Magicka DPS Set Bonus Comparison - MATH (pre-HotR)

dpencil1

@geonsocal
Haha. I have the dpencil handle also. For some reason I had to make a new forum account to access the Morrowind Closed Beta forum, so I just switched to using this one.

As far as Torgue's goes? That dips into parameters outside the scope of these calculations, at least with anything other than a weapon damage glyph. I could maybe try to calculate that part at least.

reesenorman

This is ***

Dymence

reesenorman wrote: »

This is ***

Insightful

Skagsmasha

@dpencil1 ... U da man!!

This would have to be the most constructive posts I have seen in a long time with regards to comparison.
Kudos to everyone for keeping it on the level as well - even though this is all pre-HotR.

Morgul667

It is a great post indeed

dpencil1

@catch22atplay
The question at hand seems to be: "How much does magicka actually get boosted in a realistic build, and how does that then skew our results?" To answer this, I will do an analysis on the UESP build editor with Live values first and then modify for HotR.

Totally naked lvl 50 with 64 attribute points in magicka, 0 CP, 0 skills, 0 passives: 15062
Naked with 300 CP: 18074 (20% increase)
Naked with Bound Aegis: 16266 (8% increase)
Naked with 300 CP and Bound Aegis: 19520 (29.6% increase...21.6% + 8%)

Naked with Bound Aegis and Inner Light: 17020 (13% increase...8% + 5%)
Naked with 300 CP and Bound Aegis and Inner Light: 20424 (35.6% increase...22.6% + 8% + 5%)

Naked with Bound Aegis and Inner Light and High Elf Passive: 18526 (23% increase...8% + 5% + 10%)
Naked with 300 CP and Bound Aegis and Inner Light and High Elf Passive: 22231 (47.6% increase...24.6% + 8% + 5% + 10%)

So to use the formula you got from Asayre: ((Base+AP+Gear)*%CP)+Food+Mundus(Divines))*Skills
(15062 x 1.2) x 1.08 = 19520
(15062 x 1.2) x 1.13 = 20424
(15062 x 1.2) x 1.23 = 22231
This confirms that the formula is correct (no surprise there, it is from Asayre after all)!

We also notice that every time we increase the % bonus from skills by 5% we add another 1% to the CP contribution, which makes sense since we are multiplying against 20%. Every 1% from skill percent amps adds 0.2% to the CP contribution.

So lets take a truly realistic comparison of Julianos, Spinners, and Necro with Live and then HotR values.

For this, I have set up a High Elf (10%) Mag Sorc with 2 Ilambris, 5 _____, 4 Moondancer, all divines with magicka enchants, Thief mundus, Blue food, Sharp staves. 10% Ele Expert, 15% Elfborn, 2766 Erosion. I will include Inner Light with the Mage's Guild passive (7%), 5-1-1 Undaunted Mettle (6%), CP 300 (24.6%), Major Sorcery (20% SD), and 3 Sorc skills (6% SD)
That's 47.6% boosted Magicka and 26% boosted SD.

Necro LIVE: 47682 Magicka, 2502 SD, 58.8 Crit, 65.1 Crit Dmg, 10.2% CP Magic Done, 89.4% Pen, 5% Damage Done
10071

Prepping HotR calculation:
5 Magicka bonuses increased from 967 to 1096 (3 from Necro, 1 from Ilambris, 1 from Moondancer)
967 x 5 = 4835
1096 x 5 = 5480
5480 - 4835 = 645 more magicka in HotR from 2-4 piece bonuses
4000 - 3150 = 850 less magicka from Necro 5-piece
645 - 850 = -205

47682 ÷ 1.476 = 32304
32304 - 205 = 32099
32099 x 1.476 = 47378

47682 (Live) - 47378 (HotR) = 304 less Magicka in HotR than Live

Necro HotR: 47378 Magicka, 2502 SD, 55.7 Crit, 65.1 Crit Dmg, 10.2% CP Magic Done, 84.6% Pen, 5% Damage Done
9351

Prepping HotR Necro (no nerf) calculation:
5 magicka bonuses = 645 more magicka
32304 + 645 = 32949
32949 x 1.476 = 48632

48632 - 47682 = 950 more magicka for HotR Necro (no nerf) than Live Necro

48632 - 47378 = 1254 less magicka for HotR Necro than HotR Necro (no nerf)

Necro (no nerf) HotR: 48632 Magicka, 2502 SD, 55.7 Crit, 65.1 Crit Dmg, 10.2% CP Magic Done, 84.6% Pen, 5% DamageDone
9512

