I have written a DPS simulator for melee based CM builds to compare a couple different AP spenders, namely WW, Star Pact, Meteor - Liquify, and Mistral Breeze. The simulator is written in MATLAB, which is not a freeware software program, but it uses a language similar to C++ and the m file can be opened as a text document so anyone who feels up to it can use the code to convert it to some other computing language.
The m file can be found here. Current version is 1.01. Any changes since v1.00 will be included at the end of the mfile and included at the end of this post.
Please post any ideas or suggestions you have, or any errors you see.
Features
Uses 1.0.5 spell coefficients and crowd control dimishing returns to simulate combat in D3 against 1 or more mobs for a duration that is easily adjusted. Default is usually 60 seconds.
Approximates mob movement by using an uptime coefficient that checks for each damage tic against the given probability. I've set the defaults to 100%, 90%, and 50% for WW, Meteor (both runes), and Mistral Breeze respectively.
Attempts to simulate button mashing using a Reaction Time (default is 0.01s) that assumes you press whatever button you want to at that time. It only allows spells that use a global cooldown to be used at specific intervals related to your attack speed, namely Frost Nove and the AP spenders. DS and Chain reaction can be used at any time since they do not trigger a global cooldown (i.e., a weapon strike, for those not used to the WoW lingo).
Allows for lots of variety in input ranging from basic gear stats like Crit%, AP on Crit, attack speed, max AP, various runes and passives like DS - prism and evocation, and going all the way to mob attack speed and approximate spell uptimes.
The simulator uses a Monte Carlo based approach to simulate random events over the duration specified, then repeates several times (default is 1000 simulations per skill).
The final results that are tracked are the Effective DPS multiplier of your skills and gear, % uptime of DS, the % of time the mobs are kept frozen, and the average LoH retuned per second. The Effective DPS multiplier is simply the number you multiply your char sheet dps by to get your effective dps. If you have 100k char sheet dps and your multiplier is 3.5, you would actually do around 350k dps with that skill set up and those base stats. The average LoH returned per second is basically what it sounds like. Whatever your total LoH is, you multiply it by that number to get your approximate life gain per second. For example, if the number if 3 and you have 1000 LoH you gain about 3000 life per second on average.
The simulator also outputs the % damage done by each skill for each build.
Assumptions
Mob swing timer = 1s and swing timer is reset when mob is frozen. These parameters are easily edited though.
White attacks have a coefficient of 1 and White Attacks are always performed if FN is on CD and there is insufficient AP for the AP spender. I have confirmed that the LoH coefficient for white attacks is 1.
Conduit uses spell coefficients so on a successful hit you always gain back AP = APoC*Spell Coefficient.
Evocation, when combined with CD reduction rune on Frost Nova, reduces the cooldown to 9*(1-0.15) = 7.65s.
Spell Priority is TW(if down) > DS >= FN > AP spender >= CR. CR is used on CD whenever there is enough AP, but only after spending AP on the big AP cost. Since it does not trigger a global CD or swing timer, it effectively can be used at the same time as DS and AP spender/FN. AP spender is refered to as the Prim spell in the code, for primary.
CR rune for Explosive Blast is used. Options for other runes should be added soon.
Planned Additions
I want to add functionality for additional EB runes.
Code Updates
v1.01 Changed FNdf to FNbc since I had incorrectly put deep freeze as the rune that adds 15% damage while frozen, but it is actually bone chill
v1.02 Added WWcmcoef and WWapoccoef since the CM procs and APoC returns are higher than otherwise expected. Default values are 2 for each based on best guess from limited testing.
Here are some simulation results combined with in game data to show how well my simulator is working. For each case I used the skills indicated in the tables. For the case of pinpoint barrier, I added 5 CC to the base stats. I used 114 max AP for Storm Armor and 94 without. I did a total of 3 case studies with varying overall attack speed. I used the same gear and just swapped between a 1.2, 1.5, and 1.77 APS weapon with 52% IAS on gear.
Base Stats:
APoC=19;
CC=0.505;
MaxAP=114; (94 with prismatic or pinpoint barrier)
APregen=10;
DSprism=0;
DSds=0;
Evocation=1;
CB=1;
Conduit=0;
SimDuration=VARIED;
Mobs=1;
AS=VARIED;
MobAttackRate=1;
DSabsorbs=1;
MobAttackDelay=MobAttackRate;
Simulations=300;
WWuptime=1;
ReactionTime=0.01;
Case 1: AS = 2.6904
Storm Armor
Shocking Aspect
Pinpoint
Shards
DPS Mult
Rel DPS
Simuated dps
Rel Simulated DPS
3.01
1
3.09
1.00
x
x
3.53
1.17
3.57
1.16
x
3.42
1.14
3.43
1.11
x
3.04
1.01
3.19
1.03
x
x
3.82
1.27
3.84
1.24
x
3.47
1.15
3.49
1.13
x
x
4.60
1.53
4.69
1.52
x
x
x
4.85
1.61
5.02
1.62
Case 2: AS = 2.28
Storm Armor
Shocking Aspect
Pinpoint
Shards
DPS Mult
Rel DPS
Simuated dps
Rel Simulated DPS
2.95
1
2.83
1.00
x
x
3.48
1.18
3.42
1.21
x
3.12
1.06
3.14
1.11
x
3.04
1.03
3.1
1.1
x
x
3.58
1.21
3.59
1.27
x
3.20
1.08
3.29
1.16
x
x
4.20
1.42
4.43
1.57
x
x
x
4.56
1.54
4.74
1.67
Case 3: AS = 1.824
Storm Armor
Shocking Aspect
Pinpoint
Shards
DPS Mult
Rel DPS
Simuated dps
Rel Simulated DPS
2.69
1
2.61
1.00
x
x
3.25
1.21
3.10
1.19
x
2.90
1.08
2.88
1.10
x
2.74
1.02
2.81
1.08
x
x
3.72
1.38
3.43
1.31
x
3.05
1.13
3.17
1.21
x
x
4.02
1.49
4.13
1.58
x
x
x
4.17
1.55
4.41
1.69
The dps columns are as follows:
DPS Mult = the number you multiply your char sheet dps by to get your effective dps
Rel DPS = relative fractional dps gain from the skill sets, compared to the all defensive build
Simulated dps = DPS multiplier obtained from the simulator. This should compare directly to the DPS Mult value
Rel Simulated DPS = relative fractional dps gain from the skill sets estimated by the simulator
As you can see the simulator has good agreement with the in game testing. It does appear to overestimate the shocking aspect contribution, which can be fixed by tweaking the empirical formula used. The Storm Armor results seem to not agree as well for Case 2 and 3 as Case 1, but I think that was mostly just RNG since the Storm Armor strikes can be pretty sporadic.
For reference, the empirical formula used is
WWsacoef = - 0.8737 * AS + 4.8035
where WWsacoef is the adjusted number of SA procs from WW per CM proc. In other words, if a tic from WW would proc CM, it also procs SA (WWascoef) times. This is only used for WW. CR, DS, and any other skills are all treated the same as CM at a 1:1 ratio.
