Anyone else feel that UM PC is weaker ingame than in sims?

[Valounette]

A couple of things to consider that gives SN an advantage over UM on actual boss fights.

1. On actual fights you actually end up interrupting quite a lot of spell casts, either by (accidental) movement or because you realize it was the wrong choice to cast AB or one of the many other reasons you could think off. The total UM damage you can do is limited by the amount of AB casts you can get off in a fight (which is reduced by interrupts) and your total SN damage is mostly limited by it's cooldown. Especially since it has 2 charges you're almost always gonna get the maximum amount of damage out of your SN.

I really don't think this is the issue. I don't mean to sound like I have a massive ego but I believe I can play it to near perfection; and the odd single cast break won't account for many thousands of DPS in a fight

[quote[2. SN is instant cast and can be used while moving.[/quote]

This is kinda irrelevant. Even on low movement fights like Reaver and Fel lord, or combat dummies, I had noticed the damage discrepancy was much larger than it should apparently be.

3. (speculation) in real life situations it's quite likely that you delay your SN's for longer then robomage would because your play probably isn't as optimized. Hanging on to some charges of SN gives you the ability to use them to fish for more missiles procs when you get PoF procs. I've been using my SN's for this purpose quite a lot and I strongly believe it's a big DPS increase. (to bank 1 charge in order to fish more AM during PoF) maybe someone smarter then me could sim this?

Nope. The only time supernova is ever delayed is the first one, which gets used soonafter on the crystal. My supernova never waits on 2 charges, and 2 charges get blown into every crystal. If anything, I'm certain that my play regarding Supernova is more optimal than Robomage; with using Supernova during trinket procs/high Incanters stacks, as well as keeping other factors like mana/number of arcane missiles in mind.

I'm thinking UM is just so susceptible to RNG. Getting no procs on PC, and less Arcane missiles (due to no double supernova hits and mana recovery), burn phases can be absurdly short lasting

[Methusula]

You need to post logs. Telling people you believe you play near flawlessly is not going to help your case. If you want to post anecdotal fluff, this may not be the place to do it. Post logs so we can dissect what you are doing, what UM damage you should expect with what you are doing, and what you are actually seeing. Beyond that, this discussion is pretty pointless.

[Valounette]

You need to post logs. Telling people you believe you play near flawlessly is not going to help your case. If you want to post anecdotal fluff, this may not be the place to do it. Post logs so we can dissect what you are doing, what UM damage you should expect with what you are doing, and what you are actually seeing. Beyond that, this discussion is pretty pointless.

I get that you seem annoyed over someone saying ''blabla I'm not making mistakes'' and demand proof - however you're missing the point, I was specifically asking for other peoples experiences/anecdotes, playing in their own raid settings. In my experience, the friends I have in-game have also found UM Arcane very lackluster, and I've been unimpressed myself. I came asking if a wider audience have had different results. So far, no one has said this; and a large amount of top logs also use Supernova.

Furthermore, I am not making this 'claim' (or even making a claim at all, I'm trying to gather a consensus opinion) out of nothing - there are at least two other threads which have drew this observation.

Why am I looking for anecdotes? Logs aren't incredibly helpful in this situation, just as you'll see in other similar threads asking about this and posting random comparisons, which are entirely useless to analyse due to RNG. I could post two... But you would be dissecting RNG.

I'm not really looking for guesswork at possible l2p issues (which would honestly be pretty disastrous play with simple stuff which shouldn't happen with basic use of Icefloes or knowledge of fights - things that in my mind shouldn't even be taken into consideration) - hence I dismissed those.

Still looking for people who will tell me ''no val, I'm actually finding UM PC consistently outperforms SN PC by roughly 700DPS, just like my sims show!'' (or more realistically due to RNG, at least landing in the ballpark of SN DPS, rather than being oddly a lot lower)

[Caimion]

I tried to further test by hitting dummies for about 90 minutes, alternating between pulls of UM PC and SN PC, 7:30minute pulls. Every single parse, SN PC did significantly better.

Different specs perform better depending on fight length. I wouldn't use a methodology that tests an exact 7:30 pull length over and over; sims vary fight length to account for this. You may have found a pull length that happens to favour SM over UM.

I don't know why or how the APL would overvalue Unstable magic... am I crazy? I saw another post saying they had the same problem.

SimC shows Unstable Magic on a Patchwerk sim as ~5.8% of total damage. Taking a brief look at WCL for a 'Patchwerk' fight (Mythic Iron Reaver was the closest I could think of), the parses I found with UM tended to be between 4.5-6% in total damage. So I don't believe that SimC is overvaluing Unstable Magic, although I agree with the other posters that real life situations are likely to affect UM more than SN and the WCL evidence anecdotally supports that.

[Mortiferus]

I'm not sure anecdotes help us much here, either.

