correction. After that we can sim more wait times to double check the 6 sec wait.

TLDR:
The Seraphim use in the default APL looks to be slightly less than optimal. It should probably either not wait for Divine Protection to fall off or check to make sure we won't waste HP by holding Seraphim for Divine Protection. Also a 6 second hold on SotR to build HP for Seraphim might be slightly better than a 5 second hold.

Are you also using SotR when it doesn't need to be pooled during the Seraphim uptime? Even at 10% haste, you should be able to SotR during Seraphim's uptime for even more damage reduction/DPS and still be able to pool 5 HoPo by the time Seraphim is over for 2 more SotRs. In other words, you should be able to generate 8 HoPo within the 15 seconds Seraphim is up. There should never be a point in time where HoPo is wasted because of pooling... if I'm understanding your post correctly. Nice work by the way... good info.

The chain should look something like this. The 3 HoPo that is spent during Seraphim's uptime can be used as either a WoG or an SotR. Seraphim has the same flexibility as Empowered Seals (the choice to go for more healing or more DPS) in that respect, making Seraphim superior in every single way.

Divine Protection ---> Seraphim (+ 1x SotR or WoG) ---> 2-3 charges of shield of the righteous ---> Divine Protection ---> Seraphim (+ 1x SotR or WoG) ---> 2-3 charges of shield of the righteous

In other words in HoPo terms:

Divine Protection ---> Seraphim (3 HoPo, spend, 5HoPo) ---> SotR, 1 HoPo, SotR ---> Divine Protection ---> Seraphim (3 HoPo, spend, 5HoPo) ---> SotR, 1 HoPo, SotR

In practice, there will be overlap of the CDs as haste increases giving you more room to increase the damage reduction. There will also come a point with enough haste where you will actually pool HoPo too quickly but not quickly enough to spend 3 HoPo then build the necessary amount to chain again. This is the reason why I've suggested that haste may not be as valuable as multistrike considering how multistrike seems to be simming better in DPS and HPS anyways (need confirmation on this... I don't trust my own sim work).

I wish I was good with the APL scripting so that I could come up with something myself.

Danielspally wrote:Are you also using SotR when it doesn't need to be pooled during the Seraphim uptime?...
There should never be a point in time where HoPo is wasted because of pooling...

Yes, the current APL certainly uses SotR during Seraphim uptime and downtime. The only time it doesn't use SotR is during a "lockout" period right before Seraphim is available. That was written into the sim so that we can pool HP for Seraphim use, which makes sense.

I also agree with your statement that HoPo should never be wasted because of pooling. The trick is getting the sim to model that behavior, which is harder than it sounds sometimes. In fact, just to clarify, everything I'm describing in my last post is not truly a rotational issue as much as a simulation issue. I doubt most reading this forum would think it was proper to just continue along with their rotation after hitting 5 HoPo, instead of hitting Seraphim or SotR to use it, but that's what I think the sim was doing. So in actual play, I would still agree chaining Divine Protection and Seraphim may still be better or at least equivalent, I'm just trying to get the sim to reflect that or understand why it's not.

An additional thought, though I can't sim it right now: I would expect DP procs and HA to show similar anomalous behavior, if the HA buff or DP procs are up during the lockout period. In other words, the sim might still encounter times when DP procs or HA leads to excess HoPo, but the lockout prevents SotR use and Divine Protection or Seraphim CD prevents Seraphim use. If that's true, we'd need conditionals to look for those buffs and either cast an extra SotR or cast Seraphim early and overlap it with Divine Protection slightly. In fact it would be interesting to see which use would be more optimal, though I suspect the difference wouldn't be tremendous. I'll try to look at that later if I can.

Code: Select all

Simple Action Stats    Execute    Interval
Sacred Shield          34.7       12.70sec

I'm a bit confused here. Does that mean sacred shield was actually cast 35 times? Or is this a combined cast+proc rate?

Edit: Also, does it make sense that SS was cast 2.4 seconds in, but the first tick didn't happen until 91s? First SotR was at 61.9 seconds in?