EDIT: Here are some simulations I ran based on my current tank stats (49% crit, 20 APoC, 2.5308 Attacks per second, with ReactionTime=0.1, SimulationDuration=60, Simulations=100). I used the full SNS build with Deep Freeze rune on FN and no Evocation, assuming I'd use something different but not completely sure what. What I did was change the CC, APoC, and AS by increments I call delta. I only changed one parameter at a time and kept the other two at the value listed above. Delta = 2% CC, 1 APoC, or 0.1 Attacks per second.
You can clearly see the WW breakpoints in the LoH gain per second and the FN uptime. The Effective DPS Multiplier also shows the WW breakpoints in the form of decreasing dps after the breakpoint, then a jump increase at the next breakpoint, etc. This is because increasing your attack speed while between breakpoints will increase your char sheet dps but will not lead to an increase in build dps, therefore the char sheet dps increase is misleading between breakpoints. I would guess that your effective dps would remain the same.
It's also interesting that effective dps actually seems to decrease as you raise Attack Speed past the 2.5 breakpoint, which is unexpected. As you increase Crit, it continues to increase, but APoC only has a very weak impact after the 2.5 APS breakpoint, likely because at that point you gain AP about as fast as you can spend it.
I have a couple comments about the data posted above, and some conclusions I made based on the results.
First, note that the use of Conduit does not apply the 10% bonus damage to the results since that is applied to your char sheet dps.
Raising crit% by 5% for Case 1 will raise your in game dps number but that is not accounted for in the simulator because the simulator is meant to be independent of the in game displayed dps number.
Conclusions
1) Conduit appears worthwhile for MB (5.4% dps increase) for any spec with Magic Weapon. For a WW build, it is about a break even dps alternative to Force Weapon if you already use Glass Cannon. If you don't use glass cannon it is a dps loss (4% gain vs 5% from FW). For meteor builds the gain is only about 3% dps so FW is better. Conduit also slightly improves DS uptime for all builds while also increasing frozen time for the twister specs.
2) Prism > Diamond Shards. DPS is higher, DS uptime is higher, and mobs are frozen more with Prism.
3) 5% crit > 7% IAS. Roughly speaking, if you have a 1.4 APS weapon wiht 7% IAS, your attack speed is around 1.5. To raise your attack speed to 1.6 you would need another 7% IAS. At 40% crit, you're better off gaining 5% crit than raising your attack speed to 1.6 APS. The dps is higher and the DS uptime is higher while the frozen % is roughly the same, with a slight edge to the +CC case. In short, both are good options to raise dps but CC also raises survivability. Also, CC raises your char sheet dps as well as the effective weapon dps while IAS will not raise your effective DPS beyond that shown in the table above.
4) AP on Crit is important. Just raising the amount from 16 to 17 gave a noticable increase in dps (1-2%) and a small increase to DS uptime.
5) Evocation is a promising passive both for dps and survivability since it increases DS uptime considerably. It provides a bit bigger of a dps boost for twister specs while the meteor specs get more survivability, but all specs show an increase in all parameters when Evocation was active.
6) Astral Presence is a large dps increase (5-7%), though smaller survivability increase than Evocation.
7) Calamity seems pretty bad for the build, which makes sense because it does less dmg than CR, though it is a bit cheaper, but also has about half the effective procs as CR and uses a global cooldown whereas a white attack would do less damage but refunds much more AP on a crit.
8) I took a quick look at Slow Time uptime with the new 15s CD in 1.0.5 and so far it seems like practically 100% uptime for all specs using the base stats above. This makes it higher dps gain than Magic Weapon when using the 20% damage rune, but also the option of raising your attack speed for those using twister spec. To approximate the affects of Slow Time, just add 10% of your base weapon speed to AS, so if you have a 1.4 speed weapon, add 0.14 to whatever your char sheet attack speed is, or just activate the ability and note the AS I suppose. For the damage rune, just multiply the end damage by 1.2.
Right now, my expected CM spec would be http://us.battle.net...YXOg!YWg!ZcZZYZ, assuming survivability isn't an issue. If it is, I'd likely replace Glass Cannon with Evocation and Time Stretch with either Conduit or Blood Magic, though I might end up replacing it with Teleport since I really liked using teleport as a CM wiz. Slow Time is used because the uptime is expected to be right around 100% even for low stats like those in the base case.
My reason for picking meteor over a twister build is that my current stats are pretty similar to those used in the base case above. However, it seems that the twister specs are scaling a lot better with most of the stats so with better gear I think a WW or MB spec would be more ideal. About the only appealing attributes of meteor spec is it seems like it would be slightly more effective against ranged mobs and does more dps per cast than WW or MB (though less potential DPS per AP spent than MB) and seems more suited to low attack speed gear sets.
Could you please do a favor to the all of us and create a table for WW based CM wizards with:
- attack speed along one axis
- CC along another axis
- each cell = minimum required APoC to sustain the build
This is a question i know i want an answer to, i'm pretty sure a lot of other wizards would like to know this as well - after all gearing up with all 3 sources of APoC is very limiting to the DPS and very draining financially, it would be far better to have an accurate goal to aim for in terms of APoC on gear. For example, if we are reaching 3.0 APS, would a single source of APoC suffice? Or do we still have to have 3?
Rollback Post to RevisionRollBack
"Only two things are infinite, the universe and human stupidity, and I'm not sure about the former."
If I'm understanding what you're asking, you want me to find the min APoC needed for each combination of CC and AS? I can do that but the main problem is how do you define a sustainable build? Right now my main source of reference is Morphos' video, so are you thinking something like 55% FN uptime? Or is there another % you're thinking of, or parameter, or combination of parameters?
Right now it seems the meteor build is viable with just 17 APoC and 45% crit and 1.5 attacks per second, even with low EHP and Force Armor.
I'm thinking about sustainability in terms of AP based on APS and CC. Common WW build with EB and WW as the 2 AP spenders, what's the required APoC to sustain that build? Using WW, based on your work - the higher the APS the higher the uptime on DS and nova going to be, so the only question is what is the required gear setup in terms of the 3 possible variables: APoC, APS and CC.
I don't think it would be applicable to the meteor build as the APS plays no role in that, which would make WW build more viable at higher gear levels. Quite honestly, other than a slightly different playstyle of the meteor build - i don't really see any other gearing up synergies in there: APoC is only available on 3 sources:
helm - tal's helm or mempo are going to beat any APoC helm flat out in terms of DPS any day
weapon - chantodo's have APoC but they cost 600+ mil for 1.2k DPS, while echoing fury with the same socket in 1.2k range will cost about 100m tops, rare wands with APoC cost a bomb as well and suck in DPS values as most of them lack INT, those that do have stupid costs again
offhand - is the more reliable APoC source, any OH can roll APoC
Hence the entire gearing question arising from this - if we go for a high APS and high CC would a single APoC source be sufficient to run the WW based CM build? Or do we have to inevitably invest into at least 2 sources? Like for example i'm sitting on over 300m gold right now, and i would like to have the ability to run not only archon, but also CM, especially for the ubers, so do i have a chance to run CM with just OH APoC or do i invest into MH with APoC or do i drop a chunk of DPS and get an APoC helm?
