Can you explain the effects of multiple Achilles' Heel debuff procs?

[RelativeQuanta]

4 minutes ago, Doomguide2005 said:

All critter damage resistances act as resistance to resistance debuffs vs that damage type.  So no afaik any critter can, depending on their damage resistances, resist any resistance debuffs acting against that damage resistance type not just AVs and GMs.  That's why in, for example, Hard Mode it's beneficial for any AT to hit 100% damage resistance.  That way even a Scrapper or other AT can resist all of a incoming debuff meaning no matter how large the attackers resistance debuff their resistance to that damage will be capped at 75% not 75 minus the incoming debuff as they can in fact resist 100 % of the resistance debuff.  Only unresistable resistance debuffs would apply.

My earlier reply to @tidge still stands here.  The more resistance a critter has, the more valuable a -Res debuff is.  A critter that has 50% resistance to everything that has a -20% Res debuff applied to it will take 40% more damage than it did before the debuff was applied and a critter with 90% resist all will take 3x as much damage once a -20% Res is applied.  Overcapping Resistance isn't something normal Mobs do.  

[tidge]

So I misremembered 20% as 10%, but I have been pretending (for simplicity) that the single type of debuff is "never expiring"... either it there is a power with it that never misses and always procs (and is used within 10 seconds) or more likely multiple powers have the same piece slotted...and enough of those hit and proc to make the effect "eternal".

 

There are a lot of possible "what ifs" that have to be factored in, I've also tried to not explore cases that mix %-Resistance and %damage... Only because how well this works depends on attack chains, slotting choices, proc rates, etc.

 

It really is quite situational.

[RelativeQuanta]

@tidge  Exactly!!  That's what made it interesting enough to warrant simulating.  🤓 

 

I tested a build that I'm currently leveling towards that tries to get a balance set bonuses around Global Recharge, Defense, Resistance, and Damage.  I don't think it's that far from a "typical" min/max build for a Rad/Rad Sentinel, so I'm definitely extrapolating results on that build to hold "in general."  Perhaps that means I'm overstating my conclusion?

[Sunsette]

For a sentinel in particular I think that reduces the value of -res. Not only do they come with their own source of unresistable single target -res that is a huge force multiplier on hard targets, but the most important thing that ranged characters do in difficult content is handle adds -- which resist less and have lower HP. This mechanic is at play in starred ITF, Sutter, and several incarnate trials. 

 

I would pursue -res most on single target melee specialists, personally. 

Edited March 7 by Sunsette

[RelativeQuanta]

18 minutes ago, Sunsette said:

For a sentinel in particular I think that reduces the value of -res. Not only do they come with their own source of unresistable single target -res that is a huge force multiplier on hard targets, but the most important thing that ranged characters do in difficult content is handle adds -- which resist less and have lower HP. This mechanic is at play in starred ITF, Sutter, and several incarnate trials. 

 

I would pursue -res most on single target melee specialists, personally. 

 

Quite possibly!  Though, for this particular build in an AoE situation the -Res build had basically the same personal damage as the one with an additional minor damage proc everywhere except a single high proc rate AoE.  The uptime on the -Res proc in the build with Achilles' Heel slotted all over the place on average mob was over 50% of its lifetime.  Granted, I set mob hitpoints at 2,000 in the AoE test (basically making them all LTs), so with a more realistic mix, it might not work as well.

 

Perhaps I'll test this next!

[RelativeQuanta]

Ok, thanks to feedback from @tidge, @Sunsette, and noticed other bugs in my code, I've updated the simulation.

 

TL; DR: -Res proc is ahead at even con and damage proc is imperceptively ahead at +4.  The moment you team -Res will pull ahead, by a lot.

Edit: Updated with 1000 trial results

 

1. Electron Haze had the wrong target cap of 10.  This was fixed
2. Electron haze was dropped for Neutron Bomb
3. Level shifts (purple patch) was added and a new trial is at +4 mobs
4. Group trial now is closer to an x8 group:
   1. 2 bosses (2650 hps)
   2. 4 Lieutenants (858 hps)
   3. 8 Minions (431 hps)
5. PBAoE powers now hit only random mobs and don't always hit the targeted mob
6. All trials assume the -Res debuff applies to the power that procs it
7. I've added Annihilation -Res debuff into the mix too

Build Proc Chances Per Power

Irradiate: 26.44%
Cosmic Burst:  45.87%
Neutron Bomb:   38.35%
Proton Stream: 79.74% (3.5 ppm), 90% (4.5 ppm) 
Atomic Blast: 90% 
Ground Zero: 90% 
Fire Sword Circle:  90% 
Cremate: 37.99% (3.5 ppm), 48.84% (4.5 ppm)

 

 

Trial 1

x-ray beam: no procs
Irradiate: achilles_heel 
Cosmic Burst:  achilles_heel
Neutron Bomb:  Annihilation , minor_dmg 
Proton Stream: achilles_heel, epic_dmg, minor_dmg, minor_dmg
Atomic Blast:  achilles_heel, epic_dmg
Ground Zero: achilles_heel, minor_dmg, minor_dmg, minor_dmg
Fire Sword Circle: minor_dmg, minor_dmg, minor_dmg, FotG 
Cremate: epic_dmg, minor_dmg

