Aside from the fact that you can't macro things like that anymore afaik. Clearcasting proc's when the spell hits. Clearcasting is consumed when you spend the mana. What you are seeing most likely is you are getting lag and not canceling scorch before the cast ends.

Another possibility is the crit on clearcast talent makes a secondary clearcast sound which can be confusing because you hear the sound but you have no clearcast debuff at the time that the sound happens.

I'm not sure I understand the value in comparing and trying to rank caster dps where the actual casting bonuses (dmg, crit, hit) are different among each caster. The rationale for not using a uniform "+800dmg, +16% crit, +12 hit" (just as an example) to still get a fair baseline eludes me.

Most of the time it looks like the graph is including easily obtained talents for the +hit differences and perhaps some crit differences. Dots can't crit so stacking crit on a affliction warlock probably wouldn't be your best bet.

For each class there is a difference "optimal" spread of +damage vs +hit vs +crit. I think the chart does a decent representation on how different classes might spread out their stats.

If end-game gear forced us all into the "+800dmg, +16% crit, +12%hit" then there would be no reason to bring anything but Destruction Warlocks and Fire Mages.

Fortunately, we have (some) choice in what gear to wear. So how do you compare two classes that REALLY want to wear different outfits? I chose to keep the "ilevels" of the equipment the same, and then just pick the best (IMO) combination of stats for each class.

You cant really, that was my point. :) Without a common platform you're not getting an accurate picture of how each compares to each other.


But perhaps an absolute comparison/ranking is not the purpose of this list/comparison and I'm just misinterpreting the intention(?)...

Equivalent ilevel is a common platform(in Blizzard's view). That's the entire point of the concept of ilevel in the first place.

It's the ONLY meaningful baseline we have at the moment. Dots don't crit so why would you spend ilvl on crit gear for affliction warlocks? etc...... What set of stats would you use to make a good comparison? You can not use the same damage/crit/hit on each class because different classes/builds have different values on dam/hit/crit.

Your stats should add up to the same ilvl but you should be able to add in hit/crit bonuses from talents provided you assume certain builds with certain gear combos. Comparing 800 damage 0% crit affliction warlock to a 800 damage 0% crit mage would be misleading almost like comparing a rogue with str to a warrior with str. We value different stats differently even as casters.

I disagree. We have things like wowhead and thottbot that can give us actual gear we can acquire. It would probably be a lot better to do a comparison using stats coming from an actual set of gear (that can actually be obtained) that is appropriate for the spec. In particular, it's very easy to use wowhead to filter gear by item level and its ability to be used by a class. (They have a great "Usable by" filter for that...)

It wouldn't be that hard to put together a Tier 5 (DPS set when applicable) set + similar quality non-set pieces. It's probably a half an hour of effort per major spec, and it nets far more accurate data. Gear is at least as important as spec to this question. It seems silly to use specific specs but use hypothetical gear that doesn't exist. The gear choices need not even be perfectly optimal, just competitive and of consistent ilvl.

Doing this would also give you realistic values for stats like INT, SPI and others which aren't things you can just set and forget. Just as some classes prefer crit to dmg or vice versa, every class will also have different primary stat priorities. An Arcane Mage is probably going to value INT more highly than a shadow priest or an affliction lock. A DPS holy priest is going to consider stacking spirit for spiritual guidance. A warlock will probably spend more of his item budget on stamina since INT is less useful to him. The assumption that some fixed fraction of the total item budget is always spent on dmg/crit is a flawed one.

At least to my mind, the most meaningful baseline for gear is the Tiered sets. These are, ultimately, what a whole lot of people are going to be wearing, and it seems silly to try to invent a gear baseline of consistent item level for us (and potentially do it in a stupid manner) when Blizzard has already done it for us. Admittedly, Blizzard may have done it in a stupid manner, but it's a stupid manner that matters.

