WotLK DPS spreadsheet - Page 241

davepoobond · 09/24/09, 5:39 PM

Just a couple of weird things I wanted to note when I plugged all my numbers in this week (I clicked CalcAttributes and Recalculate DPS before looking at the Gear Planner):

Valorous Scourgestalker Headpiece shows as a 0.72 dps increase over Conqueror's Scourgestalker Headpiece

Valorous Scourgestalker Tunic shows as a 6.67 dps increase over Conqueror's Scourgestalker Tunic

alienangel · 09/24/09, 8:13 PM

Originally Posted by Esoth

Well serpent sting seems to follow this formula: (2*1.045-1)*1.15+1
In other words, it's using 50% of mortal shots but 150% of the meta gem. I have no idea why it's doing this - the formula they use is probably different but this one at least matches my results.

Continuing to poke at this due to a lack of something better to do:

The fact that they're only taking 50% of mortal shots would be easy to understand if they just did 2 changes to their forumula without making them work together - for example they originally had SS with a 50% crit bonus, and with 3.2.2 they made that bonus benefit from mortal shots. Then they hotfixed yesterday to add another 50% crit bonus, but didn't make mortal shots apply to that extra 50%. Using that formula I can match my and your results exactly because obviously it works out to the "50% of mortal shots" bonus your got earlier.

MetaBonus = 0.03
MortalShots = 0.3
MortalShotsBonus = (OldBonus+OldBonus*MortalShots+NewBonus) = 1.15

I've been trying to come up with an explanation for why this kind of slip would lead to 150% of the meta gem bonus, and it seems like a similar (but in this case beneficial) issue, the old and new bonus portions of the base modifier get the meta applied separately:

They're calculating the net bonus as:

MortalShotsBonus*(1+(1+OldBonus)*MetaBonus+(1+NewBonus)*MetaBonus) = 1.2535

giving us the crit modifier of 225.35% observed.

Basically since OldBonus and NewBonus are each 0.5, and we're adding each separately to the base damage before applying the meta, it looks like the meta gets applied 1.5 times.

pabelanger · 09/24/09, 11:52 PM

Missing items:
[Snub-Nose Blastershot Launcher]

Chigau · 09/25/09, 1:16 AM

There is an item stat discrepancy for [Paragon's Breadth]

The wowhead link above correctly shows the staff having 97 Haste and 63 ArP, whereas in the spreadsheet these values are reversed and show 97 ArP and 63 Haste.

Nachti · 09/25/09, 8:38 AM

When I change the 20 Agi Gem in my Belt to a 10 Agi + 10 Crit Gem, it shows me a DPS gain (~1.5 DPS). But when I switch the 20 Agi Gem in my Gloves to a 10 Agi + 10 Crit Gem, it shows me a DPS loss. Both are yellow sockets with a +4 Agility Socket Bonus. Pretty strange.

Shandara · 09/25/09, 8:44 AM

Update:

- Added Onyxia gear
- Beast Within now 50% mana reduction while active
- Added 2-hand entry for custom gear
- Serpent Sting crit modifier as per EJ calculations (Alienangel)

Wrayth · 09/25/09, 9:59 AM

Maybe I'm retarded, but I'm not seeing Ony gear in the latest update on the OP.

Gotrage · 09/25/09, 10:19 AM

The regemmer is working really well, but it can work in strange ways if you need to mesh diferent types of gems that could share the same slot. This case happens when someone gems armor penetration untill the soft cap (with trinktet) then starts gemming for agility to fill the other slots.

One thing I guess would be nice to implement in the regemmer is an option to create a soft cap for a particular stat, just like it's done with hit. Without implementing a specific cap for each stat. That would be crazy.

For example, you can set a soft cap to armor penetration at 735 and assign Fractured Cardinal Ruby to fill gem slots until that cap. Once at the cap, the regemmer starts socketing the default dps gem. Somewhat the way gemming for hit cap works.

Guess most of the weird things people see on the gear planner could be worked with something like this.