1. 1,000 trials of a single Mob with 10,000 health and level shift of +0:
   1. Average mob defeat time 44.04s Standard Error 0.06s
   2. Mob average lifetime debuffed with Achilles' Heel: 86.53% Standard Error 0.32%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 31.16% Standard Error 0.40%
   4. Mob average lifetime debuffed with Annihilation: 1.43% Standard Error 0.18%
2.  1,000 trials of a single Mob with 10,000 health and level shift of +4:
   1. Average mob defeat time 104.61s Standard Error 0.07s
   2. Mob average lifetime debuffed with Achilles' Heel: 71.87% Standard Error 0.23%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 18.96% Standard Error 0.13%
   4. Mob average lifetime debuffed with Annihilation: 0.39% Standard Error 0.04%
3. 1,000 trials of an x8 group and level shift of +0:
   1. Average mob defeat time 10.52s Standard Error 0.07s
   2. Mob average lifetime debuffed with Achilles' Heel: 21.19% Standard Error 0.29%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 6.31% Standard Error 0.15%
   4. Mob average lifetime debuffed with Annihilation: 15.49% Standard Error 0.26%
4. 1,000 trials of an x8 group and level shift of +4:
   1. Average mob defeat time 20.59s Standard Error 0.14s
   2. Mob average lifetime debuffed with Achilles' Heel: 49.28% Standard Error 0.32%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 10.14% Standard Error 0.15%
   4. Mob average lifetime debuffed with Annihilation: 16.87% Standard Error 0.22%

 

Trial 2

x-ray beam: no procs
Irradiate: minor_dmg 
Cosmic Burst:  minor_dmg
Neutron Bomb:  Annihilation , minor_dmg 
Proton Stream: minor_dmg, epic_dmg, minor_dmg, minor_dmg
Atomic Blast:  minor_dmg, epic_dmg
Ground Zero: achilles_heel, minor_dmg, minor_dmg, minor_dmg
Fire Sword Circle: minor_dmg, minor_dmg, minor_dmg, FotG 
Cremate: epic_dmg, minor_dmg

1. 1,000 trials of a single Mob with 10,000 health and level shift of +0:
   1. Average mob defeat time 47.70s Standard Error 0.06s
   2. Mob average lifetime debuffed with Achilles' Heel: 0.00% Standard Error 0.00%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 31.57% Standard Error 0.40%
   4. Mob average lifetime debuffed with Annihilation: 0.85% Standard Error 0.14%
2.  1,000 trials of a single Mob with 10,000 health and level shift of +4:
   1. Average mob defeat time 104.01s Standard Error 0.07s
   2. Mob average lifetime debuffed with Achilles' Heel: 0.00% Standard Error 0.00%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 19.41% Standard Error 0.13%
   4. Mob average lifetime debuffed with Annihilation: 0.16% Standard Error 0.03%
3. 1,000 trials of an x8 group and level shift of +0:
   1. Average mob defeat time 10.64s Standard Error 0.07s
   2. Mob average lifetime debuffed with Achilles' Heel: 10.29% Standard Error 0.23%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 6.08% Standard Error 0.15%
   4. Mob average lifetime debuffed with Annihilation: 14.00% Standard Error 0.24%
4. 1,000 trials of an x8 group and level shift of +4:
   1. Average mob defeat time 20.22s Standard Error 0.14s
   2. Mob average lifetime debuffed with Achilles' Heel: 33.13% Standard Error 0.33%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 9.40% Standard Error 0.15%
   4. Mob average lifetime debuffed with Annihilation: 14.45% Standard Error 0.20%

 

Conclusion

Ok!  With 1,000 trials it now shows that the debuff build is ahead of the damage proc build on an even con single target by nearly 4 seconds.  However, against a +4 single target, the damage proc build pulls ahead by 0.6s.  Group defeat times match within errors at even con and damage proc build is ahead by about  0.1s outside errors against +4.  

 

-Res proc is ahead at even con and damage proc is imperceptively ahead at +4.  The moment you team -Res pulls ahead.

Edited March 8 by RelativeQuanta
1000 trial update. Gave Sunsette credit for testing against a more realistic enemy group

[Jacke]

19 hours ago, RelativeQuanta said:

My earlier reply to @tidge still stands here.  The more resistance a critter has, the more valuable a -Res debuff is.  A critter that has 50% resistance to everything that has a -20% Res debuff applied to it will take 40% more damage than it did before the debuff was applied and a critter with 90% resist all will take 3x as much damage once a -20% Res is applied.  Overcapping Resistance isn't something normal Mobs do.  

 

Ah, that's not the way -Res works.

 

First, there's the Purple Patch.  Even irresistible debuffs are changed by the Purple Patch (as @Captain Powerhouse once mentioned).

 

For -Res, look at the Combat Modifier table and it's the strength of debuffs that changes, depending on the relative Levels between the toon and the target.

