Magician %/s (KMS FLY 2 & GMS Dawnveil)

[DeanNim]

Im still sceptical about the % differences between f/p and evans. both of my evan and f/p archmage have about the same range (10~11k max range) when fully equipped but it seems like my evan is doing alot more damage per second. Am i doing it wrong or??

My illusion does almost 60k crit and 30k non crit, paralyze do a little less, never breaking 60k, but i'd imagine with arcane aim and my stacked DoTs it would still be out damaging my evan but bossing at hilla prove otherwise.

[JoeTang]

Im still sceptical about the % differences between f/p and evans. both of my evan and f/p archmage have about the same range (10~11k max range) when fully equipped but it seems like my evan is doing alot more damage per second. Am i doing it wrong or??

My illusion does almost 60k crit and 30k non crit, paralyze do a little less, never breaking 60k, but i'd imagine with arcane aim and my stacked DoTs it would still be out damaging my evan but bossing at hilla prove otherwise.

Maybe you're looking at the wrong tables.

[DeanNim]

Maybe you're looking at the wrong tables.

So my fp is supposed to be weak compared my evan?

[JoeTang]

So my fp is supposed to be weak compared my evan?

If you're not using Mist Eruption properly or don't have Speed Infusion, yes.

[DeanNim]

If you're not using Mist Eruption properly or don't have Speed Infusion, yes.

i try not to use mist eruption alot because of its delay and the constant need to keep up all of my DoTs. i usually just keep mist on and explode it once i feel like its going to dissipate or explode it once i feel it can ohko the boss.

oh well i guess i have to move back to my evan. but evans hyperskill suck.

[Arcana27961]

i try not to use mist eruption alot because of its delay and the constant need to keep up all of my DoTs. i usually just keep mist on and explode it once i feel like its going to dissipate or explode it once i feel it can ohko the boss.

oh well i guess i have to move back to my evan. but evans hyperskill suck.

Mist Eruption only has a 10 second cooldown (5 seconds with hyper skills), and DoT lasts at least 20 seconds for F/Ps. It's really not that big of a deal. :p

[DeanNim]

Mist Eruption only has a 10 second cooldown (5 seconds with hyper skills), and DoT lasts at least 20 seconds for F/Ps. It's really not that big of a deal. :p

i panic alot :P Usually i would recast all DoTs after 5 ~ 6 paralyzes >_>

[Chilly]

My brother didn't get much time to test it out, but he thinks there is some sort of chaining mechanism with Glacier Chain (chain, hehe) and Chain Lightning (chain again, hehe). To be sure, what delays are you using for GC and CL? He did some tests and GC has a massive delay if used in succession.

He didn't get multiple trials for a statistically sound result but preliminarily, a full minute of spamming results in:

• 40 casts of GC
• 74 casts of an alternating GC/CL
• 92 casts of CL

He has the Attack Speed Increase Inner Ability

Spoiler

He's fuming because his math indicates CL spam is better than GC/CL and bounds (2X) better than GC spam. He emailed me numbers specific only for his case/stats:

%Dmg. * #Hits * Crit. Multi. (Min=Max for him) * Crit. Rate * Approx. Mult. of the Hyper Skill (20%, CL only)

%CL/Cast: 200% * 7 * 1.5 * .96 * 1.1 = 2217.6%

%GC/Cast: 340% * 7 * 1.5 * .71 = 2534.7%

• 101388%/min GC Spam = 2534.7% * 40
• 176057%/min GC/CL alternating = (2543.7% + 2217.6) * (74/2)
• 204019%/min CL Spam = 2217.6% * 92

I know these number are more rudimentary than you are used to dealing with, highly unoptimized with many other excluded variables such as the 40% Ignore Def., DoT, capping, etc. He's also working on determining if you can wedge a Teleport Mastery after GC because he indicates that it's not nearly as easy as doing so after an CL.

He's asking if yourself or more importantly any other I/L can confirm the behavior of these skills.

[Bribery]

This I/L also alternates between GC and CL rather than spamming one or the other:

[hadriel]

Try GC followed by twack? Should tell if there is an additional hidden delay or not. The base delay for GC is 1050ms though... hmm...

Hadriel

[JoeTang]

My brother didn't get much time to test it out, but he thinks there is some sort of chaining mechanism with Glacier Chain (chain, hehe) and Chain Lightning (chain again, hehe). To be sure, what delays are you using for GC and CL? He did some tests and GC has a massive delay if used in succession.