One side will say simulations are devalued because they don't accurately reflect realistic scenarios. The other side will say anecdotes are devalued because of RNG. Neither side is right or wrong.

I can only go off the logs that I believe to be yours, since you haven't shared them. The mage in those logs is playing very well indeed, but doesn't have the class trinket. That has a significant impact on UM DPS, even in the sims.

Furthermore, for obvious reasons, the sims show that UM has a much higher DPS range. So, if you're arguing that, anecdotally, SN is higher (or, conversely, that UM is lower), I'd be inclined to indeed argue it's just variance.

All of that said, for me personally, I don't have the class trinket yet, and I'm simming UM as a DPS loss. With the class trinket, it sims as a DPS gain. Until I have the trinket, I can't tell you what my experiences are, but I wouldn't want to anyway—having only 2–3 weeks of logs is not enough data for me to make any conclusions, personally.

However, let's throw all that aside and say you're right—SN is better in practice than UM. The truth is, even if a Mage is playing perfectly, as you're so quick to attribute to yourself, playing perfectly in actual fights is not the same game as robomage is playing in the APLs. IIRC, Frost mages were simming way higher in HM than they were performing in practice, even at the highest levels. Playing perfectly doesn't mean you're mirroring robomage, so yes, it's entirely possible that SN is better in practice.

But 2 weeks isn't enough time for any of us to jump to any conclusions, especially since we're still gearing, learning the fights, etc. Most of us, anyway—I don't profess to start a new tier as a perfect mage.

[Komma]

In the interest of steering this conversation in a more constructive direction, I'd like everyone to hold their tongue and take a step backwards.

There are two possible situations here:
Case 1: A combination of UM being overvalued in sims and/or SN being undervalued. This would indicate an error in SimulationCraft that needs to fixed. To do this, we need to doublecheck and identify whether simulations are modelling UM/SN properly, and whether UM/SN are being used correctly. This requires posting simulation results, critiquing or accuracy, and giving suggestions on how to increase the accuracy of how things work.
Case 2: Players not using UM as well as they use SN. This means that we need to improve our execution or theory, in terms of gameplay. To do this, we would need logs from players, so that we can identify how to better use UM (which would ultimately boil down to how we can get higher AB damage.

The main point is: This thread cannot be productive unless players either link combat logs or simulation results, along with their analysis and reasoning, either to fix sims, or to improve player rotations. I ask everyone to stop simply discussing anecdotes, and post evidence for discussion - either simulation outputs with analysis, or logs that allow for gameplay critique.

All posts from now on that do not bring or discuss evidence will be removed.

[Chuk]

So, i was @ the dummy for a hour before my raid, here's what i've found.

For my gear, UM sims higher than SN fully raid buffed. Since i cannot easily sim that in dummies, i've decided to test it myself using only selfbuffs no lust / no consumables.

I'm simming on both situations about 40k DPS. But on dummies, UM is only reaching up to 36k, while SN is getting constant 40k DPS.

Here's some logs that you can see.

https://www.warcraftlogs.com/reports/F1 ... /#fight=37" target="_blank
https://www.warcraftlogs.com/reports/F1 ... /#fight=39" target="_blank
https://www.warcraftlogs.com/reports/F1 ... /#fight=40" target="_blank

I've tried to get close to 6 minutes on all tries. I'll keep logging after the raid to see if i can find any new information.

[adrift]

Something else to consider is the actual value of SN on a lot of the HFC fights. For Mythic:

Assault: Burst when you need it, can also use it to interrupt engineers and pre-meta fel casters.
Reaver: Bombs. They are #1 dps priority and often spawn close enough to cleave with SN. Its also great for bursting the quick fuse ones.
Kormok: Hands
Kilrog: Can CC the Salivating adds, also good for bursting bloods if they get close. Amazing if you go down as well.
Council: This one I could see UM, but you should be playing frost anyway.
Gorefiend: SN knockup works on the downstairs adds, good for bursting adds or even Remnants if the person is about to be digested.
Iskar: Not really sure about this one, I could see taking UM here. I feel like the SN burst for the priority adds is probably better though
Socrethar: I could see UM here, depends on your strat honestly. If you are never touching spirits then UM but its nice having the knockup in case they get close.
Fel Lord: The one and only patchwork fight. So this one should be UM, at least according to current sims.
Tyrant: Play frost.
Manno: 100% SN. As a bonus it interrupts the imps cast too.
Arch: Probably frost. Maybe UM as arcane, we don't know enough about Mythic.

So regardless of the sim accuracy, you will probably want to be running SN on most of the bosses anyway.

[Wilderness]

I don't think that how often you use UM is all that important relative to if its being modeled correctly so that SimC reflects what we see in game, and thus we can make the best informed decision on our talents per fight. I also think you're vastly overestimating the usefulness of SN's burst on most of those smaller adds on many of the fights mentioned.