Well, that's really not the point of this simulation. Both the "extreme" and "less extreme" scenarios he provides are basing their data on extremely ideal situations, which ignores proper class balance in what would likely be a successful TBC raid. All the classes are assumed to have every buff and debuff imaginable at their disposal to get what is, in theory, the absolute highest DPS threshold. Of course in reality these situations are few and far between, some are barely even remotely plausible. The purpose is not to be an accurate, in game, insert-your-spec-and-gear-here DPS calculator. It's itended to show what the classes are capable of given a set of "perfect" variables.

This would be a very interesting simulation........... but for the Shadow Priest community I'm sure it will simply validate something we already know: While Tier4/Tier5 is a marked improvement over past sets in terms of itemization, we will STILL end up using off-set items to maximize our DPS. Well..... at least any off-set items we're allowed to get anyway. :)

Well that's not really a problem. Just produce a different gear profile with similar ilvl items, and run that as yet another set of scenarios (tiered gear v. optimized gear with an average ilvl of X). Voila.

It's actually not so different for mages. Many many a mage has preferred a so-called "Clown Suit" to set pieces. (Excluding Netherwind, that's generally how I've felt about gear, actually) Pick a reasonable ilvl range (Averaging around 125 for Tier 5, say) and actually build sets with real gear. The key is that you eliminate the "Well nobody would have have gear like that" criticism. Don't get me wrong -- Your work so far has been extremely valuable; I just see your gear assumptions as one of the few serious flaws in your approach.

Adding this sort of ability to your work also gives you two interesting insights:
- First you can see how itemization affects scaling. Having hard numbers for the belief that the Tiered sets are inferior to Clown Suits is valuable.
- You can measure the effect adding the latest and greatest overpowered item has. By comparing a Tier N mage with items of mediocrity versus a Tier N mage with a (hypothetical) Talisman of Godliness you can evaluate the theoretical impact of the item in a valid context.

Anyhow, good work so far. I'm definitely interested to see where you go from here.

(EDIT: Eliminated some of my rambling.)

You will. You'll want to handle these sorts of requests with an iron-fist. Basically tell people if they want anything crazy then they're free to borrow your code and run it themselves (since you so generously provide it). Beyond a tier based set and maybe a set for the major spec types (i.e. holy v. shadow, fire v. arcane v. frost, destro v. affliction) you really don't want to go too far down that rabbit hole.

I'm sure you get a ton of talent based requests. I know I'm tempted to ask you for a 41/20 Arcane Blast/Scorch rotation simulation. But I won't (because I kinda want to play with it myself. :-)

Playing the game should come first. Save the heavyweight theorycrafting for when you hit 70 and get bored with instance grinding. :-)

Again, thanks for your work. I know I've found it informative.

(EDIT: Do not produce customer deliverables to Fortune 500 companies and theorycraft at the same time. Makes for sloppy posts and funny typos...)

I haven't looked closely at your code you are using for the simulation, but... Don't you basically take 3 numbers to represent gear? Crit, hit and +damage? You can easilly just tell people add up the values your gear gives you and plug them in.

As an unrelated aside... The methods used so far seem more in keeping with a justification to spec shadow and still raid as a priest than it does to do with an objective analysis of comparitive dps or even the effects of gear and spec on various classes dps. Using real gear will help with that, one of the big problems is that gear is often itemized quite poorly compared to optimal. I did some work on that when the origional Bonescythe set was put out and came up with a set that while the same ilvl had nearly double the +atk with more crit, more hit and sta to boot.

The methods seem in keeping with that justification because that was one of the primary motivators. :)

The audience I was trying to reach was my own class. There was too much gloom and doom on the WoW boards.

My goal was to show that Shadow has a real possibility of being competitive in the end game.

Rogues are basically #1 (& #2 in my group) for all fights in Karazhan. Rogue DPS is kinda shitty on High King Maulgar (1 add they are doing tons of kicks, mage guy they can't even touch, and High King Maulgar has a little surprise that makes rogue DPS on him hard to do at times).