Code: Select all

Dynamic Buffs             Start   Refresh   Interval   Trigger   Up-Time   Benefit
Sacred Shield             2.4     32.3      293.8sec   12.7sec   95.89%    95.89%
Sacred Shield (_tick)     91      6.7       5.0sec     4.6sec    39.12%    100.00%
Shield of the Righteous   61.9    0         7.1sec     7.1sec    41.73%    41.74%

Fundus wrote:Again, sorry I don't have a host to embed the results image (PM me a free host that let's me embed images?):

I use Gyazo, mostly because I take enough screenshots that the ability to shift+ctrl+c and instantly upload my screen to the intarweb is really handy.

marsbubble wrote:

Code: Select all

Simple Action Stats    Execute    Interval
Sacred Shield          34.7       12.70sec

I'm a bit confused here. Does that mean sacred shield was actually cast 35 times? Or is this a combined cast+proc rate?

It should mean that SS was cast 35 times, which is consistent with a cast interval of 12.70 seconds. The absorb bubbles are called sacred_shield_tick.

marsbubble wrote:Edit: Also, does it make sense that SS was cast 2.4 seconds in, but the first tick didn't happen until 91s? First SotR was at 61.9 seconds in?

Code: Select all

Dynamic Buffs             Start   Refresh   Interval   Trigger   Up-Time   Benefit
Sacred Shield             2.4     32.3      293.8sec   12.7sec   95.89%    95.89%
Sacred Shield (_tick)     91      6.7       5.0sec     4.6sec    39.12%    100.00%
Shield of the Righteous   61.9    0         7.1sec     7.1sec    41.73%    41.74%

Nope, neither of those make very much sense. But that's definitely what's happening.

e: or rather, that's what's being reported. I don't believe for a second that it's actually what the actor is doing, which means that it's probably some sort of reporting bug.

Oops, my bad. Looking at it more carefully, "Start" isn't the time at which the buff started. It's the number of times the buff was applied (i.e. when it wasn't already up). Thus, the number of casts ought to be Start+Refresh.

I think that "Interval" ought to be the interval between "Starts", while "Trigger" is the average interval between casts ("Start" or "Refresh"). I haven't checked that in the code, but it seems reasonable based on the data above (example: (32.3+2.4)*12.7=440, which is about the fight length).

Quick re-run of two sims based on the recent changes (SoR buff, Resolve changes, etc.).

Empowered Seals playoffs:
https://www.dropbox.com/s/m4hoddevspv9q ... .html?dl=0



Main thing to note here is that the SoR buff has had a significant effect. It's now more or less equivalent to prioritize RT to TR (Censure uptime should be the major difference). It's also brought up the DPS of the 3-seal rotations a bit. Overall though, nothing really changes in our recommendations that get carried over to the talent sim. Speaking of...


Talent sim:
https://www.dropbox.com/s/nyd5jvhk8t2pm ... .html?dl=0




Not a lot of change here. Focusing mostly on L100 talents, Empowered Seals' 3-seal rotation is more competitive with Sera now, though as we talked about upthread we think Sera is underperforming thanks to imperfect APL optimization. Holy Shield is definitely lagging in single-target DPS still, and will be further behind once we factor in the (up-to) 50% penalty for a tank-swap encounter. I hope to get some time to do AoE simulations this weekend to see how it stacks up in those situations (presumably better).

On the TMI side of things, all three talents are surprisingly well-balanced now. Sera lags a little bit, but again, may be underperforming slightly. I'm actually really pleased with this turnout - none of the talents is underperforming by enough to make them a bad choice under most circumstances.

Most of the rest of our conclusions still hold. SS is dominating, HA/DP/SW are all relatively close, etc. I have to do some work to test L90 talents again now that Holy Prism has been buffed still.

Update: Re-ran the Holy Prism sim. Looks like it's grown a little bit in prominence:

https://www.dropbox.com/s/uyc0f42epq87o ... .html?dl=0



Seems like the logical place to slot it in now is ahead of FW.

Aaaaand the new build contains lots of buffs that force me to re-sim. AS, Cons, HW, SotR all buffed. Time to get cranking.

Just going to link to simulation results for some of these, since I'm running them rapid-fire style.