According to your research - with, say, 3 APS WW is going to have 2*6*3=36 ticks wouldn't that be fine to run with just 10 APoC or even less? What about 2.5 APS?
Rollback Post to RevisionRollBack
"Only two things are infinite, the universe and human stupidity, and I'm not sure about the former."
What do you mean by sustaining the build? What sort of requirements do you think are necessary? For example, I could put 0 APoC and 0 CC and still run the simulator and get results. How do I indicate the results are bad, which they obviously will be? I can do the same for basically any combination of stats but without a test for what constitutes a sustainable build, I can't give a min value of APoC for a given combination of stats (namely AS and CC).
If I had to pick a requirement I would say something like 55%+ FN uptime, but if you have another threshold or another parameter in mind, I can work around just about anything.
Sustainable in terms of AP = i can go on forever without running out of AP
Do i run out of AP if
CC = 40% APS = 2.0 APoC = 15 ?
CC = 45% APS = 2.0 APoC = 10 ?
CC = 50% APS = 3.0 APoC = 8 ?
etc
Is there a way to run CM build using WW and EB as AP spenders while having only 1 APoC source? What is the minimum required CC and APS for that?
Rollback Post to RevisionRollBack
"Only two things are infinite, the universe and human stupidity, and I'm not sure about the former."
You'll always run out of AP since you can cast WW every 0.5-0.8s depending on AS, you just regenerate it back, so that isn't really a valid check for sustainability. Even with more AP regen you still run out periodically beause you're constantly casting WW and EB, so you mostly just gain more dps. I would need something a bit more concrete to check for viability.
I'm not asking about viability, i'm asking sustainability, that's 2 completely different concepts. If we have 0 APoC on gear and run out of AP within 15 seconds that's clearly not sustainable. If we have enough, whatever the amount is of APoC on gear and we can continue casting without running out of AP while hitting stuff, obviously, to generate back AP - then that is sustainable. Certainly RNG will play a factor no matter, and certainly there will be times when you get no crits for a few seconds (although i don't see how is that possible with say 5-6 ww tornadoes ticking 6 times a second with an even 40% CC), certainly mobs will sometimes move out of the WW and you get no APoC procs.
I don't know am i not explaining the concept clearly? Let me rephrase it this way:
1) Would YOU run CM build with WW with 10 APoC? Yes or no.
2) What is the correlation between the APS and CC versus minimum APoC required on gear to make the build sustainable in a sense of not running out of AP in 10 seconds?
3) Based on your conclusions the WW testing thread - at APS 3.0 WW is going to tick 32 times over 6 seconds, i do not see the point of having 30 APoC in that setup because that would be very very very extremely heavy extra.
Rollback Post to RevisionRollBack
"Only two things are infinite, the universe and human stupidity, and I'm not sure about the former."
I meant for viability and sustainability to be the same. If the build isn't sustainable to an acceptable degree, it's not viable because you'll likely just die repeatedly. The most easy to measure mark of viability to me is how much you freeze mobs, since that relates directly into survivability, so I'll just use that as the stat for comparison. Trying to use some measure of AP doesn't really work well because you'll pretty much always run out of AP very fast to start a fight because it takes time to get enough twisters down to regen AP as fast as you can spend it.
As long as the dps is fairly high, I would assume the build is sustainable because you're ragaining enough AP to frequently cast spells. From there it's just a matter of staying alive to continue to damage the mob, which is where FNuptime comes in, and the higher uptime the easier to survive.
Regarding your questions:
1) Based on a quick simulation, I still see an uptime of 56% for FN with 1.5 APS and 10 APoC and 40% crit. However, over half of the attacks done over a 60s fight are white attacks, and I probably overestimate the effectiveness of white attacks. I really should check to see what the coefficient of a white attack is. If my assumptions for white attacks is correct, the build looks sustainable for this situation.
2) As I said above, I don't think "not running out of AP in first 10s" is a valid criteria because virtually everyone will run out of AP at some point in the first 10s when facing just one mob. If you have 100 AP, you start with FN+CR then spam WW. At 1.5 aps you regen 6.67 AP between casts so WW effectively costs 28.33 AP, meaning you're out of AP after 3 casts+CR. You might have regained a little AP by the time you get your 3rd WW cast, so maybe 4-5 casts, but either way you'll run out of AP while waiting for the tics to rack up and the CR blasts to finish.
3) With 1.5 APS and 45% CC it seems 17 APoC should be enough to sustain the build to an acceptable degree, so I don't see the need for 30 APoC ever. Having more APoC is useful because it increases your effective DPS and the uptime of FN and DS because you can cast more spells
Since the video I used shows sustainability against a single target I'll just use those parameters as a measure, namely 57% FN uptime. To reduce computation time, I'll probably just use 500 simulations and 0.1s reaction time, along with 2-3% intervals for CC and 0.1 APS intervals.
I think what he is asking is at what point does energy twister (or various meteor runes) start paying for themselves and generating excess arcane power that can be funneled into explosive blast through APoC and what builds can accomplish this at what crit, APoC and attack speed values.
The calculation is fairly simple, the expected APoC return of the spell has to be greater than its cost, so:
Cost = Crit * Ticks * coefficient * ApoC
The real question is, how many targets. None of the meteor runes at real gear levels (i.e. less than BiS values) will break even on Arcane Power through APoC or enhanced regeneration on a single target, and it takes close to BiS to do so with two targets, but virtually all of them will generate surplus arcane power on 3 targets with moderate gear levels.
I've been following along with this topic. My question is if i had 2.5 aps with 45 crit chance and only 10 ap on crit. How much would i be improving my dps and nova/diamond skin uptime by increasing my ap on crit to 20, or even 30?
I think what he is asking is at what point does energy twister (or various meteor runes) start paying for themselves and generating excess arcane power that can be funneled into explosive blast through APoC and what builds can accomplish this at what crit, APoC and attack speed values.
The calculation is fairly simple, the expected APoC return of the spell has to be greater than its cost, so:
Cost = Crit * Ticks * coefficient * ApoC
The real question is, how many targets. None of the meteor runes at real gear levels (i.e. less than BiS values) will break even on Arcane Power through APoC or enhanced regeneration on a single target, and it takes close to BiS to do so with two targets, but virtually all of them will generate surplus arcane power on 3 targets with moderate gear levels.