 

For the common case of +4 non-Incarnate-content missions with Incarnated toons with a T3 or T4 Alpha boost, it's +3.  So the value of the -Res will be reduced by a factor of 0.65.

 

That means the Achilles Heel special will have its base 20% -Res reduced to 13% -Res in that case.  And it's resistible -Res.

 

Then you don't need to know the Resistances of the target.  Because of way resistible -Res is resisted by Resistance to Damage:

https://homecoming.wiki/wiki/Resistance_(Mechanics)#Resistance_to_Resistance_Debuffs

 

that 13% -Res (or whatever the -Res value is after Purple Patch) in all cases becomes 13% extra Damage outside of Damage Enhancement.

 

 

Edited March 8 by Jacke

[tidge]

Again, sticking with a solo character:

 

IMO: One of the many factors that works against %-Res beyond the marginal increases in (inherent) typed damage is that in order to capitalize on %-Res, there is a chain of several links that has to happen in order for it to work out:

1. The player has to hit with attack containing %-Resistance
2. The %-Res proc has to go off (with a ceiling chance of 90%), and hit
3. The subsequent attacks have to come within the 10 second period
4. The subsequent attacks have to hit
5. The %-Res has to be reapplied (see 1 and 2, above)

So (1) and (2) are relying on slotting (inherent to the attack, but also global set bonuses and power choices) to maximize recharge, ToHit, and accuracy... but also trying to avoid slotted recharge (to maximize %proc rates.

 

(3) and (4) are more driven by slotting choices for accuracy and damage (because it's the damage-taken that is trying to be increased, plus slotted recharge because those attacks need to be ready to take advantage of the %-Res.

 

In contrast, %damage only relies on equivalent conditions for (1), (2)... and if the %damage doesn't go off... i.e. (2) fails... it's no big deal because losing on some %damage has no effect on the rest of the attack chain, except possibly requiring a greater number of attacks.

 

On teams, or with a MM or pet class... the non-players will be more-or-less constantly dumping attacks on targets, so this is how %-Resistance turns into a force multiplier. It helps some with Damage auras too, so on something like a Tanker that has relatively low inherent DPS attacks AND can get enemies to group... every little bit helps, and Bosses tend to stick around when playing a solo Tank.

 

In my experience, the way I have been able to maximize %-Resistance effects has been to use the same %-Resistance piece in multiple attacks where those attacks:

• Have decent %proc rates and accuracy (gotta hit, gotta land the %proc)
• Hit multiple targets (so AoE), because it's all random
• The attacks can chain, so the duration of the %-Resistance is increased

And furthermore... those slotting choices are generally made on LOW DPS characters, because otherwise I'd just have the high-DPS character spamming attacks (slotted with recharge, damage... while maintaining endurance and accuracy). It takes some effort for a solo player to stay at the damage cap such that %-Resistance is boosting damage taken by critters. <-If you are here, either %-Res and %damage will increase DPS.

 

I have my Masterminds slotted to keep up %-Resistance this way, using the Annihilation set (or at least the %-Resistance piece) everywhere possible.

 

I certainly have some characters that have added the Achilles' Heel %-Resistance piece to as many attacks as possible, but those are characters focused on teaming and not solo play.

 

A solo player that just wants to have -Resistance (notice lack of %) is probably better served by taking Weaken Resolve from the Force of Will pool. All it really needs is Accuracy, and it can also be slotted with the Achilles' Heel set if that is a deal breaker. It is a single target, but as I wrote before: (hard) single targets are almost certainly the ones that a solo player wants -Resistance on. Use the slots in the attacks for more damage (%damage is usually an option) instead.

 

[RelativeQuanta]

8 hours ago, Jacke said:

First, there's the Purple Patch.  Even irresistible debuffs are changed by the Purple Patch (as @Captain Powerhouse once mentioned).

 

For -Res, look at the Combat Modifier table and it's the strength of debuffs that changes, depending on the relative Levels between the toon and the target.

 

For the common case of +4 non-Incarnate-content missions with Incarnated toons with a T3 or T4 Alpha boost, it's +3.  So the value of the -Res will be reduced by a factor of 0.65.

 

That means the Achilles Heel special will have its base 20% -Res reduced to 13% -Res in that case.  And it's resistible -Res.

 

I included purple patch shifts in my last update.  This didn't change the result.

 

8 hours ago, Jacke said:

Then you don't need to know the Resistances of the target.  Because of way resistible -Res is resisted by Resistance to Damage:

https://homecoming.wiki/wiki/Resistance_(Mechanics)#Resistance_to_Resistance_Debuffs

 

Ok, I did not know this is what people meant by resistance resisting -Res.  I still maintain that since Radiation blast is energy damage, that assuming 0% resistance IS the general case, but I'll bite.  I'll add enemy resistances in, but this will take a bit.  Unless there's a better source of enemy resistances, I'm going to test using values posted here:

 

[tidge]

Wouldn't it be true that if an enemy had ZERO Resistance to a damage type that -Resistance wouldn't add ANY damage to that enemy, for that damage type?

