First I’d like to say that the in-game representation of what percentage of shards may be succulent souls is vague enough that both players and theorycrafters quite literally have no idea what the proc rate is at all. I would welcome some clarification from blizzard on this front, as this is an integral part of two spec’s hero talent tree (demonology and affliction).
@Kaivax
Bugs:
-
This past week I noticed a lot of variance between the perceived proc rate of flat 20% and 15% that we assumed were the succulent soul proc rate for affliction and demonology respectively. Instead, it is higher than originally assumed at roughly 22%, but it isn’t clear whether this is a bug or intended behaviour using a bad luck protection formula on resource generation.
-
Feast of souls is triggering a “resource check” on both the proc itself, as well as the shard that it produces, giving it double the inherent soul shard chance to proc an additional succulent soul, up to a maximum of 2 succulent souls. This is giving the player 100% more double succulent soul events on a Feast of Souls proc than it should be, given our understanding of Feast of Souls being a single resource-producing effect. EDIT: There is also the possibility that the on-death effect is based on killing blow rather than the target dying, should probably check on that too.
-
In all events that proc multiple succulent souls more than what is intended (such as soul rot and feast of souls giving +1 or +2 extra succulent souls), it does not yield a usable soul shard for the player. For instance, when soul rot produces 4 succulent souls, it still only gives 3 total shards instead of what succulent souls reads (as in it gives the player a succulent shard to use) in the hero talent, which would be 4 total shards. If this behaviour were to be consistent with the the soul harvester talent “demonic soul”, it should read that, “some shards may contain 2 succulent souls”, although that would also not give a full understanding of what is occurring here.
I will say that if any of these bugs are fixed, soul harvester will see a massive reduction in performance as these are all propping up it’s damage significantly in game (outside of simulation).
While it’s not clear whether Blizzard would like us to know how certain procs in the game function, this is a relatively important distinction, as we can almost no longer ascertain what is a function or a bug with one of our hero talent specs.
I’ll outline some of the testing that is ongoing atm for theorycrafting:
I will preface the rest of this post with the warning that this is still ongoing theorycraft work, but would be very helpful with a response from Blizzard
After some testing in the past couple of days, a few other contributors and I did some log and target dummy testing to try and find the real proc chance for succulent souls. The underlying assumption was that affliction had a 20% proc chance on any resource-adding event (gaining a shard) with some form of bad luck protection built in to proc a succulent soul. This is also the case with demonology having a 15% chance per shard event. On the whole, this seems very demonology favoured, as the 5% less chance does not make up for the immense shard generation disparity between both specs (outside of direct spec tuning for both that skews the additional hero talent points more for one spec than the other). The underlying effect is that demonology inherently gains more in-game benefit from the passive succulent shard procs than affliction. While this might not have parity in base spec tuning vs class hero % bump (the additive bonus of hero talents on the base spec), it does mean that affliction has less gameplay that interacts with the hero talent tree than demonology, which feels bad.
Agony Generation / Resource Generation
On a four hour dummy test running only shard generation from agony, the average proc rate for a succulent soul is 22%, which is 10% higher than what our original estimation should have been – obviously the caveat being a certain bad luck protection formula that is hidden from players that increases its chance. Here are the logs labeled with what they show:
Shard Generation
https://www.warcraftlogs.com/reports/Nw3t67AqnR8xyvQB?fight=last&type=resources&spell=107&translate=true
2811 shards gained from agony
Succulent soul procs
https://www.warcraftlogs.com/reports/Nw3t67AqnR8xyvQB?fight=last&type=auras&translate=true
621 succulent souls
Here is the table of succulent souls generated, accounting for previous shards generated (as in succulent shard gained after the previous proc) The interval between succulent souls procs was capped in our testing at 10 shards being generated normally (which is also a bad feeling for affliction in particular:
| Column 1 | Column 2 |
|---|---|
| F(0) | 36 |
| F(1) | 79 |
| F(2) | 94 |
| F(3) | 121 |
| F(4) | 100 |
| F(5) | 75 |
| F(6) | 58 |
| F(7) | 38 |
| F(8) | 14 |
| F(9) | 4 |
| F(10) | 2 |
As you can see from the chart, this does not equal anywhere close to both an accumulator rng factor (where the amount of succulent souls we receive back-to-back is statistically high enough that it doesn’t have time to accumulate), or a flat 20%. In this case, there is an extremely skinny tail and really fat middle skew that’s producing a strange effect on how we accumulate succulent souls that isn’t exactly intuitive in-game, resulting in most of our succulent souls to proc on average by your 3rd or 4th shard generated (this is important for soul rot). This formula also applies to any resource generation event that does not include feast of souls (which is the bugged part). This means that souls generated by soul rot (as highlighted in the section underneath), as well as rampaging demonic soul (new 4pc) are also generating succulent souls at this perceived rate. However, even with the underlying formula on bad luck protection, both soul rot and rampaging demonic soul seem to be bugged (skewed) positively in which they are giving on average more than even the bad luck formula would generate with roughly 22% succulent soul generation per resource event. Whether this is intended, is uncertain. Again, we don’t have the formula for succulent souls.
