To cap, or not to cap, that is the question

One of the common questions that come up when dealing with stat weights is how to handle going over a particular break point, soft cap, or hard cap. In most cases, going over the break point is advisable, over/under the soft cap depends on the case, and going over the hard cap is not advisable.

However, when you’re looking at a hard cap for a stat that is your most valuable, then things start to get tricky. Some people will say that going over a hard cap like hit or expertise is a bad thing, and should never be done. While this may be true in some cases, it’s not true for all of them.

The better way to figure out if you should go over the cap is to do some simple algebra: value of capped stat under the cap multiplied by the quantity remaining until capped, compared to the value of the other stat multiplied by the total quantity, and then pick the largest of the two.

Using Elemental as an example, we have rough stat weights of Hit: 4, Haste: 2, Mastery: 1.8, Crit: 1.3. Obviously the question is between Haste and Hit. To work out how far over the hit cap we can go before haste is better, we just take the ratio of the two values, ie: 2:1. While this means that hit is worth twice as much as haste, it also means that when comparing stats for reforging, enchants or gems, you can go over the hit cap by half of that amount before it’s better to use haste instead.

eg: You are 12 hit rating short of the cap. You can either gem for 20 hit, or 20 haste. 12 hit * 4 ep = 48.  20 haste * 2 ep = 40. Hit wins out.

Often you’ll find smaller differences due to re-forging, but the same principle applies.

One site that assumes that over the hit cap is bad is www.askmrrobot.com. The following is how they tell me I should reforge with their default stat weights.

	Raid Buffed	Optimized	Difference
Total Score	43578	43596	+19
Intellect	7603.58	7603.58
Spirit	1333.00	1311.00	-22.00
Mastery	1305.00	1205.00	-100.00
Spell Power	11428.53	11428.53
Spell Hit	472.00	480.00	+8.00
Spell Crit	620.00	620.00
Spell Haste	2492.00	2606.00	+114.00

As you can see, we’re trading a net 14 hit (22 less 8) for 14 haste, and I’m currently 7 rating over the hit cap. Using their “modified” stat weights of Hit: 2.7 and Haste: 1.73, we can see that hit = 7 * 2.7 = 18.9 vs haste = 14 * 1.73 = 24.22, and haste wins. However it is noted that they manually modify their stat weights so that Int > Hit, in order to make gem/enchant selection work correctly.

Using my own stat weights generated via SimCraft using my current gear I obtained different results: Hit: 4.475, Int: 3.8562, Haste: 2.0297, Mastery: 1.8153, and Crit: 1.2781. Doing the same calculation we get 4.471 * 7 = 31.297 vs 28.4158, in favour of Hit again. To make Mr Robot do what I’m expecting, I need to over-ride their default values, and set the hit cap to 17.1%. The new results match perfectly with my gear as it’s set up (I use www.wowreforge.com to manage my reforging with a minimum hit cap).

In short, going slightly over a cap can be a good thing, as long as you consider it in context.

Doing Things Better: Updating Stat Weights

A number of sites run stat weight based systems (wowreforge, askmrrobot, wowhead, for example), but all of these sites either rely on their users to provide them with the stat weight values, or do their own work to produce those values. This means that you either need to manually update the preset EP values every time you use a site, or assume that the presets are accurate, which isn’t always the case. I know from my own experience with theorycrafting that the first, second or even third look at how things work may be incorrect as new patch notes and hotfixes get assimilated into our current understanding.

With this in mind, wouldn’t it be easier if we had a better way to connect those doing the early theorycraft/guide work with sites that use their findings? That way site administrators don’t have to manually update their stat weights, and users get more accurate information.

To do this I’m suggesting the creation of a standardized XML format for defining stat weights, which can then be distributed to item sites either by uploading them to specific community/theorycrafting sites, submitting them directly, or via some other means. They can even be turned into import strings for in-game addons like Pawn. The XML files would contain information on the class, spec, tier level, all possible stats with a use/don’t use flag, stat weights & stat caps. As this progresses we would be able to build multiple EP stat weight sets for those just starting out on 5mans, vs those working on the latest heroic raid, and adjust weights for specific breakpoints (eg: haste over X%, hit over cap being zero).

Of course, this will require a wider community participation than just myself, so if you are interested in this, leave a message here. If there is enough interest to move forward with this, I will create a sub-forum on TotemSpot to start discussing it.

VPLC proc chances

Working out double crits to reduce overall crit frequency. Warning: heavily math based

Pair of casts: Crit^2

Normal cast freq: crit^2/((lbcast * lbcount + shockcast * shockcount)/(lbcount + shockcount))

LvB + normal cast: Crit

LvB Surge offset:  0.2/fs tick time

LvB double freq with surge: Crit*(1+surge offset)/9 (9 being slightly longer than the base internal cd to allow for not casting as fast as we’d like)

Can then add both freqs together to get a cast based charge frequency (C Freq). Can then add on a value for FS freq based on crit/fsticktime less C Freq to give a dot based frequency which will be greater or equal to zero (D Freq). Subtraction occurs due to the fact that a FS crit will only have an impact if there isn’t a cast based crit (while in reality it can happen in either direction, this makes things simpler).

Add C Freq and D Freq together to get an overall proc frequency per second.

Assumptions made: ignoring fulmination as a separate spell, but it would double the crit chance for ES and increase proc chances. Add in later for a more complex model.