Basic rotations:
https://www.dropbox.com/s/10dlys74h2w3x ... .html?dl=0

No significant change. CS>J>AS>Cons>HW still wins for both TMI and DPS. Not really surprising - even though AS, Cons, and HW were buffed

Execute Range - HoW:
https://www.dropbox.com/s/987b6hbcnhrkt ... .html?dl=0

Nothing here either. HoW is still top dog for DPS, but the next best position is immediately after our HPG. TMI is sort of pointless here since as long as HoW is behind our holy power generators, the only source of TMI gain is loss of SS uptime during execute range.

Execute Range - FW:
https://www.dropbox.com/s/zqiyptqm2lg6q ... .html?dl=0

This one is a little odd. I'm thinking we may have to move this sim further down in the sequence. FW is clearly a DPS increase as high up as you can push it, but we get the best TMI performance if we put it behind HPG. That said, the difference is tiny in both departments. I think that's because there's still enough empty GCDs at this point that we're just shuffling around SS refreshes. We may see a larger impact by repeating this sim when the rotation is a little more dense, i.e. after L90 and L100 talents are included.

As far as DPS stats are concerned, the recent buffs to all of our abilities boosts the value of mastery significantly. It even starts to catch up to multistrike at the first tier of available stats, though multistrike is still better. This also improves the DPS value of empowered seals by quite a bit, unfortunately. (Unfortunately, because I want empowered seals to suck so bad that they get rid of it.)

I compared Empowered Seals using the DPS rotation and Seraphim paired with sticking with seal of truth. Seraphim wins. These conclusions were made from in game testing, not simc. I'd like to see the sim work and compare it to what I'm seeing in game, though.

L90 talents:
Since FW seemed to be ahead of AS but behind AS+GC, I decided to include that difference in the talent sims.

Execution Sentence:
https://www.dropbox.com/s/jrjox0ltfmkub ... .html?dl=0

Nothing surprising here, ES>all is top dps, followed closely by (CS-AS)>ES>(HoW-HW).

Light's Hammer:
https://www.dropbox.com/s/8kftttegfklmi ... .html?dl=0

Recent LH buffs have moved this up in priority to the same spot as ES. (CS-AS)>LH>(HoW-HW) wins both categories.

Holy Prism:
https://www.dropbox.com/s/d8a8camxfc0n3 ... .html?dl=0

The huge buff to HPr has made a difference, it's now ALSO fitting in behind AS. (CS-AS)>HPr>(HoW-HW) wins DPS and performs equitably on TMI.


So, amusing conclusion: now all the L90 talents fit in the same spot!
New base rotation:

CS>J>AS+>FW>AS>ES>LH>HPr>HoW>Cons>HW

Yeah, multiple beta builds all with changes for us. Nice to see the various talent tiers coming in pretty balanced so far.

I don't mean to rush anything, but just wondering if you think the nightly Sim-C build tonight will be caught up with all these changes? I saw the resolve updates in the change history, but not the newest ability buffs. Plenty of things to re-run all around, and I wasn't sure if I should start tomorrow or wait. I also think I recall you mentioning compiling your own builds from source, but I wasn't sure if that was something anyone could do or only those developing the sim?

Also, good point on SoT with Seraphim for off-tank/burn phases. I'll try simming that once I get my Sim-C updated, but it probably won't be tonight.

It should be. Most of the spells are coded to pull their information directly from the spell data, so as long as the spell data has been pushed to the trunk, any newly-compiled version will reflect the updates. There are a few special cases where I've had to hardcode things, but they're rare (ex: Sacred Shield's SP coefficient).

As long as you know how to download the source code and compile it, you can do so. It's not restricted to project members (though, obviously, making changes and committing those changes is).

More sims.

Sanctified Wrath:
https://www.dropbox.com/s/3x0izigm91p6x ... .html?dl=0

Looks like HW is still a little better than AS+GC, though the margin is extremely thin. There's very little DPS or TMI difference. Still, it looks like the best is CS>J>HW+SW>AS+>FW>AS>L90>HoW>Cons>HW.

Starting with these next two sims, I'm using Divine Protection and GAnK on the pull to try to make sure we're getting stochastic information and not just "what helps most on the pull."