He didn't get multiple trials for a statistically sound result but preliminarily, a full minute of spamming results in:

• 40 casts of GC
• 74 casts of an alternating GC/CL
• 92 casts of CL

He has the Attack Speed Increase Inner Ability

Spoiler

He's fuming because his math indicates CL spam is better than GC/CL and bounds (2X) better than GC spam. He emailed me numbers specific only for his case/stats:

%Dmg. * #Hits * Crit. Multi. (Min=Max for him) * Crit. Rate * Approx. Mult. of the Hyper Skill (20%, CL only)

%CL/Cast: 200% * 7 * 1.5 * .96 * 1.1 = 2217.6%

%GC/Cast: 340% * 7 * 1.5 * .71 = 2534.7%

• 101388%/min GC Spam = 2534.7% * 40
• 176057%/min GC/CL alternating = (2543.7% + 2217.6) * (74/2)
• 204019%/min CL Spam = 2217.6% * 92

I know these number are more rudimentary than you are used to dealing with, highly unoptimized with many other excluded variables such as the 40% Ignore Def., DoT, capping, etc. He's also working on determining if you can wedge a Teleport Mastery after GC because he indicates that it's not nearly as easy as doing so after an CL.

He's asking if yourself or more importantly any other I/L can confirm the behavior of these skills.

hm. I was never made aware of this, but apparently, Glacial Chains has a spam delay of 1500ms, and an attack delay of 1050ms. I was under the impression it could be fully spammed with no penalty. It's something we've never had tested since it was not possible to spam Glacial Chains indefinitely without Hyper Skills (unless you had a retarded amount of -CD?).

This means after the 1050ms of attack delay (down to 810ms at speed cap), there is still a small 450ms gap (690ms at speed cap) leading to the 1500ms where Glacial Chains can't be cast again, but another action can be performed. This explains why someone would want to alternate with Chain Lightning. My table currently uses a 1050ms attack delay with no spam delay. I'm currently in the process of updating all of my calculations to the new GMS patch, as well as the Unlimited Defense Formula for KMS, and I should get to this one after I finish the Warriors. It is a rather boring process though and I'm not hurrying.

In terms of theory, it should be beneficial to add Glacial Chains every other attack unless you're hitting the damage cap, seeing as it's the superior DPS on its own merit without the spam delay, but the spam delay prevents you from, well, spamming it. I just noticed my tables include Teleport Mastery for Glacial Chains for both charts, not just the Teleport spoiler, and I'll fix that too.
Ask him to hold down up, and mash the Teleport key while attacking and see if it works. I would imagine it to be rather difficult though if you're alternating between two skills regardless of if it's Glacial Chains + Chain Lightning, or Energy Bolt + Magic Claw.


As for the math:
CL: 200 * 7 * 1.2 * (1+0.96 * 0.5) / 0.78 = 2380%/s
CL+GL: (200 * 7 * 1.2 * (1+0.96 * 0.5) + 340 * 7 * (1 + 0.71 * 0.5)) / (0.78 + 1.050) = 3120%/s

With say 300% total damage
CL: 200 * 7 * 4.2 * (1+0.96 * 0.5) / 0.78 = 11156%/s
CL+GL: (200 * 7 * 4.2 * (1+0.96 * 0.5) + 340 * 7 * 4 * (1 + 0.71 * 0.5)) / (0.78 + 1.050) = 11804%/s

Attack Speed should scale linearly, so it's irrelevant...kind of. Assuming you take Chain Lightning Reinforce, Bonus Attack, and Glacial Chain Bonus Attack and Cooltime Reduce, these are the numbers.
It's possible that the spam delay may still be active if you have too much attack speed; i.e. at speed cap:
Glacial Chains = 810ms
Chain Lightning = 600ms
810+600 = 1410ms, so there's 90ms wasted waiting if you try to use Glacial Chains immediately, so a more accurate representation of that attack rate would be 1500ms, rather than 1410ms. In this case, I believe it is very slightly better to do GL + 2x CL.


EDIT:

I've updated both posts now, let me know if there's any discrepancies. The Glacial Chain table alternates between GL and CL, and I think at Fastest 2, it does GL + 2CL. The most notable difference is against non-0 defense.

The description I've added is wrong, and I'm too lazy to change it right now. It should read

Hyper (Glacial Chain): Uses Lightning Sphere (assumed 270ms attack speed) and the following: Chain Lightning - Reinforce, Chain Lightning - Bonus Attack, Glacial Chain - Bonus Attack, Glacial Chain - Cooltime Reduce, Teleport Mastery - Reinforce

For both tables.

[hadriel]

Confirming that Advanced Dark Chain suffers from the same delay issue.

Also, can I check with you the chain attack penalty you're using for Luminous's Reflection?