Yeah, that sounds like an easy enough calculation that doesn't really require simulating if you just use averages. With ticks = 2*APS*6 as an approximation and cost = 35 for WW, I'd also add an uptime coefficient of say 0.75 to account for the mob moving out of the WWs so we have
35 = CC * 12 * APS * 0.125 * APoC * 0.75 or 31.11 / APS / CC = APoC
It's pretty easy to make a table with APS on one column and CC as the row entry and solving for APoC for each combination. The problem is with 3 APS and 50% CC you still need 21 APoC. If you assume Prism is active it drops it to 17. This doesn't account for static AP regen, so you can change the cost from 35 to (35-7)-10/APS since you gain 10 AP per sec but you attack once every 1/APS seconds. With that equation you'd still need like 46% attack speed, 1.8 APS, and 24 APoC. That also doesn't account for extra AP to spend on EB, though when you case FN you save the AP from a WW cast that can go towards the EB.
Here's the table I threw together.
If you double the mobs, you halve the APoC needed, roughtly speaking, so it sounds a lot more reasonable with 2 mobs, i.e., Uber boss pairs.
Also, while this calculates the APoC desired to break even with WW, that doesn't mean it's the min APoC needed to sustain the build. I'd actually say it's more like the max APoC needed. As I said before, the vid clearly shows sustainability vs 1 mob with only 45% crit, 1.5 APS, and 17 APoC, whereas the table says you'd need 28 APoC, and he was using a meteor build, rather than a WW build.
If I find the time, I'll run some simulations to come up with a similar (though smaller) table based on a FN uptime of around 57%.
I've been following along with this topic. My question is if i had 2.5 aps with 45 crit chance and only 10 ap on crit. How much would i be improving my dps and nova/diamond skin uptime by increasing my ap on crit to 20, or even 30?
PPS = Average full LoH procs per second, so if your PPS = 2, you gain 2x your total LoH each second.
I only ran 1000 simulations, so there's going to be some varience to the results, but hopefully they give you some good information. Also, sorry for being a bit lazy and not rounding the results like I did above.
So you gain quite a bit of dps but you shouldn't really have any problems with FN locking mobs even at 10 APoC. Same thing with DS uptime, but that all depends on your reaction time or how fast you can mash buttons.
Just for completion, and because the run times are only like 20s, here's the results with ReactionTime=0.1s
I think those results speak to an important point about how to play CM.
Increases in DPS through increased APoC comes from additional casts of explosive blast, I assume you've modeled this to use explosive blast optimally when enough arcane power is avaliable so as not to interrupt spam of the proc generator. It is important to note however that optimal DPS is being obtained when the user is careful not to delay their next energy twister or meteor by activating explosive blast with insufficient excess arcane power.
With respect to your spreadsheet, I agree that we're basically calculating a soft cap beyond which APoC no longer provides a substantial performance increase, rather than a minimum to run the build.
Some clarifications though. Morphos was running, according to his thread 45% critical hit chance, 1.64 attacks per second, 27 APoC, using Star Pact and Prism. If we assume this gear is the minimum level of arcane power generation to make the build work, we can calculate the percentage of full APoC coverage he is getting to determine what out targets should be when reversing the process to produce required APoC values.
***NOTE: Doing the math on what I observed from the video I am led to believe that I have something wrong in my modeling of Star Pact, or the patch notes coefficient of 0.05 is off base and the actual value is higher than that.***
If we calculate assuming prism is up we find his coverage from APoC amounts to only about a little over 10% of the cost of Star Pact, which leads me to a different conclusion:
His Star Pact is not paying for itself, but his explosive blast is almost paying for itself, and actually pays for meteor casts beyond a certain level of APoC. Under a Prism his explosive blast only costs 13 arcane power, and it is generating about 4 AP per cast from APoC giving it a very low adjusted Arcane Power cost of 9.
Watching a 30 second sample of the keywarden fight (1:10 to 1:40 on the video) where only a single target is present, I count 24 meteors landing, and 13 casts of explosive blast, and prism is up for 27 of the 30 seconds (90%). His total arcane power expenditure during this time is 841 Arcane Power in 30 seconds or 28 arcane power per second. Thats a deficit of 18 arcane power per second from base regeneration. He generates 39 chain reaction blasts, each one has a 45% chance to crit, resulting in an average of 17.55 critical hits, each of which refunds 3 (2.97) arcane power, for a total of between 51 and 54 arcane power. His 24 meteors generate 24 landings and 72 DoT ticks, for a total of 96 chances to crit. On average 43 or 44 of these will crit, and that will generate 1.35 (so 1 or 2 depending on how APoC works with rounding and partial AP numbers) arcane power per crit, or anywhere from 43 to 88 arcane power.
Assuming this rounds very generously, the 122 arcane power APoC covers translates to only 4.067 arcane power per second of the 18 arcane power per second deficit, and with the reported old coefficient of 0.1 it would only be covering 8. This works out to 8 or 9 meteor casts that he should not have been able to pay for, or 50% more meteor casts, or almost five times the AP return from APoC than i would have expected based on the numbers we've been told.
Either I've made a serious error in my math, or the star pact coefficient of 0.05 is bogus, or the mechanics of star pact work substantially different from my understanding of them.
Edit:
While I don't want to base any conclusions on the above since it doesn't line up with observation yet, the calculations above show the net effect of 27 APoC on arcane power deficit to be reducing it by 4.067 (122 arcane power over 30 seconds), running Storm Armor (Power of the Storm), would have saved 3.7 arcane power per second (111 arcane power over 30 seconds), and a single -5 meteor cost item would have saved 4 arcane power per second (120 arcane power over 30 seconds) producing the same effect as three full sources of APoC.
Thank you for the information. From the results it seems like mistral breeze is out performing wicked wind in most cases. Unless im not understanding the chart correctly. It also seems that ap on crit is a bigger boost to overall performance damage wise and shouldnt be underestimated
Thank you for the information. From the results it seems like mistral breeze is out performing wicked wind in most cases. Unless im not understanding the chart correctly. It also seems that ap on crit is a bigger boost to overall performance damage wise and shouldnt be underestimated
It should be noted that these sims aren't taking into account damage that can be gained from Storm Armor (Shocking Aspect). Assuming it can be calculated as simply 35% weapon damage * PPS * CritChance / AttackSpeed, it would adjust the relative cases in favor of wicked wind, which generates more PPS than Mistral Breeze. Not having matlab to look at the code I'm not sure how the implementation of shocking aspect would look like overall.
Also note that there is an assumption in this simulation that you get 50% uptime on mistral breeze casts, 75% uptime on wicked wind casts, and 90% uptime on meteor casts. I don't personally feel that wicked wind would be any less accurate than meteor, in meaningful cases, and I highly doubt anyone is routinely getting 3 full seconds of exposure on average from moving energy twisters, so I suspect that assumption is weighting the math towards mistral breeze and inflating its numbers.