KergeKacsa · 09/25/09, 11:11 AM

I've got 2 problem with the spreadsheet:
- It doesn't save to file the "Max. time to wait for CD's" cell's value, so you cant restore it from files, you have to set it every new version.
- Gylph of Steady Shot is "build-in" in the Marked For Death talent (bug or intended, it doesn't matter at this point), but the spreadsheet doesn't calculate this.

cidica · 09/25/09, 11:57 AM

I seam to be running into an issue with this newest version also. When I load my gear from armory and goto the gear planer tab it tells me that I can regem and gain better dps. Now that is not normally an issue but when I do run the regemmer it lowers my dps by 2? It's moving my nightmare tear into my twin pack vs it being in my triumph leggings. I wouldn't mind but looking at it I do have a better gem setup so I would think.

Aker · 09/25/09, 3:10 PM

Due to the importance of the hit cap and variable talent builds with and without FA, is it possible for the Regem button to factor in FA when regemming? Perhaps defining somehow where to take the points out of and put into FA or where to put them when they come out of FA.

Or would that add too many loops into it and slow it down too much?

I find myself testing many slightly different builds by moving points in and out of FA and using and removing hit gems with slightly different builds.

Also, reporting that the regemmer tends to put Deadly Ametrine (agil/crit) gems into yellow slots over Delicate Cardinal Rubies, even when the socket bonus is not achieved either way, due to the other socket being mismatched, when using a Delicate Cardinal Ruby is a very slight dps upgrade (2-3 dps).

Kradish · 09/26/09, 7:20 PM

I've probably just missed this somewhere as I haven't been paying much attention recently but is there a reason that the L'n'L chance on the settings tab is now 4% down from 6.5% per talent point?

Koriani · 09/27/09, 2:07 AM

Originally Posted by davepoobond

Just a couple of weird things I wanted to note when I plugged all my numbers in this week (I clicked CalcAttributes and Recalculate DPS before looking at the Gear Planner):

Valorous Scourgestalker Headpiece shows as a 0.72 dps increase over Conqueror's Scourgestalker Headpiece

Valorous Scourgestalker Tunic shows as a 6.67 dps increase over Conqueror's Scourgestalker Tunic

Yes I noticed a few wierd things like that too..

For example - I have Arrowsong the bow equipped. I am a hunter. The spreadsheet is telling me that Veranos' Bane - a gun in Uld25 is an upgrade. This gun isn't even a hunter gun - it has NO hunter stats, and has defense and such on it.

But its ranking as 3rd in my Ranged Gear Planner...

*blinks*

Something seems a bit off...

Xenoborg · 09/27/09, 2:49 AM

Originally Posted by Koriani

Yes I noticed a few wierd things like that too..

For example - I have Arrowsong the bow equipped. I am a hunter. The spreadsheet is telling me that Veranos' Bane - a gun in Uld25 is an upgrade. This gun isn't even a hunter gun - it has NO hunter stats, and has defense and such on it.

But its ranking as 3rd in my Ranged Gear Planner...

*blinks*

Something seems a bit off...

Veranus' Bane has 48 more DPS than Arrowsong, how is it hard to believe it is an upgrade?

Rivkah · 09/27/09, 3:49 AM

Originally Posted by Kradish

I've probably just missed this somewhere as I haven't been paying much attention recently but is there a reason that the L'n'L chance on the settings tab is now 4% down from 6.5% per talent point?

I can't say for sure, but my guess would be that it's because the effective uptime rate is much lower than the actual uptime rate due to the cooldown so 4% is a more accurate representation. The sheet doesn't really have a good way to model in the LnL cooldown (last time I looked it seemed the rotation test basically simulates the periodic proccing but it doesn't do it in a random fashion unless specifically enabled, so it'll normally not run into the cooldown in the simulation).

The problem with having it as an estimated per talent point amount is that it doesn't actually scale evenly with amount of pts invested into LnL (1-2 pts will get much closer to the max procrate than 3 pts will), not to mention resourcefulness affects the effective uptime as well, but as long as people are only testing specs with 3 pts in LnL it should come out relatively on target.