The War Within Crafted Gear
The majority of Best-in-Slot lists that you see out there, including the ones created by our gear optimizer, will include two crafted items with embellishments. We are going to do a deep dive into the design and balance of these effects.
Current state of embellishments
Cutting to the chase: at current max item levels, you can expect around a best-case 1.5% increase to your damage/throughput from two crafted items with the better embellishments. This factors in the estimated value of the embellishment effects, the fact that you can tailor the stats on most crafted items, and that crafted items cap out a few item levels below max-upgraded gear.
1.5% is a pretty annoying gain... it's just enough that you don't want to ignore it, but just small enough that it doesn't feel like anything when you get it.
Furthermore, there are still what we consider bugs with many of the embellishments, the balance between the different embellishments is way off, and the scaling as we get into later tiers is very inconsistent. We will point these out and propose solutions.
How could they be improved?
World of Warcraft is a game where you get a lot of small upgrades that accumulate to a significant increase in your performance. There is nothing wrong with that design per se... but because each individual effect is small, it is very difficult for a player to tell if any particular gear change made a difference in-game.
Thus, we have (super awesome) gear optimizers, combat logs, and simulators. Players really have no alternative if they want to make informed decisions.
Crafted embellishments are a good example of where we think Blizzard could help players out a bit more. Firstly, being more consistent about implementations would go a long way. For example, if we knew that every damage effect would always be increased by crit, haste, and versatility (unless stated otherwise), we could make more informed decisions. Instead it's a guessing game on an effect that only triggers once or twice per minute... even if a player did want to spend enough time getting a statistically significant sample to test that, it would be intensely boring to do so.
Secondly, a little more effort to balance effects would go a long way. If we could be reasonably confident that every crafted embellishment is at least in the same ballpark, then we could make a choice based on which effect sounds most synergistic with a particular spec and style. Instead we have weird outliers where the intent of the low value isn't really clear (or just a mistake).
And lastly... better tooltips! We know this will never happen, but it is worth bringing up. Every special effect tooltip should clearly indicate the trigger rate, duration, and magnitude of the effect. If an item has non-standard behavior (like it can't critical strike or doesn't benefit from haste when normally it would), it should be noted in the tooltip.
The rest of this article will be a detailed analysis of the available embellishments and some specific proposals for improvements.
Available Embellishments
The table below has a bunch of information about each embellishment, including rough estimates of average value accounting for uptime, and how much damage/throughput you could expect from each embellishment (as a percentage gain) at the current max crafted item level (636). We also include a projection of late-tier item levels (714), which we will use to examine how these will scale going forward.
We will explain the data in more detail below, so feel free to skim past it for now. Anything in red or orange is very likely a bug on Blizzard's part, which we will explain below as well.
| Scale Type | Scale Curve | Frequency | 636 Avg | Gain | 714 Avg | Gain | Comments | |
|---|---|---|---|---|---|---|---|---|
| Adrenal Surge Clasp | item | primary | rppm (flat) | 938 prim, -89 mastery |
0.97% | 1940 prim, -89 mastery |
1.28% | The loss of mastery scales on player level whereas the primary stat buff scales on item level |
| Beledar's Bulwark | item | secondary | rppm (flat) | 266 vers | 0.27% | 266 vers | 0.27% | |
| Binding of Binding | player | secondary | rppm (flat) | 837 secondary | 0.96% | 837 secondary | 0.96% | Only affects your ally and not yourself, so technically worthless, right? |
| Fractured Gemstone Locket | player | secondary | fixed | 1269 secondary | 1.45% | 1269 secondary | 1.45% | Very generous estimate as if you have 9 gems socketed. Only affects your allies and not yourself, so technically worthless, right? |
| Sanctified Steps | item | direct | rppm (flat) | 9827 dps | 0.79% | 32255 dps | 1.17% | Splits damage, but does not indicate if it is increased per target hit or not |