Thanks a lot Loroese! That is now beginning to make sense, could you possibly just extend that table to 60% CC (which is rather obtainable) and to say 3.8ish APS (which is obtainable once again with say chantodo's set), surely i know it is going to cost a bomb to get to those levels, but just as a matter of curiosity as to whether it is at all possible to get to only 10 APoC pulling the weight of the build
Rollback Post to RevisionRollBack
"Only two things are infinite, the universe and human stupidity, and I'm not sure about the former."
That's some interesting analysis, but I think the video you were looking at was made before the meteor coefficient nerf, when it was still 0.25. Now, it's 0.1, the 0.05 is only for meteor shower, I think it's called. Also, FN returns some AP, though it only ends up being something like 1.275 AP per cast on average. I think the majority of the deficit comes from just plain old white attacks used while spamming meteor when out of AP.
The video I was refering to is this one, after which he says he's only using 17 APoC. In the first 30s of the vid, before he gets knocked back, he casts 9 CR, 7 FN, and 16 meteors.
Total AP cost is 13*9+16*28=565.
AP regen is about 300 and starts wso he gains 265 AP through crits.
CR: 9casts*3tics per cast*1/9 (coef)*17 AP per crit*0.45 crit = 22.9 AP gain
Meteor: 16casts*4 tics (incl blast)*0.1 (coef)*17 AP per crit*0.45 crit = 49.0 AP gain
FN: 7*1/6*17 = 19.8 AP gain
22.9+49+19.8 = 91.7 AP gain total, so we're still 173.3 AP short.
In 30s he has roughly 45 total attack opportunities, 23 of which are used on FN and meteor. If he lands all of the remaining as white attacks, that's about 22 white attacks. If we assume coefficient of white attacks is 1 (which I really need to test), then we get 22*0.45*17 = 168.3 AP gained, which is right around the defecit. He didn't actually land all the attacks, but he also started with some AP so the two factors mostly cancel out and we have all AP accounted for.
Regarding my model, I assume that FN is always cast when available, otherwise Star Pact (or whatever main AP spender is being modelled) is used when AP is available, and CR is used after if there is enough AP. I do not account for trying to save AP for star pact by not casting CR since CR does the most damage per AP spent and the relative cost of CR is quite small. If FN isn't cast and there isn't sufficient AP for SP, then I assume a white attack is landed, though I should probably add a factor for white attack uptime, since not every white attack will land.
The reason my WW and meteor coefficients are not the same is because the meteor tics are done after 3s while the WW is up for twice as long, which gives the mob twice the time to move out of range of it. I do agree that the actual weights used might not be the best possible ones but I felt they were reasonable for the analysis so far.
Regarding Shocking Aspect, I haven't looked at the skill so I don't know anything about how it functions. If I knew how it worked, I could try including it in my model.
@ Vibe1320
I wouldn't use the results to say absolutely how one spec performs versus another, but rather compare how each spec does vs APoC, so just compare the WW changes, or just the MB changes as you increase AP and use whichever between WW and MB you like or feel performs best. As was said in the post above, the uptime coefficients are just very rough approximations.
Personally I like WW because I just don't like the randomness of the tornados. Actually I prefer meteor because of the higher potential damage to annoying ranged mobs, but I only have an attack speed of 1.5 since I've mostly just gone Crit and Crit damage for Archon. That said, I haven't actually tried the build since the start of 1.0.4 because I never actually liked the spec. I'm only now doing this analysis because it seems like the most viable method for wizards to try and tackle the Uber bosses on MP10, which seems the best way to farm ring mats, if you can kill them before they enrage.
The m file can be found here. Current version is 1.01. Any changes since v1.00 will be included at the end of the mfile and included at the end of this post.
Please post any ideas or suggestions you have, or any errors you see.
Features
Uses 1.0.5 spell coefficients and crowd control dimishing returns to simulate combat in D3 against 1 or more mobs for a duration that is easily adjusted. Default is usually 60 seconds.
Approximates mob movement by using an uptime coefficient that checks for each damage tic against the given probability. I've set the defaults to 100%, 90%, and 50% for WW, Meteor (both runes), and Mistral Breeze respectively.
Attempts to simulate button mashing using a Reaction Time (default is 0.01s) that assumes you press whatever button you want to at that time. It only allows spells that use a global cooldown to be used at specific intervals related to your attack speed, namely Frost Nove and the AP spenders. DS and Chain reaction can be used at any time since they do not trigger a global cooldown (i.e., a weapon strike, for those not used to the WoW lingo).
Allows for lots of variety in input ranging from basic gear stats like Crit%, AP on Crit, attack speed, max AP, various runes and passives like DS - prism and evocation, and going all the way to mob attack speed and approximate spell uptimes.
The simulator uses a Monte Carlo based approach to simulate random events over the duration specified, then repeates several times (default is 1000 simulations per skill).
The final results that are tracked are the Effective DPS multiplier of your skills and gear, % uptime of DS, the % of time the mobs are kept frozen, and the average LoH retuned per second. The Effective DPS multiplier is simply the number you multiply your char sheet dps by to get your effective dps. If you have 100k char sheet dps and your multiplier is 3.5, you would actually do around 350k dps with that skill set up and those base stats. The average LoH returned per second is basically what it sounds like. Whatever your total LoH is, you multiply it by that number to get your approximate life gain per second. For example, if the number if 3 and you have 1000 LoH you gain about 3000 life per second on average.
The simulator also outputs the % damage done by each skill for each build.
Assumptions
Mob swing timer = 1s and swing timer is reset when mob is frozen. These parameters are easily edited though.
White attacks have a coefficient of 1 and White Attacks are always performed if FN is on CD and there is insufficient AP for the AP spender. I have confirmed that the LoH coefficient for white attacks is 1.
Conduit uses spell coefficients so on a successful hit you always gain back AP = APoC*Spell Coefficient.
Evocation, when combined with CD reduction rune on Frost Nova, reduces the cooldown to 9*(1-0.15) = 7.65s.
Spell Priority is TW(if down) > DS >= FN > AP spender >= CR. CR is used on CD whenever there is enough AP, but only after spending AP on the big AP cost. Since it does not trigger a global CD or swing timer, it effectively can be used at the same time as DS and AP spender/FN. AP spender is refered to as the Prim spell in the code, for primary.
CR rune for Explosive Blast is used. Options for other runes should be added soon.
Planned Additions
I want to add functionality for additional EB runes.
Code Updates
v1.01 Changed FNdf to FNbc since I had incorrectly put deep freeze as the rune that adds 15% damage while frozen, but it is actually bone chill
v1.02 Added WWcmcoef and WWapoccoef since the CM procs and APoC returns are higher than otherwise expected. Default values are 2 for each based on best guess from limited testing.