[adrift]

I'm not trying to make it out like SN is gamebreaking on those fights or something, but when the talents are as close as they are in pure patchwork sims it doesn't take much to push SN over the top for actual in-game usage. That was more my point.

People will look at Fel Reaver and think "single target" then spec UM because its the highest number in the first post. The way it plays out in game the ability to have your AoE talent on demand on that kind of fight will most likely result in higher average DPS over a large sample size. This is of course assuming the single target sims remain comparable at your current gear level. For me currently (715 + 4pc + no class trinket) the difference between full haste enchants/gems with UM/PC and full mastery enchants/gems with SN/PC is only around 350 DPS.

Edit: To try and keep this more relevant to the new thread rules, the current SN usage during conserve could probably be optimized slightly to account for PoF. I believe it should be a minor DPS gain to have it use SN when Mark of Doom is present regardless of mana % and AC=4 if AM stacks = 0.

[Wilderness]

Sure, I guess if they are that close you might make that tradeoff. Its surprising to me that your dps is that close with both options though with your gear. For me (710, 4 piece, class trinket) it’s a 4k difference in UM/PC with full haste gems/enchants and SN/PC with mastery.

[Frosted]

Due to feedback like this thread, we've investigated and found that Unstable Magic was inappropriately procing Arcane Missiles. This was causing UM profiles by overperform by about ~2.5% on average in the sim.

The fix should be in the next nightly build. Thanks for keeping the sim honest

[Zifnab]

Oh my o.O That seems like a significant change in the sims then. Due to 2/4T18 are the haste values still better to gear for than mastery?

I might have to rebuild locally again so I can check :< I just switched all my gems/enchants to haste yesterday I'd hate for that to have been a bad idea and waste of gold...

[Frosted]

Oh my o.O That seems like a significant change in the sims then. Due to 2/4T18 are the haste values still better to gear for than mastery?

I might have to rebuild locally again so I can check :< I just switched all my gems/enchants to haste yesterday I'd hate for that to have been a bad idea and waste of gold...

It does have an effect, yeah.

[Valounette]

I'm not sure anecdotes help us much here, either.

One side will say simulations are devalued because they don't accurately reflect realistic scenarios. The other side will say anecdotes are devalued because of RNG. Neither side is right or wrong.

I can only go off the logs that I believe to be yours, since you haven't shared them. The mage in those logs is playing very well indeed, but doesn't have the class trinket. That has a significant impact on UM DPS, even in the sims.

Furthermore, for obvious reasons, the sims show that UM has a much higher DPS range. So, if you're arguing that, anecdotally, SN is higher (or, conversely, that UM is lower), I'd be inclined to indeed argue it's just variance.

All of that said, for me personally, I don't have the class trinket yet, and I'm simming UM as a DPS loss. With the class trinket, it sims as a DPS gain. Until I have the trinket, I can't tell you what my experiences are, but I wouldn't want to anyway—having only 2–3 weeks of logs is not enough data for me to make any conclusions, personally.

However, let's throw all that aside and say you're right—SN is better in practice than UM. The truth is, even if a Mage is playing perfectly, as you're so quick to attribute to yourself, playing perfectly in actual fights is not the same game as robomage is playing in the APLs. IIRC, Frost mages were simming way higher in HM than they were performing in practice, even at the highest levels. Playing perfectly doesn't mean you're mirroring robomage, so yes, it's entirely possible that SN is better in practice.

But 2 weeks isn't enough time for any of us to jump to any conclusions, especially since we're still gearing, learning the fights, etc. Most of us, anyway—I don't profess to start a new tier as a perfect mage.

Ok, this thread has been solved now, but I still want to answer you, since I didn't get a chance to, and I feel that it's worth pointing out that the value of SN as an instant cast is being grossly overvalued in some of the prior posts. At least in this tier of content.

Firstly:
I didn't mean to state that I play perfectly. That would be entirely irrational. What I meant is, I am very confident that the specific factors/'counter claims' being discussed (the 'value' of an extra instant cast every 25 seconds, concerning movement) are trivial when considering a comparison between the two talents in this raid tier.

Let me elaborate:

In the current tier, I'm never finding myself screwed by movement when dropping a crystal. I never think 'damn, I wish I took Supernova for more movement abilities on this fight'. Fights like Fel Lord (not soaking crystals), Reaver, Gorefiend etc demand little more than the occasional Ice floes charge and/or Alter Time 'warp' (as an aside: Alter time is HUGE in HFC, use it to get back from dumping debuffs on Kilrog blood add debuff, Gorefiend void zone thingy, Fel lord Seeds, Iskar Chakram. We don't really have anything like Twin Ogron fire, or Hanz and Franz stampers, for example, to attribute meaningful value to the extra instant casts that Supernova provides (under the assumption that one positions properly and makes good use of Ice Floes!). There isn't really a fight which demands continuous movement that can't be accounted for by Ice floes.