Julianos LIVE: 40286 Magicka, 2879 SD, 65.1% Crit, 65.1% Crit Dmg, 10.2% CP Damage Done, 89.4% Pen, 5% Damage Done
9887

Prepping HotR Julianos calculation:
3 Magicka bonuses increased (1 from Julianos, 1 from Ilambris, 1 from Moondancer)
967 x 3 = 2901
1096 x 3 = 3288
3288 - 2901 = 387
40286 ÷ 1.476 = 27294
27294 + 387 = 27681
27681 x 1.476 = 40857
40857 - 40286 = 571 more magicka in HotR

2 Crit bonuses increased
688 x 2 = 1376
1066 x 2 = 2132
2132 - 1376 = 756
756 / 219 = 3.45% more crit in HotR

Juliano HotR: 40857 Magicka, 2879 SD, 65.5% Crit, 65.1% Crit Dmg, 10.2% CP Damage Done, 84.6% Pen, 5% Damage Done
9449

40857 ÷ 1.476 = 27680
27680 x 1.498 = 41464

Julianos HotR (+2%): 41464 Magicka, 2879 SD, 65.5% Crit, 65.1% Crit Dmg, 10.2% CP Damage Done, 84.6% Pen, 5% Damage Done
9530

Spinner LIVE: 41643 Magicka, 2664 SD, 58.8% Crit, 65.1% Crit Dmg, 10.2% CP Damage Done, 97.4% Pen, 5% Damage Done
10328

Prepping HotR Spinner calculation:
4 Magicka bonuses increased (2 from Spinner, 1 from Ilambris, 1 from Moondancer)
967 x 4 = 3868
1096 x 4 = 4384
4384 - 3868 = 516
41643 ÷ 1.476 = 28213
28213 + 516 = 28729
28729 x 1.476 = 42404
42404 - 41643 = 761 more magicka in HotR

Spinner HotR: 42404 Magicka, 2664 SD, 55.7% Crit, 65.1% Crit Dmg, 10.2% CP Damage Done, 91.5% Pen, 5% Damage Done
9664

Note that while the LIVE versions show higher ESP, that's mainly because of the nerfs to the Sharpened trait and the Thief stone, which we wanted to leave alone so we could really look at the impact of Max Magicka to the results, factoring in all the additional boosts from CP and skills.

Here is the result:

I have subtracted 9000 from the totals of each value to better highlight their differences.


Now lets compare these calculations with my Comparison G calculations. I have removed 9000 from the Realistic Build numbers and 9100 from the Comparison G numbers so we can see their interrelationship easier:


The key differences seem to be that Necro (no nerf) fairs a little better in the Realistic Build with the additional bonuses to magicka, while Julianos fairs a little worse, and Spinners remains basically the same. However, the distance from Necro to Spinners in the Realistic Build is 313, while Comparison G had them at 322. So the gap shrinks by only 9 points by adding all those extra calculations. And remember the true values of these numbers are 9351 to 9664 and 9412 to 9734, which makes the overall difference between these comparisons only 0.001%!

To my mind, the assumption that CP and magicka boosting skills drive Necropotence into a much stronger place than when considered in isolation seems to be proven false here. It is true that we notice a very slight increase in the Necro (no nerf) version, and a slight decrease with Julianos, but the overall distance between Necro and Spinner remains unchanged. If this issue were as important as catch22 claims, we would have expected to see Necro close the gap much more significantly than 0.001%.

dpencil1

....and one more graph for good measure, still drawing from the calculations in the previous post. This is simply comparing Necro, Julianos, and Spinners Effective Spell Power using the realistic build I set up on LIVE to show how they currently relate to one another in terms of ESP:

Texecutioner187

Is all of this data is presumption of all Divine traits on gear?

dpencil1

My response to catch22 (3 posts above this) has been updated. It does seem to confirm that my initial calculations are not far off realistic build conditions, even without taking into account CP and % boosting skills. The benefit that sets get from these things is universal and the difference created between builds with high magicka vs builds with lower magicka but higher SD or Pen evens out.

@jgruberman If you are referring the to Realistic Build calculations a few posts above, then yes. If referring to the OP then no. No other conditions are considered there than: Max Magicka, Spell Damage, Crit Chance, Crit Damage, and Spell Penetration, and how those variables are effected by the various sets.

Texecutioner187

dpencil1 wrote: »

@jgruberman If you are referring the Realistic Build calculations above, then yes. If referring to the OP then no. No other conditions are considered there than: Max Magick, Spell Damage, Crit Chance, Crit Damage, and Spell Penetration, and how those variables are effected by the various sets.

