|
|
I'm trying to think about when it's better to have faster attackspeed than cd of the trigger due to hit or critrate.
For CoC:
100% efficiency is having attacktime=cooldown and 100% critrate.
If you have an attacktime higher (=slower) than the cooldown, the average loss = non-critrate * attacktime + (n+1)*difference in cooldown and attacktime, where n is the number of attacks since the last crit. (=noncrit/crit right?)
Due to these scenarios:
(1) Crit-crit: loss = time from cooldown reset to the next attack
(2) Crit-hit: loss = attacktime + time from cooldown reset to the next attack
If you have an attacktime lower (faster) than the cooldown however, things change.
to help me I have two scenarios.
(3) Crit-Critfail (due to cd). loss = time from cooldown reset to next attack
(4) Crit-Hit: loss = time from cooldown reset to next attack
Average loss = critrate * time from cd to next attack + non-critrate * time from cd to next attack = time from cd to next attack = 2*cd - (cd+2*(attacktime-cd))
n is not used here since the total of (3) and (4) is always 1 time from cd to next attack?
however, theres also the possible scenario:
(5)Crit-Critfail-Hit. loss = time from cooldown reset to second attack after.
But does it matter???
|
Ideal are skills like molten strike and barrage that hit over a long enough period of time that a critroll fail is covered by a delayed hit from the prior attack. Stuff like chain (with projectile triggers) etc works well here too to get that extra proc.
with skills like cyclone you basically just want your attacktime to be barely slower than .15 and crit as close to cap as possible (keep in mind that diamond flask functionally overrides crit cap)
|
On December 06 2018 05:28 Sn0_Man wrote:
Ideal are skills like molten strike and barrage that hit over a long enough period of time that a critroll fail is covered by a delayed hit from the prior attack.
Statements like this keep coming up and make sense in theory, but if crits are rolled ONCE at the beginning of the ability rather than on each individual projectile I don't think it matters? I feel like bo1b mentioned a few pages back that MS crit is calculated once on the initial hit, then all the balls crit (or dont) based on that initial hit. If that's true then CoC will still feel "chunky", despite so many different balls/projectiles hitting at such a "high rate" (especially if the balls from a single attack all land at the same time too).
If crits are calculated separately for every projectile then yeah that's great, smoothes out the RNG and easy way to get high APS. But I'm not convinced that is how it works?
I guess one single MS "crit" technically could proc CoC twice, the initial hit and then whenever the fountain lands. But I feel like that would get smooth out with high attack rate. Basically my point being that MS doesn't offer anything unique to CoC builds over any other single-hitting melee attack depending on how high your attack speed can get. But I'm still not sure.
|
What's unclear about
a critroll fail is covered by a delayed hit from the prior attack.
?
The point is you attack at whatever rate approaching 7aps, but each attack hits repeated over a say .3s duration. That means that even repeated failed critrolls don't matter because the last successful crit is still hitting away and granting triggers.
Yes attacks are "all or nothing" on the crit roll (although it's more complex than that due to things like defender accuracy rating variations and possible variations in defender chance to receive a critical strike due to stuff like assassin's mark) but an attack that hits multiple times can still proc CoC multiple times if it hits over more than .15s total which means that it can bail you out of failed critrolls. It basically theoretically multiplies your APS in calculations above, but it's dependant on how long it keeps hitting, which is very hard to calculate in most situations.
|
Yeah that makes sense. I sort of edited that in last second. Basically a single "crit" with MS actually could count twice for CoC depending on how attack speed and the proc rate line up. So that would smooth things out, just a complicated function of attack speed and CoC proc rate and the ball hang time and stuff. But how is MS different in that regard than any other ability that hits more than once (say Earthquake, as a shitty example)
I've definitely heard people describing it (not here) as if "there are so many balls so many more chances at a crit (or many more rolls) means the RNG is smoothed out", which is definitely not the case.
|
On December 06 2018 05:02 Duka08 wrote:
Did anyone use the new Flame Dash extensively in 3.4? Is it actually a reasonable alternative to Shield Charge or Leap Slam or whatever now? Or are the charges still limiting even with the shorter CD and stuff
fortify  (
I might be stupid but:
How would you sustain mana?