Hadriel

[JoeTang]

Confirming that Advanced Dark Chain suffers from the same delay issue.

Also, can I check with you the chain attack penalty you're using for Luminous's Reflection?

Hadriel

-10% per target, so it does
100%, 90%, 80%, etc.

[hadriel]

-10% per target, so it does
100%, 90%, 80%, etc.

Doesn't that mean that the last target (8th target) receives... 30% damage? That is far, by far, too low compared to anything I've seen. I'm asking this because I was doing calculations to see how much benefit I would receive from Hyper skill: Reflection - Spread. I observed a 0.95 multiplier per successive target, so I wanted to check with you how you arrived at that penalty. Would you like screenshots for additional study?

Hadriel

[JoeTang]

Doesn't that mean that the last target (8th target) receives... 30% damage? That is far, by far, too low compared to anything I've seen. I'm asking this because I was doing calculations to see how much benefit I would receive from Hyper skill: Reflection - Spread. I observed a 0.95 multiplier per successive target, so I wanted to check with you how you arrived at that penalty. Would you like screenshots for additional study?

Hadriel

I tested this myself and observed it in some videos pre-Unlimited. It's especially noticeable for pre-Unlimited damage cap. In KMS, you would see 999999, 899999, 799999, etc. It could be different in GMS for all I know, or even changed in the most recent version of KMS.


EDIT: I just retested it, it's still -10% in KMS.

[hadriel]

I tested this myself and observed it in some videos pre-Unlimited. It's especially noticeable for pre-Unlimited damage cap. In KMS, you would see 999999, 899999, 799999, etc. It could be different in GMS for all I know, or even changed in the most recent version of KMS.


EDIT: I just retested it, it's still -10% in KMS.

I'll take some SS tomorrow for us to verify. It's midnight here and I'm dead beat. Cheers.

Hadriel

[hadriel]

@JoeTang; Data's in.

Damage range

Spoiler

First Screenshot

Spoiler

Second Screenshot

Spoiler

Calculations:

Spoiler

Damage Range: 38192 - 39783

Reflection: 420%
Dark/Light: 105%


13% Boss damage (10% DS link, 3% Kanna card)
Minimum Crit damage: 120% (base) + 16% (lvl 31 mastery) + 6% (XBM card) + 20% (Crescent Shining Rod) > 150% (max crit damage)
Therefore all crits are 150%
MDRignore = 40% (Arcane Pitch) + 10% (Beginner skill) + 3.5% (lvl 37 ambition) = 53.5%


Mob is Official Knight B, MDR = 35%


Crit damage for Reflection base = dmg * 4.2 * 1.05 * 1.13 * 1.5 * (1 - 0.35*(1 - 0.535))
= 239020 - 248978
Avg = 244000 (very small rounding)


For Bonus damage from Sunfire, it will be half the standard damage
= 119510 - 124489
Avg = 122000 (very small rounding)


Using Sunfire average damage (because it's higher up and therefore fades the last):
If damage penalty is -10% and additive [i.e. 100%, 90%, 80% etc.] then we expect
{122000, 109800, 97600, 85400, 73200, 61000, 48800, 36600}


If damage penalty is -5% and multiplicative [i.e. 100%, 95%, 90.25% ...] then we expect
{122000, 115900, 110105, 104600, 99370, 94401, 89681, 85197}


Should also note that a regression of all the critical hits collected as a single data set, and after taking logs, showed a regression of -0.0482, which sounds reasonably close to a 0.95 multiplier.

Differences between KMS and GMS I guess? o.O
This won't significantly change DPS numbers I guess, only utility and significance of Reflection - Spread.

Hadriel

[RADRaze2KX]

JoeTang, thank you for all your hard work, this is an astounding feat and mages everywhere appreciate it.

I'd like to suggest that once BaM's receive their Battle King Bar hyperskill, using teleport and twister spin reduces the DPS as they're severely outmoded. The most efficient DPS then becomes a spammy Finishing Blow/Battle King Bar combination.

[JoeTang]

JoeTang, thank you for all your hard work, this is an astounding feat and mages everywhere appreciate it.

I'd like to suggest that once BaM's receive their Battle King Bar hyperskill, using teleport and twister spin reduces the DPS as they're severely outmoded. The most efficient DPS then becomes a spammy Finishing Blow/Battle King Bar combination.

Both tables use no Twister Spin in the Hyper spoiler. This is a bug and I'm fixing it now. I would remove it completely but people should know how big of a difference the two are.

[hadriel]

Not to mention that Teleport is highly latency-dependent... sigh.

Hadriel