So then using Divines gear and Infused for a chest/legs/head pieces versus all Divines makes little difference in the overall calculations and realistic outcomes? (can you tell I'm already planning my HotR sets? )

dpencil1

@jgruberman
The difference between going all Divines vs Big Infused / Small Divines has always been pretty small and isn't really changing in HotR. Some of the Mundus Stones are getting buffed, so full Divines would provide a little more Mundus power than it does on Live, if you are using those stones. The bigger issue is which Mundus Stone you choose. You'll want to choose a Mundus that helps compliment your build, which may be different depending on the gear sets you choose and your overall goals for the build.

Texecutioner187

dpencil1 wrote: »

@jgruberman
The difference between going all Divines vs Big Infused / Small Divines has always been pretty small and isn't really changing in HotR. Some of the Mundus Stones are getting buffed, so full Divines would provide a little more Mundus power than it does on Live, if you are using those stones. The bigger issue is which Mundus Stone you choose. You'll want to choose a Mundus that helps compliment your build, which may be different depending on the gear sets you choose and your overall goals for the build.

That's a very good point, I didn't think about that. It seems like a lot of the sets outlined in OP would be better geared towards Thief if I'm not mistaken.

dpencil1

@jgruberman
Well, Thief is the one Mundus they decided to nerf. It will give 9% Crit Chance instead of 11%, which will be equal to the 9% you will get from Precise weapons. If your crit chance is high enough already (usually above 50-60%), you'll actually get better mileage out of the Shadow stone, boosting crit damage. Other options are the Lover, which will give Penetration, and the Apprentice, which got boosted to give more Spell Damage. The Mage is there too, but it is weaker than the others and only a good choice if you are specifically going for a very high max magicka build.

dpencil1

@geonsocal
If I'm going to look at Togue's Pact, can someone post the HotR set bonuses for it and also the value of a Wep Dmg glyph with both Infused and Torgue's buffs applied to it?

geonsocal

dpencil1 wrote: »

@geonsocal
If I'm going to look at Togue's Pact, can someone post the HotR set bonuses for it and also the value of a Wep Dmg glyph with both Infused and Torgue's buffs applied to it?

ya know @dpencil1 ...just lost my boss last week (moved his family to utah)...

great guy, whom not only played a bit of eso (mostly though into skyrim and fallout 4) but was an absolute data/spreadsheet animal...

which, I am most definitely not (although I can type up some fierce emails)...

most relevant though - he was a really useful human...

you wouldn't so happen to be interested in moving to so cal to become a logistics manager would ya

Betsararie

Very informative thread. I've been studying this like I'm back in calculus class back in school.

I cotinually see spinnner outperforming Julianos in this thread, found that interesting as the Julianos/Mother's Sorrow combo has been being touted. Fortunately, I have both sets.

dpencil1

@Blanco
Alcast is not going to favor Spinner's because he is making builds for a high level trials environment. There, it is much easier to get to the penetration cap, so needing more pen from Spinner's isn't important.

MLGProPlayer

jgruberman wrote: »

dpencil1 wrote: »

@jgruberman If you are referring the Realistic Build calculations above, then yes. If referring to the OP then no. No other conditions are considered there than: Max Magick, Spell Damage, Crit Chance, Crit Damage, and Spell Penetration, and how those variables are effected by the various sets.

So then using Divines gear and Infused for a chest/legs/head pieces versus all Divines makes little difference in the overall calculations and realistic outcomes? (can you tell I'm already planning my HotR sets? )

The difference was always <1% (not sure why people insisted on it since it literally made no difference in parses).

Izaki

That moment when you know that on your mag sorc which you haven't touched since homestead, both of the sets you've golded out (necro and BSW) are the worst. At least I'm going to get some dreugh wax out of deconstructing it lol

Sacredx

thanks for doing the hard work working out all the numbers.

Can you please explain how according to your cumulative chart (incl G) shows a 12% reduction between old and new necro set when the set difference in max magicka is only 6.7%. Thus I'd expect maybe around 4% global reduction, not 12%?

And how did you get 7288 max magicka for the current necro set (no nerf), based on current ingame stats isn't it 6901 (967x3 + 4000)?

dpencil1

@Sacredx
The Cumulative Comparison adds together all the values of Comparison A-G. This amplifies the difference between sets when one set continually performs worse and another set continually performs better on various comparisons. If, for example, some of the comparisons had pushed one set to the top, say because a boost to max magicka would amplify that set's power, while a boost to spell damage would amplify another set the Cumulative Comparison would even out the difference. But since Necro consistently performed worse on all comparisons and Spinner consistently performed better, the Cumulative Comparison pushes them farther apart than they are on the individual comparisons, which are available near the bottom of the OP, if you'd like to view those.