Crusader DPS and EHP Spreadsheet, meant for Crusaders
My Wizard
Base Stats:
APoC=19;
CC=0.505;
MaxAP=114; (94 with prismatic or pinpoint barrier)
APregen=10;
DSprism=0;
DSds=0;
Evocation=1;
CB=1;
Conduit=0;
SimDuration=VARIED;
Mobs=1;
AS=VARIED;
MobAttackRate=1;
DSabsorbs=1;
MobAttackDelay=MobAttackRate;
Simulations=300;
WWuptime=1;
ReactionTime=0.01;
Case 1: AS = 2.6904
Case 2: AS = 2.28
Case 3: AS = 1.824
The dps columns are as follows:
DPS Mult = the number you multiply your char sheet dps by to get your effective dps
Rel DPS = relative fractional dps gain from the skill sets, compared to the all defensive build
Simulated dps = DPS multiplier obtained from the simulator. This should compare directly to the DPS Mult value
Rel Simulated DPS = relative fractional dps gain from the skill sets estimated by the simulator
As you can see the simulator has good agreement with the in game testing. It does appear to overestimate the shocking aspect contribution, which can be fixed by tweaking the empirical formula used. The Storm Armor results seem to not agree as well for Case 2 and 3 as Case 1, but I think that was mostly just RNG since the Storm Armor strikes can be pretty sporadic.
For reference, the empirical formula used is
WWsacoef = - 0.8737 * AS + 4.8035
where WWsacoef is the adjusted number of SA procs from WW per CM proc. In other words, if a tic from WW would proc CM, it also procs SA (WWascoef) times. This is only used for WW. CR, DS, and any other skills are all treated the same as CM at a 1:1 ratio.
EDIT: Here are some simulations I ran based on my current tank stats (49% crit, 20 APoC, 2.5308 Attacks per second, with ReactionTime=0.1, SimulationDuration=60, Simulations=100). I used the full SNS build with Deep Freeze rune on FN and no Evocation, assuming I'd use something different but not completely sure what. What I did was change the CC, APoC, and AS by increments I call delta. I only changed one parameter at a time and kept the other two at the value listed above. Delta = 2% CC, 1 APoC, or 0.1 Attacks per second.
You can clearly see the WW breakpoints in the LoH gain per second and the FN uptime. The Effective DPS Multiplier also shows the WW breakpoints in the form of decreasing dps after the breakpoint, then a jump increase at the next breakpoint, etc. This is because increasing your attack speed while between breakpoints will increase your char sheet dps but will not lead to an increase in build dps, therefore the char sheet dps increase is misleading between breakpoints. I would guess that your effective dps would remain the same.
It's also interesting that effective dps actually seems to decrease as you raise Attack Speed past the 2.5 breakpoint, which is unexpected. As you increase Crit, it continues to increase, but APoC only has a very weak impact after the 2.5 APS breakpoint, likely because at that point you gain AP about as fast as you can spend it.
Crusader DPS and EHP Spreadsheet, meant for Crusaders
My Wizard
First, note that the use of Conduit does not apply the 10% bonus damage to the results since that is applied to your char sheet dps.
Raising crit% by 5% for Case 1 will raise your in game dps number but that is not accounted for in the simulator because the simulator is meant to be independent of the in game displayed dps number.
Conclusions
1) Conduit appears worthwhile for MB (5.4% dps increase) for any spec with Magic Weapon. For a WW build, it is about a break even dps alternative to Force Weapon if you already use Glass Cannon. If you don't use glass cannon it is a dps loss (4% gain vs 5% from FW). For meteor builds the gain is only about 3% dps so FW is better. Conduit also slightly improves DS uptime for all builds while also increasing frozen time for the twister specs.
2) Prism > Diamond Shards. DPS is higher, DS uptime is higher, and mobs are frozen more with Prism.
3) 5% crit > 7% IAS. Roughly speaking, if you have a 1.4 APS weapon wiht 7% IAS, your attack speed is around 1.5. To raise your attack speed to 1.6 you would need another 7% IAS. At 40% crit, you're better off gaining 5% crit than raising your attack speed to 1.6 APS. The dps is higher and the DS uptime is higher while the frozen % is roughly the same, with a slight edge to the +CC case. In short, both are good options to raise dps but CC also raises survivability. Also, CC raises your char sheet dps as well as the effective weapon dps while IAS will not raise your effective DPS beyond that shown in the table above.
4) AP on Crit is important. Just raising the amount from 16 to 17 gave a noticable increase in dps (1-2%) and a small increase to DS uptime.
5) Evocation is a promising passive both for dps and survivability since it increases DS uptime considerably. It provides a bit bigger of a dps boost for twister specs while the meteor specs get more survivability, but all specs show an increase in all parameters when Evocation was active.
6) Astral Presence is a large dps increase (5-7%), though smaller survivability increase than Evocation.
7) Calamity seems pretty bad for the build, which makes sense because it does less dmg than CR, though it is a bit cheaper, but also has about half the effective procs as CR and uses a global cooldown whereas a white attack would do less damage but refunds much more AP on a crit.
8) I took a quick look at Slow Time uptime with the new 15s CD in 1.0.5 and so far it seems like practically 100% uptime for all specs using the base stats above. This makes it higher dps gain than Magic Weapon when using the 20% damage rune, but also the option of raising your attack speed for those using twister spec. To approximate the affects of Slow Time, just add 10% of your base weapon speed to AS, so if you have a 1.4 speed weapon, add 0.14 to whatever your char sheet attack speed is, or just activate the ability and note the AS I suppose. For the damage rune, just multiply the end damage by 1.2.
Right now, my expected CM spec would be http://us.battle.net...YXOg!YWg!ZcZZYZ, assuming survivability isn't an issue. If it is, I'd likely replace Glass Cannon with Evocation and Time Stretch with either Conduit or Blood Magic, though I might end up replacing it with Teleport since I really liked using teleport as a CM wiz. Slow Time is used because the uptime is expected to be right around 100% even for low stats like those in the base case.
My reason for picking meteor over a twister build is that my current stats are pretty similar to those used in the base case above. However, it seems that the twister specs are scaling a lot better with most of the stats so with better gear I think a WW or MB spec would be more ideal. About the only appealing attributes of meteor spec is it seems like it would be slightly more effective against ranged mobs and does more dps per cast than WW or MB (though less potential DPS per AP spent than MB) and seems more suited to low attack speed gear sets.
Crusader DPS and EHP Spreadsheet, meant for Crusaders
My Wizard
Could you please do a favor to the all of us and create a table for WW based CM wizards with:
- attack speed along one axis
- CC along another axis
- each cell = minimum required APoC to sustain the build
This is a question i know i want an answer to, i'm pretty sure a lot of other wizards would like to know this as well - after all gearing up with all 3 sources of APoC is very limiting to the DPS and very draining financially, it would be far better to have an accurate goal to aim for in terms of APoC on gear. For example, if we are reaching 3.0 APS, would a single source of APoC suffice? Or do we still have to have 3?
Right now it seems the meteor build is viable with just 17 APoC and 45% crit and 1.5 attacks per second, even with low EHP and Force Armor.