Sacred Shield:
https://www.dropbox.com/s/wp90yaxi13rwq ... .html?dl=0

I added another refresh scheme here, SS8, which refreshes if there's <8 seconds left. Since SS uses HoT mechanics now, this should refresh before the final partial tick without wasting any uptime.

There isn't a lot of difference in these approaches. SS8 tends to give slightly better TMI, but lower DPS. I think the best approach is actually to combine these. One line of SS2>AS can serve as an emergency refresh, while another SS8>Cons line can attempt to refresh at a lower DPS cost whenever it's prudent. We'll have to test this more thoroughly, but the combination of both conditionals should give slightly better results than either one individually.

Selfless Healer:
https://www.dropbox.com/s/0p54wveczsv5y ... .html?dl=0

Not much difference here - clearly a TMI improvement over not using anything, but there's not a lot of gain to be had. It's a toss-up whether you consider Cons>SH or SH>Cons better - I think the ~300 DPS is probably worth more than the small TMI difference (<1k due to rounding).

So I'm going to slot SH in after Consecration (and non-SW Holy Wrath) just like before.

Seraphim:
https://www.dropbox.com/s/z4rl33izk3r1q ... .html?dl=0

I've tweaked things a little here. This is the "Rotation Configurations" text box:

Code: Select all

actions+=/divine_protection,if=buff.seraphim.down
actions+=/seraphim,if=(buff.divine_protection.down&cooldown.divine_protection.remains>0)|(holy_power>=5&(cooldown.crusader_strike.remains<=1.5|cooldown.judgment.remains<=1.5|buff.grand_crusader.react|(cooldown.holy_wrath.remains<=1.5&talent.sanctified_wrath.enabled)))
actions+=/shield_of_the_righteous,if=(holy_power>=5|buff.divine_purpose.react|incoming_damage_1500ms>=health.max*0.3)&(!talent.seraphim.enabled|cooldown.seraphim.remains>5)
CS>J>HW>AS+GC>SS+DR<2>AS>LH>HoW>SS+DR<8>Cons
name=CD_Chaining_DP>Sera

actions+=/seraphim
actions+=/divine_protection,if=buff.seraphim.down&cooldown.seraphim.remains>5
actions+=/shield_of_the_righteous,if=(holy_power>=5|buff.divine_purpose.react|incoming_damage_1500ms>=health.max*0.3)&(!talent.seraphim.enabled|cooldown.seraphim.remains>5)
CS>J>HW>AS+GC>SS+DR<2>AS>LH>HoW>SS+DR<8>Cons
name=CD_Chaining_Sera>DP

actions+=/seraphim
actions+=/divine_protection,if=time<5|(buff.seraphim.down&cooldown.seraphim.remains>5)
actions+=/shield_of_the_righteous,if=(holy_power>=5|buff.divine_purpose.react|incoming_damage_1500ms>=health.max*0.3)&(!talent.seraphim.enabled|cooldown.seraphim.remains>5)
CS>J>HW>AS+GC>SS+DR<2>AS>LH>HoW>SS+DR<8>Cons
name=CD_Chaining_Sera>DP+DPonpull

actions+=/divine_protection
actions+=/seraphim
actions+=/shield_of_the_righteous,if=(holy_power>=5|buff.divine_purpose.react|incoming_damage_1500ms>=health.max*0.3)&(!talent.seraphim.enabled|cooldown.seraphim.remains>5)
CS>J>HW>AS+GC>SS+DR<2>AS>LH>HoW>SS+DR<8>Cons
name=CD_as_available

actions+=/divine_protection
actions+=/shield_of_the_righteous,if=(holy_power>=5|buff.divine_purpose.react|incoming_damage_1500ms>=health.max*0.3)
CS>J>HW>AS+GC>SS+DR<2>AS>LH>HoW>SS+DR<8>Cons
name=No_Sera

As you can see, we still have a "No_Sera" option. "CD_Chaining_DP>Sera" is the old CD_Chaining, but using the new conditional that allows Seraphim to fire if we're about to waste a holy power generator. CD_as_available is unchanged as well. I've also added two new entries. "CD_Chaining_Sera>DP" reverses the priority, giving Seraphim no conditional but restricting Divine Protection to only be used if Seraphim is down and will not be used in the next 5 seconds.

