Archived

This topic is now archived and is closed to further replies.

lrdx

[Elemental] ESSE stat spreadsheet (updated to 4.0.1)

47 posts in this topic

As there is no Elemental theorycraft tool for 4.0.1 (yet?), I have re-written ESSE to have at least a rough image of things. (Note: ESSE isn't even trying to be the alpha and omega of Elemental Shaman theorycrafting, treat it accordingly)

It is available on Google Docs as usual:

https://docs.google.com/leaf?id=0B5clvbTTjpCDNzZjNWJkMDMtZDAzOC00NGNkLTg2NWYtOGVkYjI3YTMzOTIy&hl=hu

If you have Google account (GMail is one), log in, open the spreadsheet, and create a copy of it to edit (File -> Create copy). If you do not, you can try exporting it to Excel or OpenOffice formats (File -> Export.. -> Excel or OpenOffice. After you got an editable copy, set up the settings (yellow background cells), and receive the numbers in the green cells. If you want to know more, check other tabs.

The sheet does not model spell rotation / priority, but instead spells are weighted using the amount you cast - enter statistics from your raids (e.g. using WoL) to set them up correctly. This also means the spreadsheet has limited coverage of Elemental shaman "features", only models Lightning Bolt, Lava Burst, Flame Shock and Earth Shock (with Flumination). If you feel any other spell should be on the list, please let me know.

If you see any bugs / errors / mistypes, also let me know.

Share this post


Link to post
Share on other sites

I exported this to Excel type, but I am getting #N/A as values in many locations. This is what the formula is listing as well, so I don't know what it's missing or where the value is being obtained from (I suppose a search of that cell address may give me something) Obviously you may not be able to control this, but I know nothing about GoogleDocs and have only better than average experience with Excel.

Share this post


Link to post
Share on other sites

Hello Lrdx,

I was wondering if you would consider a toggle for enabling and disabling the usage of Bloodlust/Heroism?

I'd like to use your sheet to compare with my own prorgam to see how the results relate, and I'm not quite savvy enough to just remove the haste calculation from your sheet.

As always, the work you (And others) do for the community is much appreciated.

Share this post


Link to post
Share on other sites

I can't really do anything about the Excel export, last time I've checked it worked fine.. I'll try to make a working xls version if Google won't fix it soon and upload it somewhere.

On the settings page now you can set the fight's length and if you use BL or not.

Also, I've updated the sheet with the last PTR build, added the new-old glyph. Looks like it's better than the LB glyph. (edit: modelled the GCD with the glyph wrong, made it too good. With the fixed model, the best 3 glyphs are still FS, LvB and LB)

Share this post


Link to post
Share on other sites

I have made a couple of updates to the sheet:

  • I have moved talent / glyph / raid composition settings from settings tab to the comparator tab, now you can compare glyphs or presence of different raid buffs.
  • Created a PTR page, currently it calculates with the new talent and glyph. Is there any other change it should be aware of? It is updated to build 9637 (altho it's not different from the previous 2 PTR versions :) )
  • Talent settings are done via copying an MMO-Champion talent calculator link to a given cell. If you don't like it doing this way, just overwrite the functions that parse a given talent a couple of lines below. Getting a clean live calculator link is tricky here, as MMO only links the PTR version on the main page, but change the version in the URL to 9551 (that is the build number of 3.0.9)
  • Corrected the Thunder Capacitor's calculations, the cooldown is only applied to the release of the bolt, not to gaining charges.
  • Cleaned up a few cells, they should be more readable now.

What is yet to come (aka TODO list):

  • More consumables.
  • Better support for different professions.
  • Add raid (de)buffs that increase damage by a given percent, like CoE, Inspiration. These won't change the scales, just the final DPS number.
  • Parse glyphs from the MMO talent link as well.
  • Parse Raid buffs from an MMO RaidComp link.
  • More cleanup. Some functions are over the complexity of understanding in 5 sec.
  • Anything you suggest.

Share this post


Link to post
Share on other sites

Updates, yay!

  • Updated to build 9658
  • Added all possible raid (de)buffs I know of.

TODO:

  • More consumables.
  • Better support for different professions.
  • Parse glyphs from the MMO talent link as well. (well.. not until Boubouille fixes it)
  • Parse Raid composition from an MMO RaidComp link.

Share this post


Link to post
Share on other sites

  • Updated to build 9684.
  • Removed calculations with Elemental Mastery, as I don't know how to handle the new version in the "official" patch notes (nerfed effect and duration, one instant cast), and it is not significant DPS boost.

Share this post


Link to post
Share on other sites

Some suggestions I would love to see:

Implement JC only gems (might be covered by better profession support)

Pawn value for metagems

Thanks for your work and special thanks for not using Excel

Share this post


Link to post
Share on other sites

If I change for example draenei aura (b52) to be true the corresponding values in line 52 stay false and therefore the spreadsheets result is rather strange

Share this post


Link to post
Share on other sites

If the shaman reaches hit cap using gems, +hit on gear is still valuable, as (s)he could swap in spell power gems instead of hit ones - to model this, my spreadsheet suggests to re-gem for each set of stats, and calculates with those values. The gemming strategy will select 2 blue gems for the meta, socket +hit until cap, then fills the rest with spell power.