I'm not quite sure I follow you on the second question. The OP doesn't use Live values at all, so every magicka bonus there is 1096, not 967.

Texecutioner187

I have some questions, sorry... If I need to just put the thread down and stop playing, feel free to say so. The data is just extremely interesting to me.

-What is defined as low stats and high stats?
-When you say "fully buffed" what does that consist of?
-The percentage of penetration in your buffed calculations is based off a total amount of 18,200?
-How is the percentage of crit damage derived? Is that based off a stat or a percentage of a target value?

dpencil1

@jgruberman
I used a set of arbitrary baseline values that I then applied all the set bonuses to. You can see these in the calculations at the bottom of the OP:
Magicka: 30000
Spell Damage: 2000
Crit Chance: 40%
Penetration: 80%
Additional Crit Damage: 50%

So the very first calculation you see under the last set of graphs is Comparison A Baseline. That is simply these values plugged in to the UESP Build Editor Effective Spell Damage calculator, with the resulting number given below.

I wanted to see if increasing a particular value would have an impact on how a set performs, so I made Comparisons B-F. In each of these I boosted one value:
B boosted Magicka +10k to 40000
C boosted Spell Damage + 1k to 3000
D boosted Crit Chance +20% to 60%
E boosted Penetrarion +10% to 90%
F boosted Crit Damage +20% to 70%

The last comparison, G, included all of these stat boosts. So that's what I mean when I mention "fully buffed." And the low stat version refers to the Comparison A values.

There was nothing particularly scientific about the initial stat conditions or the amount each was buffed. They were chosen simply to approximate a reasonable set of build stats and the buffs were adjusted to all have a similar total impact on the build's ESP.

To answer the 3rd question, yes the % penetration refers to an 18200 ceiling.

Crit Damage by default is 150% of an attack's normal damage. Using things like Elfborn or certain passive skills (like on a Nightblade) can increase this percentage.

Texecutioner187

dpencil1 wrote: »

@jgruberman
I used a set of arbitrary baseline values that I then applied all the set bonuses to. You can see these in the calculations at the bottom of the OP:
Magicka: 30000
Spell Damage: 2000
Crit Chance: 40%
Penetration: 80%
Additional Crit Damage: 50%

So the very first calculation you see under the last set of graphs is Comparison A Baseline. That is simply these values plugged in to the UESP Build Editor Effective Spell Damage calculator, with the resulting number given below.

I wanted to see if increasing a particular value would have an impact on how a set performs, so I made Comparisons B-F. In each of these I boosted one value:
B boosted Magicka +10k to 40000
C boosted Spell Damage + 1k to 3000
D boosted Crit Chance +20% to 60%
E boosted Penetrarion +10% to 90%
F boosted Crit Damage +20% to 70%

The last comparison, G, included all of these stat boosts. So that's what I mean when I mention "fully buffed." And the low stat version refers to the Comparison A values.

There was nothing particularly scientific about the initial stat conditions or the amount each was buffed. They were chosen simply to approximate a reasonable set of build stats and the buffs were adjusted to all have a similar total impact on the build's ESP.

To answer the 3rd question, yes the % penetration refers to an 18200 ceiling.

Crit Damage by default is 150% of an attack's normal damage. Using things like Elfborn or certain passive skills (like on a Nightblade) can increase this percentage.

Got it... So fully buffed is simply all of the buffs added in B-F. As far as what one would need to achieve that particular level of buff all depends on variables like Mundus, CP allocation, enchants, armor/weapon traits, etc?

So when I'm in the game, how would I determine my crit damage % to better align myself with these comparisons?

dpencil1

Right. I don't spell out how you would arrive at these stats, it may not even be possible to exactly match them. The whole point was simply comparing the various sets against a static base, and since the sets behave differently with different base stats, it was important to examine more than just one stat arrangement.

I don't think anything but Elfborn is going to be effecting you, so just look at what % Elfborn is giving you, say 15% and add that to 50% = 65%.

You can then verify this by attacking a target with an ability and seeing that skill's normal and crit damage. If you have 65% more damage on crits then a 10000 non-crit attack will turn into 16500 when it crits.

Texecutioner187

dpencil1 wrote: »

Right. I don't spell out how you would arrive at these stats, it may not even be possible to exactly match them. The whole point was simply comparing the various sets against a static base, and since the sets behave differently with different base stats, it was important to examine more than just one stat arrangement.

I don't think anything but Elfborn is going to be effecting you, so just look at what % Elfborn is giving you, say 15% and add that to 50% = 65%.