| Sanctified Torchbearer's Grips | item | direct | rppm (hasted) | 6834 dps | 0.55% | 24029 dps | 0.87% | Not reduced for non-dps specs |
| Siphoning Stiletto | item | direct | rppm (flat) | 8486 dps | 0.68% | 27852 dps | 1.01% | |
| Vambraces of Deepening Darkness | item | direct 2 | rppm (hasted) | 10037 dps | 0.8% | 25855 dps | 0.94% | Increased by up to 60% vs 5 targets (+15% for each additional target up to 5) |
| Waders of the Unifying Flame | player | secondary | rppm (flat) | 419 secondary | 0.48% | 419 secondary | 0.48% | Gives your highest secondary stat |
| Woven Dawn | player | secondary | rppm (flat) | 1674 mastery | 0.96% | 1674 mastery | 0.96% | Set bonus, requires both of your embellishments to activate, so we have cut the gain estimate in half |
| Woven Dusk | player | secondary | rppm (flat) | 1674 haste | 0.96% | 1674 haste | 0.96% | Set bonus, requires both of your embellishments to activate, so we have cut the gain estimate in half |
| Ascendance | player | direct | fixed | 712 secondary | 0.81% | 712 secondary | 0.81% | Estimating this as avg 8 total stacks of random secondary stats due to ramp time |
| Blessed Weapon Grip | item | secondary | rppm (flat) | 270 secondary | 0.31% | 367 secondary | 0.42% | The spell data seems to indicate that this scales with item level, but the optional reagent in-game shows no value in the tooltip, so can't verify |
| Captured Starlight | item | direct | fixed | 2952 shield/sec |
-- | 9107 shield/sec |
-- | Being generous and estimating with a 3 min CD -- max reduction from unique gem colors |
| Dawnthread Lining | player | secondary | fixed | 756 crit | 0.86% | 756 crit | 0.86% | Only when above 80% health, so of less value for tanks |
| Duskthread Lining | player | secondary | fixed | 756 vers | 0.78% | 756 vers | 0.78% | Only when above 80% health, so of less value for tanks |
| Concealed Chaos Component | item | direct | fixed | 5830 dps | 0.47% | 19138 dps | 0.7% | The damage is per target hit, highly variable how many targets you could reliably hit |
| Elemental Focusing Lens | item | direct | rppm (flat) | 7951 dps | 0.64% | 24527 dps | 0.89% | cannot crit |
| Energy Redistribution Beacon | -- | -- | -- | -- | -- | -- | -- | Hard to really rank this, it is net-zero healing but could theoretically be beneficial |
| Pouch of Pocket Grenades | item | direct | rppm (flat) | 4601 dps | 0.37% | 15101 dps | 0.55% | Splits damage, but doesn't say it increases for more enemies hit... haven't tested whether it actually does or not |
| Prismatic Null Stone | -- | -- | -- | 0.15% crit damage | 0.15% | 0.15% crit damage | 0.15% | Giving a best-case estimate that you have 4 unique gem colors |
| Radiant Focus | item | secondary | rppm (flat) | 418 secondary | 0.48% | 418 secondary | 0.48% | Assumes that you get full value from each trigger. Scaling is capped at ilvl80 (lol) |
| Symbiosis | player | secondary | fixed | 602 vers | 0.62% | 602 vers | 0.62% | Assuming 4.125 stacks on average accounting for ramp time, fight length will impact this a little |
| Vivacity | player | secondary | rppm (flat) | 393 secondary | 0.45% | 393 secondary | 0.45% | Triggering spell school changes which stat you get |
| Writhing Armor Band | -- | -- | -- | -- | -- | -- | -- | Doubles your other Nerubian embellishment |
Quick Observations
Without getting into the weeds yet, the main takeaway from this data is the Gain column (estimated % increase to your damage/throughput). It varies quite a bit from one embellishment to the next... the better ones in the 1% range and some of the worst ones down at 0.25% or worse. Furthermore, this value is not consistent as item level increases for some of them.
It is our opinion that the goal should be roughly similar power across the different embellishments, and that the % gain should stay roughly the same as item level increases. So let's look into how these could be improved.
Background
Some of this data is a bit technical, so first we'll explain what you are seeing above.
Scaling
For each embellishment you will see a Scaling Type and a Scaling Curve. As an example take Dawnthread Lining: its scaling "type" is "player" and scaling "curve" is "secondary". This means that the magnitude of the critical strike buff scales with the equipping player's level, and it uses the secondary stat curve. The way this works under the hood is that there is a spell effect for that buff with a coefficient -- just a number. Then there is a table that says at player level 80, the secondary stat curve value is X. You multiply the coefficient by X, and that's your buff value.
There are of course annoying complications that make it a pain to determine the proper value from time to time, but that's the basic idea.