Soul Rot
For instance, when you hit soul rot, shadow of death mentions that soul rot guaruntees that each shard should be succulent, with three being generated. This should yield three suculent souls. However, each of those shard generation events, as soul rot is generating three resource-generating events, also have a separate roll that increases the chance to proc additional succulent souls following the previously outlined 22% ish proc rate.
This chart highlights another 8 hour log test on this aspect, with +1 noting four succulent souls gained, and +2 noting 5 succulent souls gained:
Soul Rot Log
(https://www.warcraftlogs.com/reports/Cx2AKXT3PYhDV9bj?fight=last&type=damage-done&source=1)
| Column 1 | Column 2 |
|---|---|
| +0 | 191 |
| +1 | 236 |
| +2 | 31 |
First, it seems that the cap of 5 total succulent souls seems to belie the understanding that each soul shard gained from both soul rot (and tyrant) could roll an extra shard. This should mean that in very extreme circumstances that the player is actually just wasting a proc as there is the potential to proc 6 souls with only a 5 shard cap. The cap is thus a limiter on extreme case scenarios, but ideally should be removed to better facilitate this type of interaction.
Unexepectedly, this test also yielded a result that highlights the same 22% additional factor noted early on the base resource generation for affliction. Moreover, this shows that the player should on average expect more than the outlined succulent shards that the talent suggests (3), which is not what the talent states. While some aspects of class design are opaque, the main function of a hero talent should probably be a little bit more transparent in how you use its design.
Feast of Souls
Feast of souls similarly suffers from a perception issue as the talent simply states “a chance” rather than a number for the player to easily digest what they ought to have during play. A problem during the alpha/beta cycle, as well as early the war within, was that without us knowing the true proc rate, or some aproximation of a proc rate, we actually just straight up didn’t know this talent didn’t function properly until an unknown date when it was bug fixed. This shouldn’t happen. A secondary issue is that the proc rate is tied to both nightfall as well as an rng effect, something which is vastly different than its choice-node counterpart shared fate. I can say that it doesn’t feel great to have a proc come from a proc. It’s double rng, and even bad luck protection can’t help feeling the extreme lows from this talent. In an ideal world, the on death aspect of feast of souls is removed and then the proc becomes 100% chance on nightfall cast.
In all our data run, this proc chance follows the same rough formula as the other succulent shard changes, but has an estimated base proc of 12.5% on nightfall consumption to proc a succulent shard, with some form of added bad luck protection. In game, this is closer to 13.5%. If we were to apply, on average, our 22% base succulent soul modifier to this proc rate, we should see a yield of at least 16.5% total succulent shards per proc gained over combat (13.5% for the proc, and an extra 3% from double procs). However, we’re seeing this chance be roughly 19.5% - 20% total in-game. The problem is that feast of souls, it seems, is proccing both on the resource event itself (the shard gained) as well as the check of whether the player has procced feast of souls. The result is roughly 100% more double procs that the player can expect than what the talent reads.
The net result is that on average players are receiving more succulent souls than they should be, but almost everybody can’t tell (as the current assumption is that this has been an expansion-spanning issue/intention. I think this is both a hallmark of poor agency design for succulent souls in general, as well as not having enough “feel good” moments for affliction in particular to have with one of their hero talents.
As a final point of feedback, something that could be a good stop-gap measure would be to allow rapture to have an instant cast time if it will consume a succulent soul shard.
TLDR
- Feast of souls has some unintended bonus effect that is skewing results positively in game and increasing our chance at “bonus” succulent souls from a resource-generating events.
- Succulent shard data has shifted throughout the expansion unexpectedly, and at this point it’s difficult to find a pattern of what the proc rate should be
- The addition of more resource-generating effects for soul harvester affliction has made this problem a much more apparent issue with a lack of transparancy for this hero talent as a whole. It’s like going to mcdonalds and ordering a meal, but not knowing any of the ingredients and trusting that they are correct (or just flat out guessing).