Too much mana reserved
8 aps is too high
Dropping haste, changing middle right cluster to attackspeed to drop it fixes two of those.
btw i just got spammailed about poe psw reset :E
|
I accidentally placed empower instead of enlighten, the attack speed I need is something like 7.02 aps due to server tick rate, it will be far easier to achieve that through a targetted faster attacks + quality than anything else, I want the most aps possible for shield charge. I don't think mana will be a problem tbh, if it is against bosses I'll cwdt warlords.
|
Is your only source of life leech the helm corruption, and no source of mana leech? I wonder how you would sustain without the helm corruption. Maybe would have to use a life leech gem in the CoC setup or something?
|
On December 06 2018 05:24 Duvon wrote:
I'm trying to think about when it's better to have faster attackspeed than cd of the trigger due to hit or critrate.
For CoC:
100% efficiency is having attacktime=cooldown and 100% critrate.
If you have an attacktime higher (=slower) than the cooldown, the average loss = non-critrate * attacktime + (n+1)*difference in cooldown and attacktime, where n is the number of attacks since the last crit. (=noncrit/crit right?)
Due to these scenarios:
(1) Crit-crit: loss = time from cooldown reset to the next attack
(2) Crit-hit: loss = attacktime + time from cooldown reset to the next attack
If you have an attacktime lower (faster) than the cooldown however, things change.
to help me I have two scenarios.
(3) Crit-Critfail (due to cd). loss = time from cooldown reset to next attack
(4) Crit-Hit: loss = time from cooldown reset to next attack
Average loss = critrate * time from cd to next attack + non-critrate * time from cd to next attack = time from cd to next attack = 2*cd - (cd+2*(attacktime-cd))
n is not used here since the total of (3) and (4) is always 1 time from cd to next attack?
however, theres also the possible scenario:
(5)Crit-Critfail-Hit. loss = time from cooldown reset to second attack after.
But does it matter???
I'm having alittle trouble parsing the math at the moment, but im just gonna point out here that the value of attack speed is not going to be linear, it increases linearly until attack time is the same as the cooldown, then the value is gonna plummet to nearly half because on successful crits the next attack that could trigger it is twice your attack time and if you don't crit your next attack is basically 1 cd away anyway so you don't gain very much. After that it will increase again because obviously attack time being half a cd is the same as being equal to the cd but you get 2 chances to crit.
I'm not sure any of that was coherent but what im getting at is theres no one answer to is it better to have lower attack time than the cd, but you can find out whether its better for a specific attack time.
|
Hmm so with new ele flasks (ruby/topaz/sapphire) it's possible to have over 100% reduced elemental damage taken from hits, but very conditional (requires pathfinder, flask uptime, stationary, fortify, 2x nebuloch and at least 6 endurance charges). Also still quite vulnerable to DoT (such as what nebuloch is dealing you) but an interesting thought exercise nonetheless.
|
On December 06 2018 06:27 Sn0_Man wrote:
Hmm so with new ele flasks (ruby/topaz/sapphire) it's possible to have over 100% reduced elemental damage taken from hits, but very conditional (requires pathfinder, flask uptime, stationary, fortify, 2x nebuloch and at least 6 endurance charges). Also still quite vulnerable to DoT (such as what nebuloch is dealing you) but an interesting thought exercise nonetheless.
Looks like you can do it with 2x Ichimonji instead of Nebuloch if you stack both flask effect and fortify effect (requires tons of gear, though, Elder belt, Horror crafted boots and amulet, balanced resists for Wise Oak, still have to be stationary). Might be preferable so you're not dealing DoT to yourself that you can't deal with.