I am currently hitcapt (with gems) and the value for more hit is 1.36 and for spellpower 1.21 and i am using 16 hit gems and 9 spl 8 hit gems.

If I would aquire 8 more hit on gear I would win 10 spellpower (from the gem): 8*1.21 = 10.88 // 10*1.21 = 12.1

If I would aquire 16 more hit on gear I would win 19 spellpower (from the gem): 16*1.21 = 19.36 // 19*1.21 = 22.99

I am rather confused cause I thought the left and the right side should be equivalent

Share this post


Link to post
Share on other sites
If I would aquire 8 more hit on gear I would win 10 spellpower (from the gem): 8*1.21 = 10.88 // 10*1.21 = 12.1

Uhm.. purple/orange colored gems have 9 spell power, 9*1.21 = 10.89. The other comes from that currently, gems are "half gems" for the table, so the 19 spellpower gem is actually just 2*9. Fix will come with the support of JC gems.

Share this post


Link to post
Share on other sites

new change new bug ;) You missed a "," (I think it is called semicolon) in the Pawnstring between yellow socket and meta socket

Share this post


Link to post
Share on other sites

A good addition to the pawn string might be a second one witch rates hit=0. It helps me to find out the best item-combination from the gear currently availably to me (and I have done it manually but have to re change it every update)

Share this post


Link to post
Share on other sites

Updated to build 9704: Added T8 set bonuses. No other changes (give me more free time!)

Share this post


Link to post
Share on other sites

Updated to live.

Removed the gemming system, as it just created a gazillion of input parameters, and I don't think it was worth it. Now there are two scales, one with, one without +hit - use the second only if you don't have any +hit gems. This means, now the value of hit is: max(actual value ; spell power value / 19*16)

Also revamped the raid composition settings. Please tell if you see any errors.

I still need to add the new relic.

Share this post


Link to post
Share on other sites

Irdx,

Rank 14 Lightning Bolt base damage range is 719-819. Information on wowhead.com and the like about Lightning Bolt damage on all ranks is outdated. You are using 715-815 lower and upperbound in your model.

And since I am unable to phantom your math in the spreadsheet I would very much like to know how you calculate Lightning Bolt and Lightning Overload damage from spellpower and base damage in simple math for validation.

Also would like to see the option of excluding raid buffs and just being able to enter a simple final bonus damage number.

Thanks in advance and I would like to hear from you soon!

Share this post


Link to post
Share on other sites

The old bounds 715-815 were true until 3.1, I have updated this in the beta sheets, looks like this change was left out when I merged changes back to the live sheets, thanks.

I use the standard

( base_damage + coefficient * spellpower ) * damage_multiplier

formula for all spells. For Lightning Bolt,

  • base damage is the average of 719 and 819
  • coefficient is 2.5/3.5+0.1 (from Shamanism)
  • damage multiplier consist of Concussion, LB glyph, and LO. Extra damage done/taken (de)buffs are calculated separately, seen as a last multiplier in "xy spell average damage" rows.

Bonus damage from critical strikes are calculated later (around line 150), together with CC/EO bonus damage multiplier.

LO's effect is considered also when calculating Clearcasting chances: every LB (or CL) are actually two different spells (the normal and the one from overload) with the same crit chanse, increasing the chance of proccing CC.

Edit: typos

Share this post


Link to post
Share on other sites

Lrdx,

Thanks for the explanation.

I am under the impression that your Lightning Overload modelling/multiplier is erroneous.

My Lightning Overload model:

719 - 819

+ bonus damage * (2.5 / 3.5 + 0.2)

+ Totem of Hex bonus damage * (2.5 / 3.5)

* 1.09 = Concussion + Glyph of Lightning Bolt

divided by 2

Using 2390 bonus damage + Totem of Hex.

Lightning Overload non-crit proc, no CC: 1646-1701.

Lightning Overload crit proc, no CC: 3442-3556.

Lightning Bolt non-crit, no CC: 3032-3141.

This correlates to about 99.98-99.99% of my in-game lower and upper bound Lightning Bolt and Lightning Overload non-crit and crit procs. It can differ a few damage points due to in-game conversion of floating point to whole integers.

Lightning Overload is using a higher coefficient than Lightning Bolt so you can not use a static multiplier of 10.92%. Indeed you will see your Lightning Overload multiplier to be scaling positively with bonus damage.

Calculating Lightning Overload at .9142857 versus Lightning Overload at .8142857 shows a multiplier of about 10.84% This would be a correct multiplier to use at 2390 bonus damage and having the relic Totem of Hex ONLY if all spells will crit.

After some close observations I also noticed the following.