There are two scaling types for embellishments: ones like Dawnthread Lining scale on the equipping player's level. Others like Siphoning Stiletto scale with the item's item level. Then there are four curves of interest:
- Primary: used for primary stats like strength
- Secondary: used for secondary stats like crit -- they scale up slower than primary stats
- Direct: used for direct damage/heal/shield effects
- Direct 2: less commonly used for direct damage/heal/shield effects -- it scales up slower than the regular Direct curve
Frequency
The frequency is how often you get the embellishment's buff or trigger its damage effect. Most of them are rppm, which stands for "real procs per minute". It is just a fancy technique that Blizzard put in a while back to make an effect trigger randomly, but about N times per minute on average. Sometimes rppm will be increased by your haste, and sometimes it won't. We make note of that for each embellishment.
Other embellishments will say "fixed" -- that just means they trigger on a fixed interval (e.g. every 8 sec), or are always up, like Dawnthread Lining.
Average Value
We have determined an average value for each effect above. For stat buffs this is straightforward: we determine the amount of the buff and multiply it by the uptime.
For direct damage effects we similarly determine how often they trigger and for how much to get a damage per second (dps) value, but some assumptions need to be made here: most direct damage effects can crit, happen more often with haste, and get multiplied by your versatility.
At item level 636 we did these estimates at 25% avg crit, 40% avg haste, and 10% avg versatility.
At item level 714 we did these estimates at 33% avg crit, 50% avg haste, and 13% avg versatility.
The haste value includes averaging things like Bloodlust, temporary buffs, etc. Critical strike and versatility more or less include only what you would have on gear: any non-gear sources of critical strike don't affect items, e.g. a Warrior's Recklessness. (And yes, it is really annoying that the game does not clearly indicate what affects what.)
The increase in stats from item level 636 to 714 we determined using the actual item budgets at those item levels. The base values at item level 636 will of course vary from spec to spec, but this is good enough to do a relative comparison.
Gain
The gain was determined using some rough approximations:
- Against a single-target, someone in the 90th percentile will be around 1,250,000 DPS at ilvl636 and 2,750,000 DPS at ilvl714
- Using these DPS estimates, a 12,500 DPS embellishment is a 1% gain at ilvl636 and a 27,500 DPS embellishment is a 1% gain at ilvl714
- Around 875 rating of a secondary stat will give a 1% increase in damage/throughput, or 975 for Versatility which scales slightly worse; this is pretty consistent no matter your average item level
- Around 875 primary stat is a 1% increase in damage/throughput at ilvl636 or 1400 primary stat at ilvl714
Remember the goal here is to do relative comparisons, so perfect approximations are not necessary for this analysis. If we say an embellishment is a 0.9% gain but NO MAN it's like, DEFINITELY 1.1%... noted. Doesn't matter for our purposes. To get a more accurate ranking for your specific character you should use our gear optimizer.
Make scaling consistent
The goal here is twofold: an embellishment should give roughly the same percent increase to your damage/throughput no matter your current average item level, and similar effects should scale in the same way. We will note where this is not the case and how it should be fixed.
General note for secondary stat buffs: most of the secondary stat buffs scale on player level. For all practical purposes that means they are static and will not go up as item level goes up. Our guess as to why they did this, which is backed up by our estimated projections for higher gear levels, is that it keeps the value of the embellishment relatively consistent no matter the average item level of the player's gear.
General note for direct damage: most of the direct damage effects use the more aggressive scaling curve and scale with item level. By our estimates, the flatter scaling curve would give a more consistent result as we get into later tiers. As it is right now, we predict that the direct damage effects will tend to scale up and outclass the secondary stat effects by using the more aggressive curve. It is very possible that our rough approximations are a little off... it will be interesting to keep an eye on this.
Specific fixes per embellishment:
-
Adrenal Surge Clasp:
scale the mastery loss with item level
All embellishments should be roughly the same value at any player average item level.This is a tough one... we predict that its value will go up due to how primary stat amounts and value will scale at higher item levels. One potential way to keep its value more consistent would be to also scale up the mastery penalty with item level. It would not completely balance it out, but it would get closer.