After simming this once, I was surprised to find that this was worse than not using Seraphim at all, which was... odd, to say the least. I guessed that it was probably due to the large spike at the beginning of the fight not having DP coverage, so I created "CD_Chaining_Sera>DP+DPonpull", which just adds an "|time<5" part to the conditional on DP, allowing it to be cast on the pull regardless of Seraphim. As expected, this brought it back into line with the others.

That said... there's virtually no differences between the three strategies in terms of TMI or DPS. Given the circumstances, I'm inclined to go with the "CD_Chaining_Sera>DP+DPonpull" option, simply because the conditionals are a lot simpler and the DPS change is negligible.

Empowered Seals:
I decided not to re-run the entire gamut of sims for EmpS. Instead, I made a few assumptions - namely that the current playoffs sim would be close enough. I updated the rotation base accordingly though, and made a few guesses at where things fit in (example: in most cases, if X>AS before, I changed it to AS+>X>SS2>AS to fit the new structure and keep our HPG prioritized).

Here's the results:
https://www.dropbox.com/s/3fyy8gd3mvgfq ... .html?dl=0



Not a lot has changed. SoI is still our defensive winner, (CS-AS)>TRd>(HoW-TRe) is still our offensive winner. IRTd>Base>IRTe is still our middle-of-the-pack 3-seal performer, giving about 4k DPS at the cost of ~3k TMI.

Random thought: Glyph of Immediate Truth may actually be a slight DPS increase in the Empowered Seals rotation. It depends on how long we stay in Truth, though. The thought being that if we swap to Truth just long enough to Judge, we won't always build up a full 5-stack of Censure and it won't be running the entire time. Thus the short-term buff to the immediate damage component may actually be more important.

Not 100% sure on that of course, just something I thought of while looking at glyphs. At any rate, we should sim it to see.

Talent comparison - 45/75/100:
https://www.dropbox.com/s/ctwwgeji5r7ro ... .html?dl=0

No major change from before. Holy Shield still lagging in single-target DPS. Worth noting that you need to be in "survival" mode with Empowered Seals to really keep up with Seraphim or Holy Shield this time around. Probably an effect of the GAnK/DP addition on the pull filtering out that source of errant spikes.

Talent Comparison - L90:
Just for kicks, I ran this to se how they stack up. It was quite interesting:
https://www.dropbox.com/s/hrb7xdr76bk0d ... .html?dl=0

This obviously doesn't include ES being used for healing, but realistically I don't think anyone does that anyway. So ES is still the damage leader, with LH at the back of the pack, and HPr in the middle somewhere. TMI differences are slight (between LH and HPr anyway).

Another random one: seals.
https://www.dropbox.com/s/ifo7am9ft7rrz ... .html?dl=0


It appears that SoT is a non-trivial DPS increase again (7%-8% or so). This is with SW+Holy Shield, not with Empowered Seals.

AoE Rotations.

Cons position, against 5 targets:
https://www.dropbox.com/s/0axzpt7jgqddk ... .html?dl=0

Boosting Cons up in the priority is obviously advantageous. I'm going with (CS-LH)>Cons>HoW here, because the DPS difference between that and (CS-AS)>Cons>(LH-HoW) is within error, but there's a ~2k TMI advantage for the former.

In fact, this leads me to believe that the proper ordering of the middle of that rotation ought to be:
CS>J>HW>AS+>SS2>AS>LH>HPr>Cons(aoe)>ES>HoW>SS8>Cons>SS

Repeating for 2-4 targets, and excluding the possibilities where Cons>AS+ or higher:
2: https://www.dropbox.com/s/u9sutmwfjox4x ... .html?dl=0
3: https://www.dropbox.com/s/h73te07zfv2v8 ... .html?dl=0
4: https://www.dropbox.com/s/svo5cfbsx1rz5 ... .html?dl=0

It looks like there's no reason to elevate Consecration against 2 targets, but for 3+ it appears that (CS-LH)>Cons>HoW is a solid increase. So we'll elevate Cons against 3 or more targets.