DPS Spreadsheets

[Motenten]

Ugh. Yep, an unfortunate bug. The acc total (both melee and ws, both sets) included the total of gear att, rather than gear acc. It will likely radically alter your results.

Thanks for finding the problem.

[Xilk]

Is there a good way to use the spreadsheets to compare dps when you want to alternate weaponskills? Mainly to keep up AM3 from a mythic.

I see the ratio section for averaging the dps totals of the 2 different sets for comparison, but that does not appear sufficient.

I'm sure it applies to plenty of jobs, but often the mythic WS is not as fantastic as something else, but the AM3 is ofcourse unparalleled.

[Motenten]

Set 1: Mythic weaponskill. Set minimum TP to 300.
Set 2: Normal weaponskill. Set minimum TP to 100.
Determine time interval for Set 2, and figure out how many weaponskills you can manage in the time that AM3 is up, after deducting the time for get TP for activating AM3 from Set 1. Depending on job, you may be able to improve on the basic numbers by use of special abilities.
Use the two time periods for the averaging values.

[augus]

On the Mnk spreadsheet on the Data tab, formula on cells H15 and H16 is missing a 'if(isblank(...' statement on the first part of the sum, in this case if there is no footwear equipped on the second gear set, both cells produce #N/A error that goes up to results.

I don't post often but thanks a lot to all people that works on this information.

[Motenten]

Fixed. Thanks.

[Byrthnoth]

Skadi body is still set to 5% Dual Wield in the DNC spreadsheet. (JP reports indicate that it should be 7%)
Evisceration is probably 10%/20%/40% instead of 10%/15%/15%.

For some reason switching from Potestas to Charis at low Haste is a better decision than at high Haste. I'm Acc capped and don't see how critical hit rate/damage vs. attack could possibly depend on attack speed. Any idea how it could?

[Motenten]

Updated Skadi body to 7% DW for thf/bst/dnc
Evis should be 10%/10%/20% if using my formatting (each one is the delta with the one before), unless you mean it progresses from 10% to 30% to 70%?

For some reason switching from Potestas to Charis at low Haste is a better decision than at high Haste. I'm Acc capped and don't see how critical hit rate/damage vs. attack could possibly depend on attack speed. Any idea how it could?

Assuming high haste includes marches, you're probably getting an extra 8-10 dex from the brd AF3 bonus, pushing you past the dDex cap. That means the 5 dex on Charis isn't contributing to crit rate, lowering its net value.

[Motenten]

Ok, found your conclusion post on FFXIAH, so I'll update the spreadsheets that have Evisceration.

[Stanislaus]

Hey, it's me again. Before I start, I'll again say thanks for the work you put into creating and maintaining this project. I've got nothing but respect for the massive undertaking that you... undertook.

My query today is about your implementation of regain in the WS cycle. Are you sure it's mathematically valid to use average regain per ws, when regain comes in discrete increments? It seems to me that because of regain's incremental nature, you can't have a regain contribution that isn't a multiple of the increments that regain comes in. E.G. the game doesn't let you WS at 99 TP with 2/3 of an embrava tic complete.

So I guess my full question is: Is there some precedent you're operating under that makes the average regain a valid model of in-game regain's incremental nature?

[Motenten]

That would be a valid argument if the TP return from the weaponskill and melee were a discrete value. However since those values are averaged across a wide variety of possible hit combinations, it's impossible to say which particular incidence maps to the amount of TP you get from Regain. As such it just uses the averaged return value.

[Fredjan]

I re-downloaded the DRK DPS spreadsheet and noticed an error when trying to enable Souleater. Both options result in Err:504 (I'm using OpenOffice Calc) for anything related to WS damage and total DPS, so the error seems to involve the Weaponskill section. I looked into that and everything related to the SE Dmg, Ttl Dmg, and P(SE Dmg) results in Err:504 when Souleater is turned on for both "WS only" and "on". This applies to both Set 1 and Set 2.

[Motenten]

I'm assuming you're referring to the Weaponskill page. Can you check the Melee page's Souleater sections, see if anything in there is generating an error? I'm not able to find anything wrong when using Excel, so it's probably an incompatibility with Calc's handling of the code.

[Byrthnoth]

Thanks for all the updates! One minor and one major request (either of which you're free to ignore):

Minor - Your skillcap calculator here still has DRG's parrying skill being C, though now it is B-.

Major - If you get a chance, could you adjust the DNC spreadsheet so that using a non-weapon sub would work?

[Motenten]

Your skillcap calculator here still has DRG's parrying skill being C, though now it is B-.

Easy enough to tweak, but BG wiki still lists it as a C as well.

If you get a chance, could you adjust the DNC spreadsheet so that using a non-weapon sub would work?

Does setting the offhand weapon to "None" not work?

[Motenten]

Also, added Fencer for dnc if /war with no offhand weapon.

[Byrthnoth]

Doh! I just kept hitting delete on the sub weapon instead of looking for a "none" option. Also, discovering that it's wrong on bgwiki is how we found out to tell you.