[RelativeQuanta]

44 minutes ago, tidge said:

Wouldn't it be true that if an enemy had ZERO Resistance to a damage type that -Resistance wouldn't add ANY damage to that enemy, for that damage type?

Nope!  You can have negative resistance.  You can see it by activating rest.  It puts a huge -Res debuff on your character.

[Mister Mass]

I just wanted to add that, because RES resists -RES, mathematically -RES increases damage by the same percentage at every RES level.  To demonstrate this, let's quickly look at some examples:

 

Inflicted Damage = Power Damage x (100 - [adjusted] Target RES)

Proc -RES = 20%

 

At 0% Target RES, Inflicted Damage is 100% of Power Damage without the Proc active. 

With the Proc active, the Target's adjusted RES becomes -20%, so the Inflicted Damage is 120% of power Damage.

120% is 120% of 100%, so the Proc increases Inflicted Damage by 20%.

 

At 25% Target RES, Inflicted Damage is 75% of Power Damage without the Proc active. 

With the Proc active, the effective Proc value is 15% after 25% is resisted, so the Target adjusted RES becomes 10%, & the Inflicted Damage is 90% of Power Damage. 

90% is 120% of 75%, so the Proc increases Inflicted Damage by 20%.

 

At 50% Target RES, Inflicted Damage is 50% of power Damage without the Proc active.

With the Proc active, the effective Proc value is 10% after 50% is resisted, so the Target adjusted RES becomes 40%, & the Inflicted Damage is 60% of Power Damage. 

60% is 120% of 50%, so the Proc increases Inflicted Damage by 20%.

 

At 75% Target RES, Inflicted Damage is 25% of power Damage without the Proc active.

With the Proc active, the effective Proc value is 5% after 75% is resisted, so the Target adjusted RES becomes 70%, & the Inflicted Damage is 30% of Power Damage. 

30% is 120% of 25%, so the Proc increases Inflicted Damage by 20%.

 

At 90% Target RES, Inflicted Damage is 10% of power Damage without the Proc active.

With the Proc active, the effective Proc value is 2% after 90% is resisted, so the Target adjusted RES becomes 88%, & the Inflicted Damage is 12% of Power Damage. 

12% is 120% of 10%, so the Proc increases Inflicted Damage by 20%.

 

At 100% Target RES, Inflicted Damage is 0% of power Damage without the Proc active.

With the Proc active, the effective Proc value is 0% after 100% is resisted, so the Target adjusted RES becomes 100%, & the Inflicted Damage is 0% of Power Damage. 

0% is 120% of 0%, so the Proc increases Inflicted Damage by 20%. 😉

 

So while the amount of additional Inflicted Damage by the Proc decreases as Target RES increases, the percentage of additional Inflicted Damage by the Proc remains constant.

 

This of course ignores the effects of the "Purple Patch" & the like, as other posters have noted previously in this thread.

Edited March 8 by Mister Mass

[ZekeStenzland]

I ran around attacking single hellions and prisoners with fireball (positron's blast %dam + annihilation %-res) and lotus drops (achilles heel %-res and fury of the gladiator %-res), watching my combat log.

1. The damage was the same, regardless of whether Achilles Heel procced or not. So it looks like the instant attack damage does not gain the benefit of lowered damage resistance. Later damage (the fire dot of fireball on prisoners) does benefit.
2. I see rolls to hit for the attack. I see 1 notice each time the posi %dam proc goes off. I see 6 notices each time one of the %-res procs goes off. For lotus drops, I see either 0, 6, or 12 notices in the log.
3. At no point is a "to hit roll" mentioned for any of the procs. AFAICT, if the attack hits and the proc goes off, then the proc hits as well. This might only apply to procs in attacks. Not sure if there's a to hit roll for abilities that are autohit for some aspects and have a tohit roll for others.

Quote

IMO: One of the many factors that works against %-Res beyond the marginal increases in (inherent) typed damage is that in order to capitalize on %-Res, there is a chain of several links that has to happen in order for it to work out:

1. The player has to hit with attack containing %-Resistance
2. The %-Res proc has to go off (with a ceiling chance of 90%), and hit
3. The subsequent attacks have to come within the 10 second period
4. The subsequent attacks have to hit
5. The %-Res has to be reapplied (see 1 and 2, above)

 

IMO, this just seems like an unrealistic analysis. All you need to do to capitalize on %-res is attack in the 10sec window. That's it. It's no different than capitalizing on Aim or Build Up.

 

Just like for %dam procs, the %-res proc does not appear to need a separate to hit roll. So items #1 & #2 are just as much drawbacks for %dam as they are for %-res.

 

What you want to compare is how much extra damage you expect on average to get from the %damage proc vs how much extra damage you expect on average while the %-res proc is up. That's it. 