-
Sanctified Steps:
rppm should be hasted
All damage effects should scale with crit, haste, and versatility.Blizzard probably didn't haste the rppm because it also gives a speed buff... but we would argue that a small speed buff getting its trigger rate hasted is less of a scaling problem than the damage effect not benefiting from haste
-
Sanctified Torchbearer's Grips:
damage should be reduced for non-DPS specs
All pure damage effects should be reduced for non-dps specs.Maybe they just missed this one because tanks can't use it... but healers can.
-
Siphoning Stiletto:
rppm should be hasted
All damage effects should scale with crit, haste, and versatility.Blizzard may not have hasted this because it also hurts yourself... but the self-damage is very tiny and the trigger rate is very low.
-
Vambraces of Deepening Darkness:
change to the same curve as all other damage embellishments
All damage effects should use the same curve.We see no reason why this one effect should scale differently than all the others. As noted above, we think there is a case for changing all embellishment damage effects to use the "flatter" curve used by this effect. But even if our prediction is off on that front, this effect should at least be changed to match all the others.
-
Ascendance:
change to secondary scaling curve
All secondary stat effects should use the secondary scaling curve.This is kind of a hidden one... only data nerds like us would even know that they used the wrong curve.
-
Blessed Weapon Grip:
change to player level scaling
All secondary stat effects should use player level scaling.We are not 100% sure if this is actually using item level scaling because we have not tested it in-game. The spell data seems to indicate that it does though. Every other secondary stat embellishment is using player level scaling, so this one probably should too.
-
Concealed Chaos Component:
should scale with haste in some way
All damage effects should scale with crit, haste, and versatility.Perhaps this doesn't scale with haste because it also does a snare. We would argue that all damage effects should scale with haste though, so if not the trigger rate, then perhaps the number of mines, or they could make the explosion a short-duration DoT with hasted ticks.
-
Elemental Focusing Lens:
rppm should be hasted, damage should crit
All damage effects should scale with crit, haste, and versatility.This is a pretty straightforward damage effect... we can't think of why it should not scale with crit and haste.
-
Pouch of Pocket Grenades:
rppm should be hasted
All damage effects should scale with crit, haste, and versatility.Another basic damage effect... there is no reason the trigger rate should not be hasted.
-
Radiant Focus:
should use player scaling
All secondary stat effects should use player level scaling.Another kind of hidden one... they fixed a few of these but this is a holdout... it is using item level scaling fixed to item level 80 rather than player level scaling. Should be changed for consistency.
Balance the embellishments
Now that we have consistent scaling across the embellishments, we can look at balancing some of them. Below are specific suggestions for ones that seem the most off.
NOTE: the suggestions below assume that the goal is for all embellishments to be of similar power. It is possible that some of these are intentionally weak because they are cheaper to craft, but there is a lot of obfuscation to the cost of crafting with all of the different reagents and materials now, so it is tough to say.
Our guess is that the intentionally cheap and weak embellishments would be Blessed Weapon Grip, Concealed Chaos Component, Pouch of Pocket Grenades, and Prismatic Null Stone. We have not checked if these are actually cheap, but they are definitely weak.
- Beledar's Bulwark: multiply value by 3x
- Sanctified Steps: fix haste scaling and multiply value by 0.85x
- Sanctified Torchbearer's Grips: multiply value by 1.5x
- Siphoning Stiletto: fix haste scaling and multiply value by 0.95x
- Blessed Weapon Grip: multiply value by 3x
- Concealed Chaos Component: multiply value by 1.5x, keeps it a little weaker due to snare
- Elemental Focusing Lens: fix crit/haste scaling and multiply value by 0.8x
- Pouch of Pocket Grenades: fix haste scaling and multiply value by 1.75x
- Prismatic Null Stone: make it increase Blasphemite effects by 100%
- Radiant Focus: multiply value by 1.25x
- Symbiosis: multiply value by 1.25x
- Vivacity: multiply value by 1.33x
Final thoughts
It will be interesting to see if Blizzard makes adjustments to the embellishments as time goes on. Most likely we would see rebalancing when a new tier comes out and new item levels of crafted items become available.
A lot of the balancing in this game seems to be reactive: Blizzard sees how people are playing and how certain effects are performing, and then makes an adjustment when necessary. That is a decent approach... but sometimes we wonder if it also skews the data a little. How many people are actually going to use the embellishments that every guide, optimizer, and simulator says are bad?
Thus we think there is a good case for doing pure theory like this article. If you calculate the numbers and they are way off... it is probably worth taking a look.