Elemental DoT appears in the wild often enough that it still won't be enough to be functionally immortal, though.
|
On December 06 2018 06:25 ThaddeusK wrote:Show nested quote +On December 06 2018 05:24 Duvon wrote:
I'm trying to think about when it's better to have faster attackspeed than cd of the trigger due to hit or critrate.
For CoC:
100% efficiency is having attacktime=cooldown and 100% critrate.
If you have an attacktime higher (=slower) than the cooldown, the average loss = non-critrate * attacktime + (n+1)*difference in cooldown and attacktime, where n is the number of attacks since the last crit. (=noncrit/crit right?)
Due to these scenarios:
(1) Crit-crit: loss = time from cooldown reset to the next attack
(2) Crit-hit: loss = attacktime + time from cooldown reset to the next attack
If you have an attacktime lower (faster) than the cooldown however, things change.
to help me I have two scenarios.
(3) Crit-Critfail (due to cd). loss = time from cooldown reset to next attack
(4) Crit-Hit: loss = time from cooldown reset to next attack
Average loss = critrate * time from cd to next attack + non-critrate * time from cd to next attack = time from cd to next attack = 2*cd - (cd+2*(attacktime-cd))
n is not used here since the total of (3) and (4) is always 1 time from cd to next attack?
however, theres also the possible scenario:
(5)Crit-Critfail-Hit. loss = time from cooldown reset to second attack after.
But does it matter??? I'm having alittle trouble parsing the math at the moment, but im just gonna point out here that the value of attack speed is not going to be linear, it increases linearly until attack time is the same as the cooldown, then the value is gonna plummet to nearly half because on successful crits the next attack that could trigger it is twice your attack time and if you don't crit your next attack is basically 1 cd away anyway so you don't gain very much. After that it will increase again because obviously attack time being half a cd is the same as being equal to the cd but you get 2 chances to crit. I'm not sure any of that was coherent but what im getting at is theres no one answer to is it better to have lower attack time than the cd, but you can find out whether its better for a specific attack time.
Ofcourse. Thats what makes it interesting. But first i need to see if my logic is ok, which I suspect its not.
|
On December 06 2018 07:16 ASoo wrote:Show nested quote +On December 06 2018 06:27 Sn0_Man wrote:
Hmm so with new ele flasks (ruby/topaz/sapphire) it's possible to have over 100% reduced elemental damage taken from hits, but very conditional (requires pathfinder, flask uptime, stationary, fortify, 2x nebuloch and at least 6 endurance charges). Also still quite vulnerable to DoT (such as what nebuloch is dealing you) but an interesting thought exercise nonetheless. Looks like you can do it with 2x Ichimonji instead of Nebuloch if you stack both flask effect and fortify effect (requires tons of gear, though, Elder belt, Horror crafted boots and amulet, balanced resists for Wise Oak, still have to be stationary). Might be preferable so you're not dealing DoT to yourself that you can't deal with. Elemental DoT appears in the wild often enough that it still won't be enough to be functionally immortal, though.
well there's always the moving part too, kinda need to do that to be useful. Maybe cast on crit lightning warp 
Are you calculating alchemist flasks? I'm not fully clear on how that increased effect adds to other sources of increased flask effect. However, I only get 97% reduction with dual ichimonji, neck, full PF stuff etc. must be alchemist flasks.
|
On December 06 2018 07:33 Sn0_Man wrote:Show nested quote +On December 06 2018 07:16 ASoo wrote:On December 06 2018 06:27 Sn0_Man wrote:
Hmm so with new ele flasks (ruby/topaz/sapphire) it's possible to have over 100% reduced elemental damage taken from hits, but very conditional (requires pathfinder, flask uptime, stationary, fortify, 2x nebuloch and at least 6 endurance charges). Also still quite vulnerable to DoT (such as what nebuloch is dealing you) but an interesting thought exercise nonetheless. Looks like you can do it with 2x Ichimonji instead of Nebuloch if you stack both flask effect and fortify effect (requires tons of gear, though, Elder belt, Horror crafted boots and amulet, balanced resists for Wise Oak, still have to be stationary). Might be preferable so you're not dealing DoT to yourself that you can't deal with. Elemental DoT appears in the wild often enough that it still won't be enough to be functionally immortal, though. well there's always the moving part too, kinda need to do that to be useful. Maybe cast on crit lightning warp Are you calculating alchemist flasks? I'm not fully clear on how that increased effect adds to other sources of increased flask effect. However, I only get 97% reduction with dual ichimonji, neck, full PF stuff etc. must be alchemist flasks.