 

I'm a fire/ninja blaster. If I swap out Achilles Heel on Lotus Drops for the Lady Grey %dam proc, Mids says my damage should go from ~217 to ~252. On average. The 35pt difference means the proc is going off about 50% of the time. Sweet! Of course, if I'd had Achilles Heel in there instead, also going off about 50% of the time, how much bonus damage should I expect? Well, in 10seconds, I can reasonably expect to dish out maybe 2000 damage? 20% extra on 2000 damage from the -damRes is about 400 damage, take off 50% because the proc doesn't go off all the time, so... I can reasonably expect the Achilles Heel to give me ~200 pts extra damage vs Positrons Blast ~35 pts extra damage.

 

As for #3, yep, they do have to be within 10secs. Unless I'm at -40% recharge because Lady Winter's unresistable slows have everything including brawl, kick, and ghost slaying axe on cooldown, finding attacks to throw in that 10sec window is not an issue. Yeah, under the 30sec malta stun I do not have a seamless T1/T2 attack chain. Not something I worry about.

 

As for #4, meh. Between enhancements, set bonuses, kismet, tactics, etc, I normally run at 95% chance to hit unless I'm fighting piles of Ring Mistresses, Death Mages, and Dark Servants. That's what range is for.

 

As for #5... %-damRes doesn't need to be applied continuously. If it falls off, it's no big deal. It's probably better if you aren't keeping it up continuously, tbh. If  you're trying to keep it up all the time, that means you're attempting to proc it again while the debuff is still up. If you do that, it means you didn't gain the full benefit of the previous -damRes.

 

 

There's lots of room for nuance and edge cases here. Depending on the exact mob you're fighting slotting a 50dam IO is a better choice than either, or moving that slot to another power is better. The breakpoints change if you're using a VR %dam -- more PPM and more damage. Since the debuff updates when it procs by anyone, there's less value if you're playing on a team where lots of other people have slotted it as well. But if you're on a team, you get to include the bonus damage from other players as to whether it was worth it. You won't see much benefit if you proc Achilles Heel and then spend the next 10 seconds on healing aura, putting up force fields, clicking glowies, resting, or /em boombox.

 

 

But  my quick, off the cuff, YMMV opinion is that yes, slotting one Achilles Heel is not likely to be a dps loss. I haven't really looked into other -damRes powers, like Weaken Resolve or Surveillance.

 

[shortguy on indom]

Random thoughts:

-Res = +Dam

-Res Proc enhancements = +Dam Proc enhancements = +Dam enhancements

---------------------------------------------------
However a difference in how they work: (all i can think of atm, lol).

 

-Res Procs same as a global +Dam enhancement for all powers on self.
-Res Procs same as a global +Dam enhancement for all powers on teamates.
-Res Proc benefits team objective.
-Res Proc (one enhancement) = Pool Power: Leadership Assault.
-Res Proc does not stack for self.
-Res Proc does not stack for team.

 

+Dam Proc works on single power on self.
+Dam Proc benefits team objective.
+Dam Proc stacks for self.
+Dam Proc stacks for team.

 

+Dam enhancements work on single power on self.
+Dam enhancements have varying benefit depending on power.
+Dam enhancements benefit team objective.
+Dam enhancement benefit works with -Res Proc
+Dam enhancement benefit works with +Dam Proc

----------------------------------------
Possibly an argument could even be made where Procs:  +Rech = -Res = +Dam
----------------------------------------
FWIW: (right or wrong).
I run just a single -Res proc in my builds.
If satisfied I have enough chance-to-hit in every attack, will throw in an +Dam Proc in on rare occasion, but usually
opt for a +Acc.  Would rather be certain my attack hits even if it is lower in damage.
Need my chance-to-hit up there for my controls for survivability reasons, figure I cannot damage them when dead.

[tidge]

9 hours ago, ZekeStenzland said:

1. At no point is a "to hit roll" mentioned for any of the procs. AFAICT, if the attack hits and the proc goes off, then the proc hits as well. This might only apply to procs in attacks. Not sure if there's a to hit roll for abilities that are autohit for some aspects and have a tohit roll for others.

 

 

The need for %proc to also hit doesn't show up in the logs. IIRC this was confirmed through code review. I convinced myself this was the case when I had a power with somewhat poor native accuracy... I replaced a %damage piece with an accuracy IO and via the logs I could see that %damage was hitting more frequently.

 

Conversely... IIRC if you collect enough data from an attack with a %damage where the %proc rate is at the ceiling, you can eventually see that you get fewer %damage records (i.e. below 90%) for each hit. This is somewhat painful, because the %proc ceiling chances are in the slower recharging powers. It was with a Tanker's Taunt (loaded with %Psi from Perfect Zinger, and nothing else) where I first noticed that the %damage occurrences were not quite reaching the levels I expected from the %proc formula.

 

As near as I can tell, not all "autohit" powers are created equal, especially when it comes to %procs. The differences that I observe are in the ones that can accept Accuracy IOs (like Tanker Taunts) and those that cannot.

 

Auras/Toggles/pulsing pseudopets are particularly bad for %damage (or other %procs), mostly because the way the "tick" affects the %proc calculation but also because those types of powers are often not slotted for accuracy or cannot take accuracy enhancements.