You can then verify this by attacking a target with an ability and seeing that skill's normal and crit damage. If you have 65% more damage on crits then a 10000 non-crit attack will turn into 16500 when it crits.

Got it. That makes sense. 50% is the base crit damage percentage so it's just a matter of increasing beyond the baseline.

STEVIL

dpencil1 wrote: »

@SodanTok
Thanks for clarifying. As I said in the OP, how someone decides the value of max magicka over other stat boosts is largely subjective. If you know of a way to quantify it objectively I would honestly love to see that because I have been wracking my brain.

I showed earlier that the amount of extra shields compared to another set with 2 max magicka and 1 spell damage bonus is about 3300 in PvP. 3900 witout a nerf. I'm not sure how you would quantify how much lower on the Effective Spell Damage ranking Necro should be because of this. Same thing with pet damage or having more resources. And should we put points back on for it scaling badly with weapon attacks and forcing you to have a living pet?

If shield size is the problem they should address that directly with diminishing returns rather than hurting PvE dps.

The key is that you cannot.

"value" is tied to need and need is tied to challenge and challenge is tied to content.

if you run solos (dungeons or vma or whatever) you are responsible for your own healing and defenses and survival completely - so in those cases shields are very important as are pet heal size and pet durability etc. The maxmag sets help more than dps for that content, right?

if you run organized g/t PVE content then your own defense and healing is less important ads dedicated and skilled healers and tanks draw so much of that pressure off you.

if you run PVP, its a different world and even within that world whether you are talking duels, open pvp, zergs, mega-keeps or whatever drastically all shift those aspects in importance. Just like in that content DPS vs burst shifts compared to PVE.

much the same way that target dummy parses are not a "true" reflection but more of a funhouse mirror, the math alone or isolated set pieces cannot account for all of these variables.

This is not to dismiss all math and all analysis but the more you isolate one element and use that one element to reach conclusions that really show relatively minor differences even within that shoebx, the less relaible or convincing your conclusions become.

The best "analysis" is of play results and what happens there.

LiquidPony

dpencil1 wrote: »

@Subversus
There is not going to be one BiS or go home this patch. The difference between all of these set combinations is very small. If you look at the OP, you'll see that I mention above the Combined Set (fully buffed) chart that I remove 10000 points from each to better show their difference. This is what that same chart looks like with the full values applied:



So the difference from the highest value and the lowest value:
Spinner/Mother:11012
BSW/Necro: 10428
is only 584...that's only a 5.6% difference.

And if we exclude the lowest performing set in the list, Necropotence, the gap is even smaller:
Spinner/Mother: 11012
Julianos/BSW: 10733
Only 279!...that's only a 2.6% difference!

So just about all of the combinations listed will be viable. It will come down to what will work best for you given other factors you can use to increase the effectiveness of those sets, and what your overall playstyle is, as well as group composition.

Correct me if I'm wrong ... but you say you removed "10,000" points from each Effective Spell Power result to better show the difference?

As in, from the graph above, the actual values are Spinner/Netch: 20,959, Spinner/Julianos: 21,005, Spinner/Necro: 20,667, etc.?

If that's the case, you should normalize the values, not subtract some arbitrary number. In the example above, Spinner/Julianos is actually ~1.6% better than Spinner/Necro, but if you just subtract 10,000, it appears that Spinner/Julianos is ~3.2% better.

So that turns a 40k parse vs a 40.6k parse comparison into a 40k parse vs 41.3k parse comparison.

Or, in the worse case of BSW/Necro (10,428 vs. 20,428) vs. Spinner/Mother's Sorrow (11,012 vs. 21,012), the latter option is represented as being 5.6% better (40k vs. 42.24k) when it should actually be ~2.9% better (40k vs. 41.16k), if I understand your methodology correctly.

catch22atplay

Don't give up on Necro yet. I certainly am not getting rid of mine. If you just use the eso build editor as Pencil suggests you can come to your own conclusions. It's more accurate then anything i've seen yet. I'm at 12,730 SP with 451cp atm and would be 14,067 SP at cp630. This will just increase come monday.

Sacredx

@dpencil1

Thanks for clearing that up! Yes that makes sense now, comparing future necro (as proposed) vs future necro (with 4k 5th bonus variation).

Interesting that individual set comparison is very close together. Given that the scores represent effective spell power only you should consider adding pros/cons to each set describing other benefits that they bring. eg. Necro set gives highest max magicka which allows more casts and effectively more short term dps. It also gives higher shields and provides a global buff that is not tied to damage type or targets armor (pen) or if the target has a shield active (crit).

Just some food for thought, it's not just about effective spellpower, every aspect needs to be considered to be able to make a sound decision.