4.1 AoE Changes

Or “How I Learned To Start Worrying And Not Love The AoE”.

To begin, here’s a post I recently made over at EJ.

So, after last nights Heroic Maloriak attempts, I have come to the following conclusions about Elemental AoE damage.

1. Earthquake is virtually useless, and probably only be used for specific heroic encounters via a secondary spec
2. Fire Nova is fairly useless too.
3. Tab spamming Flame Shock to get more Fire Nova bursts out doesn’t make much of a difference.

I compared tab FS with FN spam, single target + CL with FN spam, and my usual single target + CL rotation on the Dark Phase, and there was virtually no difference in damage output, making allowances for things that happened in each attempt (~2.8m for the first, with two dark and nearly going OOM plus aboms, ~1.5m with one dark and aboms, ~1.25m for one dark without aboms respectively).

In short, I’m going to continue not using FN, and I’m going to have a second ele spec for the odd occasions I actually need Earthquake (which is virtually never). If I use both, my AoE rotation will be 7 abilities with differing cooldowns (3, 4, two on a shared 6, 8, 10) vs a much more manageable 5 (3, two on a shared 6, 8). I’m not sure how complicated other caster AoE is, but surely it can’t be as complex as this with such little gain over single target damage.

Expanding on this, I’ve loaded up the three attempts in RaidBot’s CompareBot to have a look at things. Unsurprisingly, damage output for Fire Nova and Earthquake is surprisingly low, less than 1000 dps over the whole encounter. In fact, across all attempts that night, Fire Nova did less damage than the direct damage explosion from Living Bomb (on Vile Swills), and Earthquake was much, much less.

At this stage, the AoE changes for Shaman appear to be worthless.

Doing some basic math, and getting these values unbuffed, Magma does 456 dps per target, Fire Nova 521 assuming you get it off every 4 seconds, and Earthquake does 869. The problem is that the overall value doesn’t get up past 15k dps until you get to 5 FS targets/5 AoE targets, 3 & 6 or 2 & 7. Sure we can still use CL, LB and LvB between Novas, but it then becomes an awkward mishmash of rotation types and cooldowns get pushed back (never mind the fact that having to run in to drop Magma goes against every ranged AoE mechanic except for Demo Locks). In addition, the scaling of each ability is mostly restricted to Int and Crit, with EQ getting some small benefit from Haste. Still, Magma gets an extra 8 dps per 100 Int, Fire Nova 17, and Earthquake a paltry 6. It’s a plausible statement to say that I could do more damage single target than I can with a double flame shock/fire nova AoE setup on 7 targets.

Generally speaking, any AoE damage is good AoE damage, but in terms of output and efficiency, I’d say that Elemental is better sticking to single target damage with some on-CD Chain Lightning use.

The only way forward, to be perfectly honest, is going to be the removal of Magma Totem. Enhancement don’t use it, thanks to their Searing Totem interactions, and the action of running in, dropping a totem, and then running out again is counter productive (a big reason why Fire Nova was de-coupled from totems). Once it is gone, improvements can be made to the rest of the AoE rotation. These are my suggestions

1. Restrict Fire Nova to the current target affected by Flame Shock. This means no more having to worry about multi-dot AoE situations
2. Add a new AoE rotation filler, be it a modified CL, a Forked Lightning ability, or some other channelled ability with no cooldown. Otherwise we end up in the situation of having to run our single target rotation in amongst the multi-target rotation, and that’s already complicated enough on it’s own.

Once these play style changes are made, then a damage adjustment can be made, so that a multi-target rotation results in better damage output just like the other casters.

Lava Surge, Part Two

In the middle of the comments on a post of Lava Burst by Charles (aka Zamir) over at Planet of the Hats, there’s a little gem by Santux. It reads as follows:

What do you think about a Lock and Load type treatment with Lava Surge? *Flame shock ticks have a 30% chance to make your next Lava Burst not trigger its cooldown* Would definately solve the problem of the 8 second window having meaningless Surge procs.

A Preliminary Post on Lava Surge

Been working a bit tonight on updating my Lava Surge spreadsheet that I posted on EJ a few weeks back.

I need to write up some stuff to do with it, but here’s a graph of the DPS gain. It’s important to know I made a few assumptions on the math, and they are listed in order of “priority”.

1. If a surge proc occurs within a single LB casts time of the cooldown resetting anyway, it is “wasted”
2. If a proc occurs you are able to take advantage of it straight away
3. Procs will occur exactly 10%, 20%, and 30% of the time

The (Elemental) Cataclysm and You! Part Two: Talent Builds

Expanding on my previous entry about talents, I’m now going to look at what we’ll do with them. I’ll look at a “standard” elemental build, a Dual Wield build, and a Searing Flames build.

Before I start I should point out that the Tier 2 Enhancement talents are due for a re-work. This has been confirmed by a few Blizzard poster comments about Enhancing Totems being turned into a self-buff when totems are out, and Ghost Wolf being usable indoors (whether this means it will be an instant cast by default or not remains to be seen).

Modelling Searing Flame

Hello readers!

Since I’m trying to be a bit more… approachable & social with this new blog, I thought I’d run you through how I do some of the things I do. The first of which is a little run-through of how I ended up modelling the Searing Flame dot effect.