 

 

[tidge]

10 hours ago, ZekeStenzland said:

I'm a fire/ninja blaster. If I swap out Achilles Heel on Lotus Drops for the Lady Grey %dam proc, Mids says my damage should go from ~217 to ~252. On average. The 35pt difference means the proc is going off about 50% of the time. Sweet! Of course, if I'd had Achilles Heel in there instead, also going off about 50% of the time, how much bonus damage should I expect? Well, in 10seconds, I can reasonably expect to dish out maybe 2000 damage? 20% extra on 2000 damage from the -damRes is about 400 damage, take off 50% because the proc doesn't go off all the time, so... I can reasonably expect the Achilles Heel to give me ~200 pts extra damage vs Positrons Blast ~35 pts extra damage.

 

Even if we accept the math above, my main argument against the blanket recommendation of %-Resistance is that (for this example above) is this: In order to see the hypothetical benefit of "200 extra damage", that benefit won't show up until the next attack. If the critter gets defeated because of %damage, it isn't necessary to make multiple attacks. It makes no difference if a critter with 10HP left is finished off with an attack that does 2000 or 2200.

 

For giant sacks of HP, this is noticeable. For other critters, less so.

[ZekeStenzland]

3 hours ago, tidge said:

The need for %proc to also hit doesn't show up in the logs. IIRC this was confirmed through code review. I convinced myself this was the case when I had a power with somewhat poor native accuracy... I replaced a %damage piece with an accuracy IO and via the logs I could see that %damage was hitting more frequently.

 

Conversely... IIRC if you collect enough data from an attack with a %damage where the %proc rate is at the ceiling, you can eventually see that you get fewer %damage records (i.e. below 90%) for each hit. This is somewhat painful, because the %proc ceiling chances are in the slower recharging powers. It was with a Tanker's Taunt (loaded with %Psi from Perfect Zinger, and nothing else) where I first noticed that the %damage occurrences were not quite reaching the levels I expected from the %proc formula.

 

I graciously tip my hat towards your superior testing methodology! Still, I'll have you know I spent well over 5 minutes -- nay, over 10 minutes! -- obliterating  arresting hellions and prisoners and combing my combat logs for information. The idea of me being wrong, on the internet, at such a loud volume, after such extensive testing should have been inconceivable!

 

I reserve the right to be grumpy that there's a secret additional to hit check for procs though.

[ZekeStenzland]

2 hours ago, tidge said:

Even if we accept the math above, my main argument against the blanket recommendation of %-Resistance is that (for this example above) is this: In order to see the hypothetical benefit of "200 extra damage", that benefit won't show up until the next attack. If the critter gets defeated because of %damage, it isn't necessary to make multiple attacks. It makes no difference if a critter with 10HP left is finished off with an attack that does 2000 or 2200.

 

For giant sacks of HP, this is noticeable. For other critters, less so.

 

Yep. I agree. Maybe I just run too many itrials where almost everything including the minions are big sacks of HP? I almost never run into mobs that I can just one-shot with Lotus Drops even if a %dam went off. And it's just so glorious to run into a group of mobs, Smoke Flash >> Lotus Drops >> Inferno for about 975 damage at -50% damRes if all the procs go off.

 

I dunno.

 

At least the %-damRes procs I'm using now (one of each) aren't needed for set bonuses and I could easily unslot them for %dam and see if there's a noticeable difference. 

[tidge]

1 hour ago, ZekeStenzland said:

I reserve the right to be grumpy that there's a secret additional to hit check for procs though.

 

Some back of the envelope calculations... if you want to see if there is a slight difference between %proc rates (which could be attributed to a "secret additional check")... let's say 88% instead of 90%... I think in order to have 95% confidence with 80% power if you saw at least 5239 successful procs out of 5842 you could "disprove" the 88% hypothesis (at those levels)

 

This is a somewhat difficult experiment to construct because of what gets logged. We aren't concerned with actual hits, just hits that trigger %damage, so it isn't impossibly tedious.

 

If the secret additional check had a larger effect, say 0.95 (the 'always miss' threshold) times 0.90 (%proc ceiling) = 0.855, then the numbers for a 99%/95% confidence/power are smaller... something like 748/847 to 'disprove' the 0.855 fraction. I don't really like this construction because it's the 0.90 we are trying to disprove... if it were possible to reverse the problem and easily record number of No Procs (which would never be logged!) the experiment takes less data: 75 no results out of 572 possible results would be enough to tell the difference between 1 10% no-proc chance and a 15.5% no-proc chance at 99/95. The logs would have to be translated to go this way.

 

Smaller difference require lots more data... so that last construction would require almost 15K total proc chances to tell the difference between 10% and 11%.

 

I guess take that Tanker to Perez Park and start spamming Taunt with just Perfect Zinger %damage and Recharge IOs? I don't remember if hits of Taunt get logged. If they do, just make sure no other Taunt effect is in play (like an aura, or attacks). If you don't care so much about type two errors the experiment could probably stop after about 400 hits of Taunt. If 48 chances to apply %psi damage turned out to not do damage out of 397 chances, that would be a 90%/50% result.

