Confirmed Pet Melee Lag

Discussion in 'Bug Reports' started by Sancus, Oct 25, 2019.

  1. Mithra Augur

    TLP doesn't have raid lag until around TBL expansion
    Fenthen likes this.
  2. Fenthen aka Rath

    Budget servers hosted in a budget datacenter tend to have that effect.
  3. Razorfall Augur

    Do you have parses for these? I'm curious how much difference it makes without the lag

    I'm wondering if I should set an add block hotkey for when my bst is in raids, and a remove block hotkey when she isn't
  4. Hamshire Augur

    Any update?
  5. Lodestar The Undefeated

    Recently ran extensive parsing for 4 weeks comparing all melee procs on pet, vs. none and all blocked (Bristlebane) (raids only).

    There is still an exceptionally significant impact running procs on pets vs. all blocked (only in raids). The margin of error is not small here. Therefore, it's a perpetually pronounced issue consistently identifiable across any sample size. Additional parsing was not needed to clearly identify the issue remains today on Live.
  6. Bunkerbones New Member

    With the recent patches, is it still wise to be blocking these pet buffs? Or has this been fixed?
  7. Lodestar The Undefeated

    Refer to my last post directly before yours. That is the state of affairs.

    Daybreak continues to put in exceptionally minor changes (such as reducing the amount of swarm pets in the last patch), but it has not moved the needle. All pet buffs that induce a proc or spell cast should be blocked still.
    Bunkerbones likes this.
  8. Rexa Ask someone else for buffs

    I, too, would love to see an update on this.
  9. Lodestar The Undefeated

    The update was already provided. What update do you want? I extensively parsed this in Q2 2023, and you can as well to prove it yourself.

    There is still a massive parse difference as of Q2 2023 in running all pet buffs/spells, vs. blocking everything. The margin of error is massive, requiring minimal and entirely imprecise parsing. This is why I suggest you do your own parse, as it's an obvious variance requiring minimal time to self prove. It's also instantly identifiable and corroborated by your pet doing hardly no animations (with all proc spells loaded on your pet), vs continuous and normal animations with all proc spells blocked (i.e., proc spells on your pet generate melee swing delay/lag).

    You will yield more than a 10% gain in DPS as result of blocking all pet buffs/spells. As an immensely dumbed-down example: If you are a beastlord and usually #3 in the parse, you'll drop to #10 instead.
  10. Warpeace Augur

    This is another problem thread with Darkpaw that affects several classes and was labeled "Confirmed" and then no communication.

    They will want people to buy the Friends and Family bundle for the new expansion with the same and more bugs.
  11. Lodestar The Undefeated

    Well that's a great point indeed. The update needs to come from Darkpaw now on next steps.

    Frankly, I'd expect exceptionally minor improvements (decreasing AA swarm pet counts, proc spells, etc.) is the best they can do without breaking the game at this point.
  12. Missiny Journeyman

    Many thanks @Sanctus for all the effort you put into the issue.
    Regarding the groupgame: I understood that you surpress all casts/proccs even in the groupgame apart from a defenisve procc if tanking. Did I understood this well?
    While raiding petclasses are affected most, even in the groupgame it is detrimental to allow any casts or proccs for the main-pet? So any rusty dagger will result in a better performing pet, than using any of the good mage-summoned weapons?
  13. Kosha Journeyman

    I've heard mixed messages on pet procs in group content. Hopefully someone who has taken the time to parse can tell you.

    This has the feel of being something they won't be able to fix due to lack of resources.
  14. Lodestar The Undefeated

    The issue is only in raids.

    In groups, you should run all spells/procs. I parsed that out as well, and anyone can parse it on a combat dummy too to self-prove if desired.
  15. Sancus Augur

    No, it's not only in raids. There's parseable melee degredation even against combat dummies, although to a very minor degree. In low-lag scenarios, it mostly manifests in degradation of backstab rate, so unless you test with a pet that can backstab you need very long parses to see the difference. That said, for pets that backstab, it is still a (very small) DPS loss to have any procs. In group content, it depends on the lag (population & EQ Zone Box) of the zone you're in. In most cases your experience will be similar to that on a dummy, but there are specific scenarios in group content (based on exogenous variables) where the effect can be more severe.

    At the end of the day, it's much less important that you block procs in group content than in raids, and there are some procs (e.g., Prism Skin) that provide enough benefit you likely want them in group content. For pure DPS purposes, at least with backstabbing (Mag/Nec Rog) pets, it is still optimal to limit procs.
    Barton, Missiny, Swiss and 3 others like this.
  16. Sancus Augur

    To put some numbers to it, here's a somewhat extreme scenario:

    No procs:
    Mage water pet (EM 31) buffed with Burnout XV Rk. III and Frenzied Burnout XVII equipped with two Summoned: Daggers, spellhold on:
    Combat Dummy Meda II in 3697s, 1.51B Damage @407.66K, 1. Puksu = 1.51B@407.66K in 3697s

    Same scenario but with two Summoned: Shadewrought Staffs, Spellhold off, Primal Fusion, Iceflame Barricade Rk. III, Companion's Elements IV, Apex of the Elements IV, Night's Endless Terror Rk. II, Arcane Ballad, and Travenro's Song of Suffering (in-group Bard playing two-song melody):
    Combat Dummy Meda II in 3620s, 1.35B Damage @371.83K, 1. Puksu = 1.35B@371.83K in 3620s

    Pierces declined from 3.797 per second to 3.197 per second, while Backstabs declined from 0.232 per second to 0.199 per second. That's a decline of 15.8% and 14.2% respectively. Typically in less extreme situations I've seen quicker declines in Backstab rates as stated, but this was enough to degrade everything pretty significantly. Procs here did 7K DPS in total, nowhere near enough to make up for the decline in hit rate.