Having 0 charges of CC up before casting a Lightning Bolt crit will not increase your Lightning Overload spell damage by 10%.

Having 1 charge of CC up before casting a Lightning Bolt will not increase your Lightning Overload spell damage by 10% even if your Lightning Bolt or Lightning Overload crits.

Having 2 charges of CC up before casting a Lightning Bolt will increase your Lightning Overload spell damage by 10%

Lightning Overload will not consume CC charges.

This also causes your Lightning Overload multiplier to scale positively with crit.

In general it would be more correct not to use a multiplier but an actual corresponding simulation or model.

Share this post


Link to post
Share on other sites

I was sceptic that Lightning Overload procs scale better with spell power than standard Lightning Bolts (+0.2/2 coefficient instead of +0.1/2), as it does not feel right, and no one in this site noticed this so far (or the forum's search function suck) so I went testing a bit on the lvl 80 dummy.

First, I've tried shooting at it naked: 0 spell power, to find out the total damage multiplier, what should be

C = (1+0.05+0.04)/2 = 0.545
(0.05 from Concussion, 0.04 from Lightning Bolt Glyph) The expected range of LO with any coefficient is: from 719*C = 391.855 to 819*C = 446.355 Doing a 10 min shooting, I found that the minimum hit (not crit) was 391 damage, max was 445. The sample is far too small to say anything with 99.99% certainty, but it looks like multiplier everyone assumed is correct, I don't have more time for further testing. (Although it would only need one hit out of the 391-446 range to confute it.) Next stop, fully geared except Totem of Hex, no buffs, 1925 spell power. With my original model, the range of the LO procs should have been:
((719 to 819) + (2.5/3.5+0.1)*1925 )*C = 1246,1425 to 1300,6425
With what you suggested:
((719 to 819) + (2.5/3.5+0.2)*1925 )*C = 1351,055 to 1405,555
With an another 10 min casting, min hit was 1353, max was 1403. Looks like you are right, I'll update the sheets according to this. I have tried recording CC's effect on LO procs, but I successfully managed to delete the log file before I could parse for it. Lame, huh? :rolleyes: Edit: I'd like to ask readers to confirm / confute that the LO damage is within the range
((719 to 819) + (2.5/3.5+0.2)*spellpower )*(1+concussion+LB glyph)/2

Share this post


Link to post
Share on other sites

Mainly readability changes, all formula should be straightforward now, but there are a few logic changes:

  • Lightning Overload is modelled as separate spells, instead of a flat modifier to LB / CL.
  • Removed hit calculation, that column was basically just a copy of the base column.. You should be hit capped before anything, anyway. Hit value is derived from other stats, modelling if you can replace a hit gem to another type.
  • Removed Thunder Capacitor calculation. Will be coming back soon with other trinkets.
  • Reverted setting Talents/Raid composition via a calculator link. Was a nice idea, but the old way is faster.

Share this post


Link to post
Share on other sites

Lrdx,

Thanks for the update.

I have been playing around with your spreadsheet comparing it to my own model and I was able to fix some of my own errors through that.

At the moment I'm using a basic priority rotation in my simulation with the option of waiting for Lava Burst which seems to be benefical up to even 0.5 seconds of wait time at my current interaction of bonus damage, haste and crit rating. Adding Chain Lightning cast on CD lowers DPS, also evident by your spreadsheet. I will try to find out a way to use Chain Lightning intelligently. For example, I feel casting Chain Lightning before FS or Lava Burst to be a bad move since it reduces your chance of getting LO versus a LB-LO and that reduces chance of getting CC uptime before Lava Burst cast.

I saw some "questions" in the red fields and I believe I have answered the question about how LO interacts with CC in my previous post already. Also, Totem of Hex does add bonus damage to LB-LO and CL-LO.

I was wondering how much chance Chain Lightning has to proc LO on 1st, 2nd and 3rd target. Seems you put it at 13% for 1st target. You did any testing on this? I have been to lazy so far. Also it's nice to see you have noticed that LO from Chain Lightning is just half damage, calculated at the normal Shamanism coefficient of 10% instead of 20% for LO from Lightning Bolt.

About Call of Thunder affecting LO. If Blizzard has any consistency in their tooltip's, which I believe they generally have. It would be logical to assume that LO is affected by Call of Thunder since LO is a second similair spell to the one casted.

Share this post


Link to post
Share on other sites

Binkenstein,

If you look back to my 2nd post in this topic you will see some of my "comments" at the end relating to CC bonus spell damage being applied to Lightning Overload.

After close observations I only saw CC bonus spell damage being applied to Lightning Overload only when 2 charges of CC were up before the casting of the spell Lightning Overload was following. In this case it was only tested with Lightning Bolt. I'd be happy for you to confirm it.

hit hit crit+lo (no cc-bonus spell damage applied)

hit crit hit+lo (cc-bonus spell damage applied)

crit hit hit+lo (no cc-bonus spell damage applied)

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.

  • Recently Browsing   0 members

    No registered users viewing this page.