I actually wasn't using Alchemist's flasks. My math is:
-----
-20% base reduction from normal elemental flasks
-10% base reduction from The Wise Oak
..scaled by 94% flask effect:
-26% from Alchemist cluster
-30% from Pathfinder
-10% from Elder belt mod
-8% from Conqueror's Potency
-20% from 2x Ichimonji
...gets you:
-38% from normal elemental flasks
-19% from Wise Oak
=57% total from flasks
-----
-20% base reduction from Fortify
...scaled by 50% fortify effect:
-15% from Rampart cluster
-15% from Horror amulet
-20% from 2x Ichimonji
...gets you:
30% total from Fortify
-----
So the total is:
-57% (flasks)
-30% (Fortify)
-6% (Nature's Boon)
-5% (Gorgon's Gaze, while stationary)
-5% (Horror boots, while stationary)
= 103% total
-----
Looks like with Alchemist's flasks, you could maybe do it without Ichimonji (which is important because it means you might be able to recreate unkillable Poet's Pen).
*edit* No, Alchemist's still isn't enough to do it with Poet's Pen, unless it scales multiplicatively with non-local flask effect (I don't think it does).
*edit 2* Wait not it is just barely enough to hit exactly 100% if you use Alchemist's. Poet's Pen flaskfinder lives (except that you'll die instantly to Shaper beam).
|
ichimonji doesn't grant flask effect
|
On December 06 2018 07:59 Sn0_Man wrote:
ichimonji doesn't grant flask effect
Aaaaah, I thought flask effects were buffs, but now I found the discussions of them changing it way back.
Okay never mind!
|
dear lord imagine the expense of trying to get CDR Horror boots for a poet's pen immortal build lmao
also idk how u get fortify with poet's pen pathfinder
|
On December 06 2018 08:04 Sn0_Man wrote:
dear lord imagine the expense of trying to get CDR Horror boots for a poet's pen immortal build lmao
also idk how u get fortify with poet's pen
Oh yeah also a good point. I was originally thinking about trying to make some kind of Ichimonji Flicker CoC thing with this, which was why I included Fortify, but that definitely doesn't work with Poet's Pen.
Good riddance, honestly.
|
Hmm... Winter Orb damage is pretty weak, it really depends on how good the projectiles overlap. If they mostly or all do, it would be very nice with GMP and things I guess. Multiple Totems scaling to 21% makes it very interesting, however I don't think it's worth it if you have 4+ totems since you also give up an support gem for it.
|
On December 06 2018 05:39 Sn0_Man wrote:What's unclear about ? The point is you attack at whatever rate approaching 7aps, but each attack hits repeated over a say .3s duration. That means that even repeated failed critrolls don't matter because the last successful crit is still hitting away and granting triggers. Yes attacks are "all or nothing" on the crit roll (although it's more complex than that due to things like defender accuracy rating variations and possible variations in defender chance to receive a critical strike due to stuff like assassin's mark) but an attack that hits multiple times can still proc CoC multiple times if it hits over more than .15s total which means that it can bail you out of failed critrolls. It basically theoretically multiplies your APS in calculations above, but it's dependant on how long it keeps hitting, which is very hard to calculate in most situations.
Back on this topic, because I've been thinking about the math in a bit more detail: What is a reasonable crit chance to expect someone to have with a build like this? Including power charges, flasks, etc. I don't play crit builds (let alone CoC) often enough to have a good feel in my head for what a "good end game crit chance" would look like for various ascendancies.
|
|
|
|
|
|
EPT