    In most group scenarios you're going to have fewer procs (and you may or may not have overhaste), so the impact is probably not this large unless you're in a particularly laggy zone. The issue is definitely not solely consigned to raids, though.
    Missiny, Swiss and Barton like this.
  17. Lodestar The Undefeated

    Sancus: I'd recommend controlling for all possible variables, including weapon. Run your parse comparisons by only changing spells/procs, and absolutely nothing else. Don't provide a different weapon, or change anything other than the specific variables that need to be controlled for. I'd be interested to see your results by using Summoned: Daggers in both cases instead, while changing nothing other than spells on the pet. Run it to the extreme, with zero spell procs vs. all spell procs.

    When I controlled for all variables, in a Guild Hall with 30+ players I saw absolutely no statistically significant difference in hit rate attempts of a beastlord pet with all pet procs running vs. none. Hit rate was identical with every pet proc running vs none. Specifically, the hit rate attempt is all that matters to parse in this case--DPS will always have a margin of error and is therefore irrelevant here. Hit rate attempt is what's directly impacted as result of lag, and is the propulsive variable that drives DPS and should itself be measured.

    I will say that I could believe that going into a laggy zone could induce the same issues as seen in a raid. However, that's beyond my concern and I intentionally did not measure it, as group zones rarely experience the same issue as raid instances. At the point one begins removing spells in group content, then you are intentionally creating the risk of lowering your DPS rather than increasing it; or, spending extra time shuffling spells that may make no difference, given evidence that hit rate attempt doesn't change in a normalized non-raid environment with all pet spells vs. none.
  18. Sancus Augur

    Here's the exact same proc test except with Summoned: Daggers equipped:
    Combat Dummy Meda II in 3620s, 1.33B Damage @367.72K, 1. Puksu = 1.33B@367.72K in 3620s

    My pet performed 3.175 pierces per second and 0.198 backtabs per second, declines of 16.4% and 14.7% compared to the non-proc tests. This is actually slightly worse swing rates vs the parse with proc weapons, but that's likely a combination of RNG and potential differences in serverside lag (it is Saturday now).

    And here's the above scenario but with spellhold on (so just proc buffs):
    Combat Dummy Meda II in 3632s, 1.36B Damage @375.24K, 1. Puksu = 1.36B@375.24K in 3632s

    My pet performed 3.278 pierces per second and 0.203 backstabs per second, declines of 13.7% and 12.5% compared to non-proc tests. This shows some improvement vs spellhold off (as expected), but still degradation vs avoiding procs.

    The reason I included proc weapons was to properly include the sources of pet procs that people would typically use but have to avoid due to this bug: proc buffs, innate pet casting, and proc weapons. Pet weapons are (counterintuitively) not an input into pet melee DPS beyond allowing Rogue pets to backstab from all angles (all of these parses were from the rear, so that effect doesn't matter). Therefore, the only impact of the weapon change was adding an additional proc, and using weapons with procs in the group game is exactly what someone would do if they were under the incorrect belief that procs only degraded pet melee in raids.

    Anyhow, however you slice it, these parses show a clear, demonstrable impact of procs on pet melee rates in non-raid content. YMMV based on the quantity of procs you include and exogenous zone/serverside factors (this is a lag-induced bug after all), but it's simply not a raid-only phenomenon.
    Fenthen, Rexa, Raccoo and 3 others like this.
  19. Missiny Journeyman

    many thanks @Sanctus
  20. Lodestar The Undefeated

    I re-parsed all of this on Bristlebane over the weekend outside of raids, and outside of the guild hall. Margin of error must be taken into consideration. If you are seeing gains or decreases < +/- 5%, then that's all well within margin of error in which you can also argue that there's risk of intentionally decreasing DPS by removing spell/weapon procs. I'd argue that ~ +/- 10% is still within the margin of error. Some of the deltas Sancus is quoting above are closer to 2-3% rather than 12%, so context is not really sufficient. Guild Hall lag also is not consistent with some lag swings seen in dynamic zones.

    In the group game: If you want to argue against running BST 2.0 epic that has a 200 damage proc, then that makes sense and I'd concur (unless you want gains from defensive avoidance). However, arguing against running 15K procing pet weapons and Night's Perpetual Terror procs does not make sense or correlate with the margin of error in the data. You're going to have to run real-world multi-hour parses, which is essentially not something anyone is going to successfully accomplish when controlling for all variables in this game, in zones that continually fluctuate on +/- minimal lag during the very same parses. I don't expect anyone will be finding any mobs with enough HP to hit for 3 hours straight, or proper meta-analysis aggregation of multiple mobs without extensive error in their data.

    Ultimately the consensus is: If you're raiding, remove every single pet/spell proc, as the effect on pet hit attempts is overwhelmingly detrimental and obvious. However if you're grouping, only remove pet/spell procs that are exceptionally minimal in damage output; or, that activate very often (e.g., BST 2.0 epic) that are also damage-minimal in nature. The goal in the group game would be to only eradicate high proc-per-minute (PPM) pet spells/procs, or very low damage effects that are essentially unparseable.