Edited March 9 by tidge
Brain fog: don't slot Taunt with Recharge! derp

[RelativeQuanta]

9 hours ago, tidge said:

The need for %proc to also hit doesn't show up in the logs. IIRC this was confirmed through code review. I convinced myself this was the case when I had a power with somewhat poor native accuracy... I replaced a %damage piece with an accuracy IO and via the logs I could see that %damage was hitting more frequently.

Any chance you have a source for this?  It's extremely counter intuitive. 

 

 

[tidge]

3 minutes ago, RelativeQuanta said:

Any chance you have a source for this?  It's extremely counter intuitive. 

 

 

 I did some searching, but I couldn't find the connection. I remember being made aware of this in one of the AT subforums and from there being directed to a different forum for details. This was years ago.

[RelativeQuanta]

Ok, here are my latest results after including the Average Adjusted Resistances for level 40+ minions, LTs and bosses from Galaxy Brain's post (linked above)

 

TL; DR: The purple patch is now showing a greater difference.  -Res proc wins against even con enemies and minor damage wins against +4.  Likely due to "double dip" penalty against -Res that drops the duration AND the magnitude.  However, a difference in performance is unlikely to be noticeable for a solo player and still hugely in favor of -Res in groups.

    Boss_resists = {
        "Smashing": .0874,
        "Lethal": 0.0973,
        "Energy": 0.0510,
        "Negative": 0.0002,
        "Psychic": 0.0681,
        "Fire": 0.0435,
        "Cold": 0.0396,
        "Toxic": 0.0966
    }
    LT_resists = {
        "Smashing": .0781,
        "Lethal": 0.0973,
        "Energy": 0.0453,
        "Negative": 0.0013,
        "Psychic": 0.0608,
        "Fire": 0.0234,
        "Cold": 0.0211,
        "Toxic": 0.0954
    }
    Minion_resists = {
        "Smashing": .0314,
        "Lethal": 0.0454,
        "Energy": 0.0135,
        "Negative": 0.0029,
        "Psychic": 0.0897,
        "Fire": 0.0044,
        "Cold": 0.0256,
        "Toxic": 0.0995
    }

    super_boss_resists = {
        "Smashing": 0.5,
        "Lethal": 0.5,
        "Energy": 0.5,
        "Negative": 0.5,
        "Psychic": 0.5,
        "Fire": 0.5,
        "Cold": 0.5,
        "Toxic": 0.5
    }

Changes to the simulation:

1. Simulation tic rate changed from 0.05s to 0.132s to match CoH tic size
2. All powers and damage procs have the correct damage type associated
3. Electron Haze is in again and Neutron bomb is out (Electron Haze has a higher adjusted DPA and I had a mistake on the damage of Neutron Bomb before)
4. I've, yet again, changed when -Res is applied
   1. Power damage is applied first
   2. The -Res is applied after the cast time of the activating power
   3. Why: I didn't want to make the simulation any more complicated by considering things like travel time, multiple hits, and DoTs.  Applying the -Res at the end of the power cast time keeps that debuff time from being "lost."

 

Trial 1

x-ray beam: no procs
Irradiate: achilles_heel 
Cosmic Burst:  achilles_heel
Electron Haze:  Annihilation, PosBst_dmg 
Proton Stream: achilles_heel, Apoc_dmg, ToLG_dmg, ShdB_dmg]
Atomic Blast:  achilles_heel, Arm_dmg
Ground Zero: achilles_heel, ToN_dmg, Obli_dmg, ToLG_dmg
Fire Sword Circle: Obli_dmg, Erad_dmg, SciDer_dmg, FotG 
Cremate: Hec_dmg, ToD_dmg

1. 1,000 trials of a single Mob with 10,000 health, super boss resists, and level shift of +0:
   1. Average mob defeat time 76.78s Standard Error 0.07s
   2. Mob average lifetime debuffed with Achilles' Heel: 91.80% Standard Error 0.23%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 41.87% Standard Error 0.27%
   4. Mob average lifetime debuffed with Annihilation: 0.00% Standard Error 0.00%
2.  1,000 trials of a single Mob with 10,000 health, super boss resists, and level shift of +4:
   1. Average mob defeat time 188.04s Standard Error 0.09s
   2. Mob average lifetime debuffed with Achilles' Heel: 68.73% Standard Error 0.16%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 20.70% Standard Error 0.08%
   4. Mob average lifetime debuffed with Annihilation: 0.00% Standard Error 0.00%
3. 1,000 trials of an x8 group and level shift of +0:
   1. Average mob defeat time 6.87s Standard Error 0.06s
   2. Mob average lifetime debuffed with Achilles' Heel: 24.38% Standard Error 0.32%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 15.39% Standard Error 0.22%
   4. Mob average lifetime debuffed with Annihilation: 3.38% Standard Error 0.09%
4. 1,000 trials of an x8 group and level shift of +4:
   1. Average mob defeat time 19.02s Standard Error 0.15s
   2. Mob average lifetime debuffed with Achilles' Heel: 48.96% Standard Error 0.31%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 17.42% Standard Error 0.20%
   4. Mob average lifetime debuffed with Annihilation: 8.80% Standard Error 0.13%