Crusader DPS and EHP Spreadsheet, meant for Crusaders
My Wizard
I don't think it would be applicable to the meteor build as the APS plays no role in that, which would make WW build more viable at higher gear levels. Quite honestly, other than a slightly different playstyle of the meteor build - i don't really see any other gearing up synergies in there: APoC is only available on 3 sources:
helm - tal's helm or mempo are going to beat any APoC helm flat out in terms of DPS any day
weapon - chantodo's have APoC but they cost 600+ mil for 1.2k DPS, while echoing fury with the same socket in 1.2k range will cost about 100m tops, rare wands with APoC cost a bomb as well and suck in DPS values as most of them lack INT, those that do have stupid costs again
offhand - is the more reliable APoC source, any OH can roll APoC
Hence the entire gearing question arising from this - if we go for a high APS and high CC would a single APoC source be sufficient to run the WW based CM build? Or do we have to inevitably invest into at least 2 sources? Like for example i'm sitting on over 300m gold right now, and i would like to have the ability to run not only archon, but also CM, especially for the ubers, so do i have a chance to run CM with just OH APoC or do i invest into MH with APoC or do i drop a chunk of DPS and get an APoC helm?
According to your research - with, say, 3 APS WW is going to have 2*6*3=36 ticks wouldn't that be fine to run with just 10 APoC or even less? What about 2.5 APS?
If I had to pick a requirement I would say something like 55%+ FN uptime, but if you have another threshold or another parameter in mind, I can work around just about anything.
Crusader DPS and EHP Spreadsheet, meant for Crusaders
My Wizard
Do i run out of AP if
CC = 40% APS = 2.0 APoC = 15 ?
CC = 45% APS = 2.0 APoC = 10 ?
CC = 50% APS = 3.0 APoC = 8 ?
etc
Is there a way to run CM build using WW and EB as AP spenders while having only 1 APoC source? What is the minimum required CC and APS for that?
Crusader DPS and EHP Spreadsheet, meant for Crusaders
My Wizard
I don't know am i not explaining the concept clearly? Let me rephrase it this way:
1) Would YOU run CM build with WW with 10 APoC? Yes or no.
2) What is the correlation between the APS and CC versus minimum APoC required on gear to make the build sustainable in a sense of not running out of AP in 10 seconds?
3) Based on your conclusions the WW testing thread - at APS 3.0 WW is going to tick 32 times over 6 seconds, i do not see the point of having 30 APoC in that setup because that would be very very very extremely heavy extra.
As long as the dps is fairly high, I would assume the build is sustainable because you're ragaining enough AP to frequently cast spells. From there it's just a matter of staying alive to continue to damage the mob, which is where FNuptime comes in, and the higher uptime the easier to survive.
Regarding your questions:
1) Based on a quick simulation, I still see an uptime of 56% for FN with 1.5 APS and 10 APoC and 40% crit. However, over half of the attacks done over a 60s fight are white attacks, and I probably overestimate the effectiveness of white attacks. I really should check to see what the coefficient of a white attack is. If my assumptions for white attacks is correct, the build looks sustainable for this situation.
2) As I said above, I don't think "not running out of AP in first 10s" is a valid criteria because virtually everyone will run out of AP at some point in the first 10s when facing just one mob. If you have 100 AP, you start with FN+CR then spam WW. At 1.5 aps you regen 6.67 AP between casts so WW effectively costs 28.33 AP, meaning you're out of AP after 3 casts+CR. You might have regained a little AP by the time you get your 3rd WW cast, so maybe 4-5 casts, but either way you'll run out of AP while waiting for the tics to rack up and the CR blasts to finish.
3) With 1.5 APS and 45% CC it seems 17 APoC should be enough to sustain the build to an acceptable degree, so I don't see the need for 30 APoC ever. Having more APoC is useful because it increases your effective DPS and the uptime of FN and DS because you can cast more spells
Since the video I used shows sustainability against a single target I'll just use those parameters as a measure, namely 57% FN uptime. To reduce computation time, I'll probably just use 500 simulations and 0.1s reaction time, along with 2-3% intervals for CC and 0.1 APS intervals.
Crusader DPS and EHP Spreadsheet, meant for Crusaders
My Wizard
The calculation is fairly simple, the expected APoC return of the spell has to be greater than its cost, so:
Cost = Crit * Ticks * coefficient * ApoC
The real question is, how many targets. None of the meteor runes at real gear levels (i.e. less than BiS values) will break even on Arcane Power through APoC or enhanced regeneration on a single target, and it takes close to BiS to do so with two targets, but virtually all of them will generate surplus arcane power on 3 targets with moderate gear levels.
Yeah, that sounds like an easy enough calculation that doesn't really require simulating if you just use averages. With ticks = 2*APS*6 as an approximation and cost = 35 for WW, I'd also add an uptime coefficient of say 0.75 to account for the mob moving out of the WWs so we have
35 = CC * 12 * APS * 0.125 * APoC * 0.75 or 31.11 / APS / CC = APoC
It's pretty easy to make a table with APS on one column and CC as the row entry and solving for APoC for each combination. The problem is with 3 APS and 50% CC you still need 21 APoC. If you assume Prism is active it drops it to 17. This doesn't account for static AP regen, so you can change the cost from 35 to (35-7)-10/APS since you gain 10 AP per sec but you attack once every 1/APS seconds. With that equation you'd still need like 46% attack speed, 1.8 APS, and 24 APoC. That also doesn't account for extra AP to spend on EB, though when you case FN you save the AP from a WW cast that can go towards the EB.
Here's the table I threw together.
If you double the mobs, you halve the APoC needed, roughtly speaking, so it sounds a lot more reasonable with 2 mobs, i.e., Uber boss pairs.
Also, while this calculates the APoC desired to break even with WW, that doesn't mean it's the min APoC needed to sustain the build. I'd actually say it's more like the max APoC needed. As I said before, the vid clearly shows sustainability vs 1 mob with only 45% crit, 1.5 APS, and 17 APoC, whereas the table says you'd need 28 APoC, and he was using a meteor build, rather than a WW build.
If I find the time, I'll run some simulations to come up with a similar (though smaller) table based on a FN uptime of around 57%.
Crusader DPS and EHP Spreadsheet, meant for Crusaders
My Wizard
PPS = Average full LoH procs per second, so if your PPS = 2, you gain 2x your total LoH each second.
Base case, 2.5 APS, 45% CC, 100 MaxAP, SimDuration = 30, 10 APoC, ReactionTime=0.01s
20 APoC
30 APoC
I only ran 1000 simulations, so there's going to be some varience to the results, but hopefully they give you some good information. Also, sorry for being a bit lazy and not rounding the results like I did above.
Base case, 2.5 APS, 45% CC, 100 MaxAP, SimDuration = 60, 10 APoC, ReactionTime=0.01s
20 APoC
30 APoC
So you gain quite a bit of dps but you shouldn't really have any problems with FN locking mobs even at 10 APoC. Same thing with DS uptime, but that all depends on your reaction time or how fast you can mash buttons.
