Riftworld Codex Analysis

The Riftworld Codex is a new tanking trinket obtained from Portal Keeper Hasabel in Antorus. It is one of those trinkets where the in-game tooltip doesn’t tell you a whole lot about how it works, and we’ve seen some pretty weird stuff said about it on the internet, so we thought it would be worthwhile to test the item and describe precisely how it is working.

##How does it work?
Here is a recent log of someone with this trinket: https://www.warcraftlogs.com/reports/31H6Vz8cLjXTbBNC#fight=6&type=healing&source=98&translate=true

From this log (and we have compared it to others as well), we can determine a few of the more ambiguous mechanics:

  1. The trinket has a flat 3 RPPM chance to trigger, not increased by haste. Also, we have seen it said somewhere that each effect has its own chance to trigger, effectively tripling the number of procs you would see. No logs that we have examined support this claim, not sure where it came from. If you have a log that supports a different proc rate, please link it!

  2. The ticks of the heal-over-time proc (Winds of Kareth) are hasted, and can crit, and are modified as expected by vers, healing multipliers.

  3. The ticks of the DoT (Flames of Ruvaraad) are not hasted (always ticks 4 times), and can crit. The heal is a multiplier on the damage, and thus never crit independently (otherwise it would “double dip” on crit).

  4. The absorb shield (Light of Absolarn) total amount never “crits” for a bigger shield. It is increased by versatility like most trinket absorb effects.

##Is the trinket any good?
We usually let simulation data answer this question for us – short version: it is a decent trinket compared to the other Antorus tank trinkets – it might be slightly weaker than some of the others, but not significantly so. It can depend a lot on the fight and your spec, etc. We have ranked it via our adaptive strategies, and you can also run your own simulations against specific boss fights and try it out. Ultimately you want a personalized ranking for this – there is no blanket answer that works for everybody (and that’s exactly what our adaptive tank strategies do).

Many people overestimate how much it will overheal. That depends a lot on your healers and how hard the fight is, and a bit on luck. The highest we have seen is around 60% overheal – specs that heal a lot like blood DKs will tend to be higher on the overheal. We have also seen as low as 10% in logs. A good “average” to expect is around 30% overheal.

Many people also dislike this trinket because it has crit on it, a stat that is generally not as desirable as other stats for many tank specs. The general community tends to exaggerate the gap in value between secondary stats – while maybe not the best stat for say a protection paladin, crit is by no means a bad stat – it is pretty close in value to anything except haste usually (it improves SotR uptime, increases average size of certain healing effects, gives some parry, etc.). We just let the data handle stats – adaptive strategies will figure out how much of a difference in value the stats will cause. We don’t need to wave our hands or be so black-and-white about it: each individual gets a personalized estimate of the value of that crit compared to available alternatives.

The fact that it is a random proc instead of an on-use doesn’t really impact its value as much as people tend to think either. (You can easily do this experiment in our simulator by removing the proc and making it an active trinket with a 20-second cooldown.) This trinket essentially contributes “background” healing. Raising the overall amount of background healing that a tank receives will lower their chance to die by lowering the probability that they will be in danger when a hit comes in.

For the average player, a proc trinket can often be more valuable as there is a natural tendency to “save” on-use trinkets until things go really terrible, when in fact this is often not the optimal way to play. Being more proactive and aggressive with trinket use will generally perform better, even if you guess wrong and “waste” a use from time to time. Random procs essentially force you to be “aggressive” with a trinket, getting the maximum uses per fight at the expense of perfect timing.

The tank community in general tends to have a bias against effects like this that are harder to “feel” except in an aggregate, statistical sense. For the longest time, many tanks avoided stamina when given the choice – they felt that getting a little extra stamina made no real difference in toughness. This was proven to be false – more stamina is in fact the single best way to get tougher for all tank classes, even in small amounts. “Background healing” serves a similar purpose to stamina and armor – it makes hits relatively smaller, on average, to the size of your current health pool at the time you get hit.

Due to the complexity of how things can interact with background healing, and how “opaque” it can seem just sitting here and reasoning about it, simulation is by far the best way to determine the value of such effects. If it reduces your chance to die on a difficult fight by more than some other effect, we generally consider it better. That is the most straightforward way to evaluate this trinket, and that is how our simulator and strategies handle it.