Test Discussion GU45 Celestial AM DPS Update

every power should get the buff. You can keep the buff up with just doing the might part of a power. Just need one "multi-input combo starter"

Yeah, basically the only things that don't get the buff are long DoTs not called DL or CI, Defile, and non-damaging abilities or combos. The full list is at the beginning of the thread.

This pretty much. Prec/Might mods were a dmg lose in every single test I've done. The general sacrifice is around 500-600 DPS.

Can you guys be more specific about the rotations and tests you used?

On the new 3 sparring target setup I've gotten my best results with prec might in reds and mixed prec in the remaining slots. My best rotation so far is alternating RcW with ScH, and clipping DL into the 3 buffs when they're up. Using DL instead of CI has performed better for me due to CI's long cooldown, but I suspect RcW+DL+CI+3 buffs will be better.

On the 8 regular LH sparring targets I get my best results by essentially keeping PcDL up as much as possible.

Not sure why folks are alternating RcW and ScH. The latter is weaker and only 400 msec quicker clipped, so it's a DPS loser against pure RcW spam. Sure looks cool, though.

Because it performed better for me than pure RcW spam. I try not to be biased about rotations; I simply perform different rotations and record the time it takes to KO the sparring targets. I have different sets of modded gear to test pure prec vs prec might vs pure might.

Alternating RcW with ScH performed better for me than pure RcW spam for sure, but was only marginally better than RcW with DL. I'll need to run it a few more times to determine if it's really worth it for me to carry Smite.

Yea ScH is actually losing damage. Generally just start with buffs, then RcW, then PcDL, RcW, RcW, PcDL, repeat. I even put RS back into the loadout.

I'm new enough to DCUO that I made this account to post here. After trying Nature for free and learning that optimal healing and damage both required shape shifting, I subbed for Celestial. Battle Healer. Excellent, right? No.

GU45. Celestial revisited! How fortunate. Except the concept of the power set is ignored in favor of more damage (which was lacking, and the buffs are nice, sure - the battle healer is still not able to battle heal, which is my crux). There was a time where McB converted 100%? I wish I was around for those days, although if I was, I'd probably have unsubbed by now.

25%.

Why do devs insist on over nerfing abilities when the consequence of doing so is rendering the ability next to useless? It happens in every MMO, and even bleeds over into MOBAs. Rather than bring things up to what is considered the best as intended, things are brought down so that only x, y and z can shine (usually by being easier to accomplish the same) while everyone else shares the misery of "toughing it out means more skill" or some such nonsense.

This will probably be written off as unhelpful, and that's fine, since "this GU is about Clestial DPS". But what good is Celestial DPS if Celestial is still broken? The ONE ability that makes the concept viable is left in the graveyard of "post nerf syndrome", and the class will, for the foreseeable future, not be what it should be.

25? You couldn't even take a half measure, just straight leap to 75%. This creates more work for you, as you will, guaranteed, have to revisit Celestial again in the future. Why devs create these op>nerf>useless>revsit cycles for themselves is beyond me when careful steps toward the sweet spot (aka, balance, aka, the whole point of these teams), is safer to ensure the desired result.

Sad. I don't even know what to do with this new subscription. Do I PvP? Absolutely. Do I understand Celestial requires a lot of SP to be effective? Definitely. Was I willing to invest that time? Certainly. To see that this power has been neutered, is living in a shadow of itself, and will continue to do so... not so sure. This game is a lot of fun, but the main reason I want to play it was thrown off a cliff before I even got here.

@Karasawa, it's pure math. Retribution and Smite have identical might damage. Initially, it's single target, but in both cases, it converts to AoE, when used after a multi-input combo. They are both max range. Thus, the only differences are power cost (Retribution being more expensive) and the fact that comboing into cWither can set up the AM buff, whereas cHaunt cannot.

Now, to the combos. Both consist of a precision AoE burst and two trailing ticks of half that burst damage apiece. The aggregate base precision damage of cHaunt is exactly half of cWither. How is the cHaunt combo at half the aggregate damage going to outperform cWither, even with 400 msec clipped duration advantage? It's just not.

@Remander

You may be right, I honestly am only interested in finding effective ways to DPS with Celestial. To that end I run tests and use shiny's combat log analyser as much as possible. I encourage you to at least run the rotations a few times and record the results. What you describe may be entirely true, or you may be missing something because my results are different.

I like your concern, but like the devs said earlier this is a dps only update. No part of celestial healing will be updated. Many of us like to pvp, but as of right now it might be best to go into an early retirement or use a different power for pvp. They might change it in the future, but as of now I dont think there are any plans to change celestial pvp/healer in gu45.

Have you compared the crit chance on each rotation? I have the same rotation going for a few houndred DPS difference due to the simple fact that each rotation results in a different crit chance on the powers. It's not by much but its there and it can favor a rotation who may looks better but was actually worse just because the crits where in favor on this one.

When I done the mod comparisons I tried to replicate the crit chance on the powers as close as possible (same rotation, different mods). The Might bulk is so small, you not gain anything with Prec/Might in red sockets just by looking at it. Had hopes I don't have to remod the red sockets to full Prec but it was already obvious on paper that prec is the key and might nothing worth to be considered, the rotation comparison with equal crits prooved the theory.

I totally understand your goal. I, too, parse runs for comparison. Elusian makes an excellent point about how crits can bias the results. A lot of the testing that folks do is using max gear, mods, and buffs to look for peak DPS. That's fine, but also needs to be interpreted with caution. I do a lot of theory crafting to come up with rotation possibilities, before I test, since I have already grinded out the base values and timings. What I explained above is sound. In theory, there's no way RcW spam should lose to RcW/ScH with the latter having 50% less precision damage and only a 17% time advantage. Doesn't compute.

Isn't RcW spam overwriting 1 tick of prec damage? I'm getting the same results as Karasawa.

Out of curiosity do you know how base numbers are calculated? I see them quoted and used for different powersets on the forums but I'm not sure how they're being determined.

For Might take your gear off and use the equations 1 might =.45% increase in damage to back calculate the base. For prec you need at least some prec for the prec to hit so taking everything off won't work but prec is damage=(weapon dps +(prec/10)*prec constant of the power. You calculate the prec constant by taking off your weapon and having a small amount of prec that you can then back calculate then of course there is variance of around .9 to 1.1 on prec damage so you factor that in to get a base damage range should you have 10 prec(1 dps).

Nvm, got confused with all the logs I have saved.

I also think going prec with:
Anoint, Bene, RcW, PcDL, CI and then DP/WcR/CScWotP/RS something like this

No worries, dude. For those reading, RcW doesn't overwrite.

Yeah, the base values I have are reduced to a precision or might of 1. They need to be refined by averaging more runs, actually. They're close, but it's clear there is some statistical noise still. I just haven't had time, and I wasn't sure if there would be base value changes with the update. Once we have a finished product, I'm going to rerun the numbers for better accuracy.