Just for completion, and because the run times are only like 20s, here's the results with ReactionTime=0.1s
Base case, 2.5 APS, 45% CC, 100 MaxAP, SimDuration = 60, 10 APoC, ReactionTime=0.1s
20 APoC
30 APoC
Crusader DPS and EHP Spreadsheet, meant for Crusaders
My Wizard
Increases in DPS through increased APoC comes from additional casts of explosive blast, I assume you've modeled this to use explosive blast optimally when enough arcane power is avaliable so as not to interrupt spam of the proc generator. It is important to note however that optimal DPS is being obtained when the user is careful not to delay their next energy twister or meteor by activating explosive blast with insufficient excess arcane power.
With respect to your spreadsheet, I agree that we're basically calculating a soft cap beyond which APoC no longer provides a substantial performance increase, rather than a minimum to run the build.
Some clarifications though. Morphos was running, according to his thread 45% critical hit chance, 1.64 attacks per second, 27 APoC, using Star Pact and Prism. If we assume this gear is the minimum level of arcane power generation to make the build work, we can calculate the percentage of full APoC coverage he is getting to determine what out targets should be when reversing the process to produce required APoC values.
***NOTE: Doing the math on what I observed from the video I am led to believe that I have something wrong in my modeling of Star Pact, or the patch notes coefficient of 0.05 is off base and the actual value is higher than that.***
If we calculate assuming prism is up we find his coverage from APoC amounts to only about a little over 10% of the cost of Star Pact, which leads me to a different conclusion:
His Star Pact is not paying for itself, but his explosive blast is almost paying for itself, and actually pays for meteor casts beyond a certain level of APoC. Under a Prism his explosive blast only costs 13 arcane power, and it is generating about 4 AP per cast from APoC giving it a very low adjusted Arcane Power cost of 9.
Watching a 30 second sample of the keywarden fight (1:10 to 1:40 on the video) where only a single target is present, I count 24 meteors landing, and 13 casts of explosive blast, and prism is up for 27 of the 30 seconds (90%). His total arcane power expenditure during this time is 841 Arcane Power in 30 seconds or 28 arcane power per second. Thats a deficit of 18 arcane power per second from base regeneration. He generates 39 chain reaction blasts, each one has a 45% chance to crit, resulting in an average of 17.55 critical hits, each of which refunds 3 (2.97) arcane power, for a total of between 51 and 54 arcane power. His 24 meteors generate 24 landings and 72 DoT ticks, for a total of 96 chances to crit. On average 43 or 44 of these will crit, and that will generate 1.35 (so 1 or 2 depending on how APoC works with rounding and partial AP numbers) arcane power per crit, or anywhere from 43 to 88 arcane power.
Assuming this rounds very generously, the 122 arcane power APoC covers translates to only 4.067 arcane power per second of the 18 arcane power per second deficit, and with the reported old coefficient of 0.1 it would only be covering 8. This works out to 8 or 9 meteor casts that he should not have been able to pay for, or 50% more meteor casts, or almost five times the AP return from APoC than i would have expected based on the numbers we've been told.
Either I've made a serious error in my math, or the star pact coefficient of 0.05 is bogus, or the mechanics of star pact work substantially different from my understanding of them.
Edit:
While I don't want to base any conclusions on the above since it doesn't line up with observation yet, the calculations above show the net effect of 27 APoC on arcane power deficit to be reducing it by 4.067 (122 arcane power over 30 seconds), running Storm Armor (Power of the Storm), would have saved 3.7 arcane power per second (111 arcane power over 30 seconds), and a single -5 meteor cost item would have saved 4 arcane power per second (120 arcane power over 30 seconds) producing the same effect as three full sources of APoC.
Also note that there is an assumption in this simulation that you get 50% uptime on mistral breeze casts, 75% uptime on wicked wind casts, and 90% uptime on meteor casts. I don't personally feel that wicked wind would be any less accurate than meteor, in meaningful cases, and I highly doubt anyone is routinely getting 3 full seconds of exposure on average from moving energy twisters, so I suspect that assumption is weighting the math towards mistral breeze and inflating its numbers.
The video I was refering to is this one, after which he says he's only using 17 APoC. In the first 30s of the vid, before he gets knocked back, he casts 9 CR, 7 FN, and 16 meteors.
Total AP cost is 13*9+16*28=565.
AP regen is about 300 and starts wso he gains 265 AP through crits.
CR: 9casts*3tics per cast*1/9 (coef)*17 AP per crit*0.45 crit = 22.9 AP gain
Meteor: 16casts*4 tics (incl blast)*0.1 (coef)*17 AP per crit*0.45 crit = 49.0 AP gain
FN: 7*1/6*17 = 19.8 AP gain
22.9+49+19.8 = 91.7 AP gain total, so we're still 173.3 AP short.
In 30s he has roughly 45 total attack opportunities, 23 of which are used on FN and meteor. If he lands all of the remaining as white attacks, that's about 22 white attacks. If we assume coefficient of white attacks is 1 (which I really need to test), then we get 22*0.45*17 = 168.3 AP gained, which is right around the defecit. He didn't actually land all the attacks, but he also started with some AP so the two factors mostly cancel out and we have all AP accounted for.
Regarding my model, I assume that FN is always cast when available, otherwise Star Pact (or whatever main AP spender is being modelled) is used when AP is available, and CR is used after if there is enough AP. I do not account for trying to save AP for star pact by not casting CR since CR does the most damage per AP spent and the relative cost of CR is quite small. If FN isn't cast and there isn't sufficient AP for SP, then I assume a white attack is landed, though I should probably add a factor for white attack uptime, since not every white attack will land.
The reason my WW and meteor coefficients are not the same is because the meteor tics are done after 3s while the WW is up for twice as long, which gives the mob twice the time to move out of range of it. I do agree that the actual weights used might not be the best possible ones but I felt they were reasonable for the analysis so far.
Regarding Shocking Aspect, I haven't looked at the skill so I don't know anything about how it functions. If I knew how it worked, I could try including it in my model.
@ Vibe1320
I wouldn't use the results to say absolutely how one spec performs versus another, but rather compare how each spec does vs APoC, so just compare the WW changes, or just the MB changes as you increase AP and use whichever between WW and MB you like or feel performs best. As was said in the post above, the uptime coefficients are just very rough approximations.
Personally I like WW because I just don't like the randomness of the tornados. Actually I prefer meteor because of the higher potential damage to annoying ranged mobs, but I only have an attack speed of 1.5 since I've mostly just gone Crit and Crit damage for Archon. That said, I haven't actually tried the build since the start of 1.0.4 because I never actually liked the spec. I'm only now doing this analysis because it seems like the most viable method for wizards to try and tackle the Uber bosses on MP10, which seems the best way to farm ring mats, if you can kill them before they enrage.
Crusader DPS and EHP Spreadsheet, meant for Crusaders
My Wizard