[Note, since we're only 2 weeks into the tier, the above is referencing Heroic. Mythic mechanic additions might make that irrelevant; but my experiences in comparing the talents have been from heroic progression!]

Due to feedback like this thread, we've investigated and found that Unstable Magic was inappropriately procing Arcane Missiles. This was causing UM profiles by overperform by about ~2.5% on average in the sim.

The fix should be in the next nightly build. Thanks for keeping the sim honest

You're brilliant. This, combined with the higher variance in RNG with UM certainly makes things a lot clearer.

Lastly @ those who have shown distaste to an arguably unconstructive discussion:

Gathering opinion and anecdotes, with reasonable suspicion that something is wrong, will lead to an increased interest in the subject area. This in turn leads to data and results and potential for change.

[Lak]

Due to feedback like this thread, we've investigated and found that Unstable Magic was inappropriately procing Arcane Missiles. This was causing UM profiles by overperform by about ~2.5% on average in the sim.

So until SC is updated I'll divide my results using UM by 1,025? (I'm on OS X so can't use the nightly builds).
For my character that puts Frost (UM) on ST top, by a small margin over Arcane (SN), Frost (FBomb) third and Arcane (UM) forth. I'm 704 with 2T18 and no BRF trinkets (goren+crucible).

[Mortiferus]

Due to feedback like this thread, we've investigated and found that Unstable Magic was inappropriately procing Arcane Missiles. This was causing UM profiles by overperform by about ~2.5% on average in the sim.

So until SC is updated I'll divide my results using UM by 1,025? (I'm on OS X so can't use the nightly builds).
For my character that puts Frost (UM) on ST top, by a small margin over Arcane (SN), Frost (FBomb) third and Arcane (UM) forth. I'm 704 with 2T18 and no BRF trinkets (goren+crucible).

I believe Frost was referencing an average, meaning that certain UM profiles are likely more and others less than 2.5%. So while not unreasonable, I wouldn't take the approach you did to draw any definitive conclusions.

It's possible to compile the nightly build from source and run it without the GUI, even on a Mac (I'm a fellow Mac user, so I can relate)—that's probably ridiculous, however. Instead, if you'd like, post or message a link to your Armory profile and I can run the sims for you.

[Zifnab]

I did build from source on my Mac, and it wasn't very hard.... But since I'm already a developer and already had most of the required tools installed on my system, my barrier for entry was probably lower than the average user.

I'm on my PC now, so don't have the results handy - but I did find that SN with arcane is now outperforming UM. Haste still appears to be good to gear for, and with the reforge plot I still saw the noticeable jump at 20% that Frosted has mentioned in the past.

I think Frosted's conclusions about haste are still valid, but at the same time I think SN is the talent to use with PC.

The sims always say to grab Rune of Power, but with the amount of movement I never sim RoP... because I almost never take it. In any case, haste with SN was ahead regardless of whether you use RoP, or IF.

I wish I'd saved a screenshot of these results, I'm sure frosted or someone else will get to posting the new results before long (if they haven't already, I haven't checked the main general discussion forum in a bit).

[martyr]

I did build from source on my Mac, and it wasn't very hard.... But since I'm already a developer and already had most of the required tools installed on my system, my barrier for entry was probably lower than the average user.

I'm on my PC now, so don't have the results handy - but I did find that SN with arcane is now outperforming UM. Haste still appears to be good to gear for, and with the reforge plot I still saw the noticeable jump at 20% that Frosted has mentioned in the past.

I think Frosted's conclusions about haste are still valid, but at the same time I think SN is the talent to use with PC.

The sims always say to grab Rune of Power, but with the amount of movement I never sim RoP... because I almost never take it. In any case, haste with SN was ahead regardless of whether you use RoP, or IF.

I wish I'd saved a screenshot of these results, I'm sure frosted or someone else will get to posting the new results before long (if they haven't already, I haven't checked the main general discussion forum in a bit).

Also have to consider with and without Tome. With Tome you aren't using SN on PC anyway. So the value then shifts to either UM or NT. NT scales really well with Haste, so UM / NT should bounce back and forth as the best based on Haste once you have the class trinket.

[indomitus]

Due to feedback like this thread, we've investigated and found that Unstable Magic was inappropriately procing Arcane Missiles. This was causing UM profiles by overperform by about ~2.5% on average in the sim.

The fix should be in the next nightly build. Thanks for keeping the sim honest

Out of curiosity, is there any possibility the same thing is happening for UM in frosts sims proccing FoF procs? (I realize this isn't the frost boards but I felt it was more appropriate within the topic here.) Im not asking based on any personal observations as of yet.