 

Trial 2

x-ray beam: no procs
Irradiate: achilles_heel 
Cosmic Burst:  achilles_heel
Electron Haze:  Annihilation, PosBst_dmg 
Proton Stream: achilles_heel, Apoc_dmg, ToLG_dmg, ShdB_dmg]
Atomic Blast:  achilles_heel, Arm_dmg
Ground Zero: achilles_heel, ToN_dmg, Obli_dmg, ToLG_dmg
Fire Sword Circle: Obli_dmg, Erad_dmg, SciDer_dmg, FotG 
Cremate: Hec_dmg, ToD_dmg

1. 1,000 trials of a single Mob with 10,000 health, super boss resists, and level shift of +0:
   1. Average mob defeat time 83.42s Standard Error 0.06s
   2. Mob average lifetime debuffed with Achilles' Heel: 0.00% Standard Error 0.00%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 38.45% Standard Error 0.25%
   4. Mob average lifetime debuffed with Annihilation: 0.00% Standard Error 0.00%
2.  1,000 trials of a single Mob with 10,000 health, super boss resists, and level shift of +4:
   1. Average mob defeat time 184.11s Standard Error 0.07s
   2. Mob average lifetime debuffed with Achilles' Heel: 0.00% Standard Error 0.00%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 21.96% Standard Error 0.09%
   4. Mob average lifetime debuffed with Annihilation: 0.00% Standard Error 0.00%
3. 1,000 trials of an x8 group and level shift of +0:
   1. Average mob defeat time 6.77s Standard Error 0.06s
   2. Mob average lifetime debuffed with Achilles' Heel: 6.30% Standard Error 0.14%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 15.34% Standard Error 0.22%
   4. Mob average lifetime debuffed with Annihilation: 3.35% Standard Error 0.09%
4. 1,000 trials of an x8 group and level shift of +4:
   1. Average mob defeat time 18.29s Standard Error 0.14s
   2. Mob average lifetime debuffed with Achilles' Heel: 14.05% Standard Error 0.18%
   3. Mob average lifetime debuffed with Fury of the Gladiator: 18.65% Standard Error 0.21%
   4. Mob average lifetime debuffed with Annihilation: 7.90% Standard Error 0.12%

 

Conclusion

1. One change from earlier, now -Res wins by 4s on even con single target but loses by 4s on +4 single target. 
2. The purple patch appears to be the downfall of -Res, likely due to its "double dip" penalty on duration and magnitude.

 

Notes:

1. RelativeQuanta/CityOfHeroes-Damage-Sim: A python damage simulator for City of Heroes (github.com) if anyone wants to take a look at how I did everything

 

Edited March 10 by RelativeQuanta
Fixed conclusion about purple patch effects.

[RelativeQuanta]

16 hours ago, tidge said:

 I did some searching, but I couldn't find the connection. I remember being made aware of this in one of the AT subforums and from there being directed to a different forum for details. This was years ago.

That's unfortunate.  A separate to-hit roll is mathematically the same as a lower PPM count but computationally expensive.  😕  If it exists, it should be removed IMHO.

[aethereal]

14 hours ago, RelativeQuanta said:

Ok, here are my latest results after including the Average Adjusted Resistances for level 40+ minions, LTs and bosses from Galaxy Brain's post (linked above)

 

TL; DR: The purple patch is now showing a greater difference.  -Res proc wins against even con enemies and minor damage wins against +4.  Likely due to "double dip" penalty against -Res that drops the duration AND the magnitude.  However, a difference in performance is unlikely to be noticeable for a solo player and still hugely in favor of -Res in groups.

 

I have no doubt that the first Achilles Heel proc that you have in a group leads to a big gain in group DPS.  However, if you're analyzing it for its results in a group, you have to weight the chance that other people are also running the same proc, and it does not stack.  It's a popular proc!  This clearly mitigates its power, in probably unquantifiable ways since it's hard to know how many people will be packing it.

[tidge]

1 hour ago, RelativeQuanta said:

That's unfortunate.  A separate to-hit roll is mathematically the same as a lower PPM count but computationally expensive.  😕  If it exists, it should be removed IMHO.

 

So based on some searching and visiting City of Data (which has had some changes in the way data is represented over the recent years) I think the ToHit roll for %procs was made part of the %proc themselves with the primary(?) purpose of making sure that %damage didn't automatically hit when slotted into an autohit power.

 

I'm ok with this, as there is a often stated feeling that %damage is overpowered to the point of needing to be revisited(*1). Requiring a ToHit check for %damage is a completely reasonable mechanic (on top of the %proc formula), and I don't think it is unreasonable to ask players to slot powers for Accuracy if they want to leverage this extra source of potential damage.

 

(*1) For the record: I don't believe %damage is overpowering; I consider the opportunity to slot %damage one of the more democratizing design choices the game offers.