For some reason I get different results for number of attacks per round when I do it by hand than when I plug it in to the spreadsheet. I can't exactly follow what is going on in your matrix, but here's what I have:
4% QA, 8% TA, 35% DA, Mythic OA2-3, 95% hit rate

Plugging it in to the spreadsheet, it tells me I am hitting 1.939 times per swing, but when I work it out in a different excel sheet I cannot get that number out:

Spoiler: show

4% QA * 3 = 0.1200
8% TA * ( 1-4% QA) * 2 = 0.1536
35% DA * (1-4% QA - 8% TA*(1-4% QA) ) = 0.3102
Mythic OA2-3 (1.8 attacks per swing) * (1-4% QA - 8% TA*(1-4% QA) - (1-4% QA - 8% TA*(1-4% QA))*35% DA ) = 0.4891
Sum + 1 base = 2.0729
95% hit rate = 1.969
Spreadsheet says: 1.939

Cases:
* If I delete all the gear and just use OA2-3 with /NIN, it tells me 1.708 (1.8*.95 = 1.71), so that's only slightly off.
* If I put on Saber Dance (21% DA in the sheet), it says 1.150 (1.1495, so correct)
* If I put on Thaumas Coat with no other multi-attack, it says 1.118 (1.1176, so correct)
* If I put on Saber Dance and OA2-3, it tells me 1.748 when I get 1.7499 (should round up to 1.750)
So it's something about the interaction of OA2-3 and DA/etc. Each trait behaves fine alone, but together they give the wrong results or results that use different assumptions somehow.

Edit: Solved a minor problem. It was listed as 67%/33% for Mythic OA2-3 in the spreadsheet, so I changed it to 2/3 and 1/3. That fixed the OAT-alone case but not the combination case.



Also, I'm now fairly convinced that Terpsi/None is worse than Terpsi/Twash even at the delay cap. >:-/ It's a combination of the fact that WSing takes time (approximately 1/3 of the time it takes us to build 100TP) and that Twashtar's 20 DEX is useful for dDEX. That's kind of lame. I was looking forward to having a new way to play. Oh well~!

[Motenten]

Also, discovering that it's wrong on bgwiki is how we found out to tell you.

Ok, Skill Calculator has been updated.

Reviewing the hits-per-round stuff.

[Motenten]

Math:

Spoiler: show

Single-wielding Terps 99. 95% hit rate.

4% QA - 1.0640
Remaining: 100% - 4% = 96%
4% @4 + 96% @1 = 1.12 * 95% = 1.064
match

Add 8% TA - 1.2099
8% of 96% = 7.68%
Remaining: 96% - 7.68% = 88.32%
4% @4 + 7.68% @3 + 88.32% @1 = 1.2736 * 95% = 1.20992
match

Add 35% DA - 1.5036
35% of 88.32% = 30.912%
Remaining: 88.32% - 30.912% = 57.408%
4% @4 + 7.68% @3 + 30.912% @ 2 + 57.408% @1 = 1.58272 * 95% = 1.503584
match

Add AM3 (60% proc, 2/3 2x, 1/3 3x) - 1.9399
60% of 57.408% = 34.4448%
Remaining: 57.408% - 34.4448% = 22.9632%
4% @4 + 7.68% @3 + 30.912% @ 2 + 34.4448%*(2/3)*2 + 34.4448%*(1/3)*3 + 22.9632% @1 = 2.041984 * 95% = 1.9398848
match


Only 35% DA + AM3 - 1.7765
35% @2
65% * 60% = 39% AM proc
Remaining: 100% - 35% - 39% = 26%
35% @2 + 39%*(2/3)*2 + 39%*(1/3)*3 + 26% @1 = 1.87 * 95% = 1.7765

So manual calculations seem to match what the spreadsheet is saying.

35% DA * (1-4% QA - 8% TA*(1-4% QA) ) = 0.3102

4% QA
(1-4%)*8% TA = 7.68%
(1 - 4% - (1-4%)*8%) * 35% DA = 0.30912

0.30912 != 0.3102

Error occurred here?

[Byrthnoth]

Ah! Yes, I see it now. My excel calculations were slightly different than what I wrote. They calculated everything as independent chances of occurring, which doesn't make sense. Good deal then.

[Fredjan]

I'm assuming you're referring to the Weaponskill page. Can you check the Melee page's Souleater sections, see if anything in there is generating an error? I'm not able to find anything wrong when using Excel, so it's probably an incompatibility with Calc's handling of the code.

Aye, I was referring to that. Seems to be so, yes. I looked into the error itself, that error represents a "parameter list" error, "Function parameter is not valid, for example, text instead of a number, or a domain reference instead of cell reference." So yeah, it's just incompatibility. And no, nothing in the melee page was generating errors.

Edit: Just tried LibreOffice, the error is not occuring with that program, so it's OpenOffice specific. Incompatibility confirmed.