The Roguecraft Spreadsheet - Page 82

Kospire · 02/13/09, 5:27 PM

Originally Posted by Zeyk_Shade

When I put this into my spreadsheet(I did update it with the proper SND/HFB changes), I lose 345.28dps.

Changed "Use Envenom instead of Eviscerate" to FALSE
same cycle as you mentioned
same glyphs as you mentioned
Undead Boss
No TotT

This nets 5228.34dps on the sheet vs. 5573.62dps with WD/SR and IP/DP using Envenom instead.

Is there something I am overlooking / doing incorrectly?

Modify the sheet to make the target always poisoned

MentalPROblem · 02/14/09, 9:02 AM

Originally Posted by Zeyk_Shade

When I put this into my spreadsheet(I did update it with the proper SND/HFB changes), I lose 345.28dps.

Changed "Use Envenom instead of Eviscerate" to FALSE
same cycle as you mentioned
same glyphs as you mentioned
Undead Boss
No TotT

This nets 5228.34dps on the sheet vs. 5573.62dps with WD/SR and IP/DP using Envenom instead.

Is there something I am overlooking / doing incorrectly?

Great thanks for noticing my post/question!
Here's where you can check how to "trick" the spreadsheet to think that there is always poison applied
http://elitistjerks.com/1097801-post1103.html
Would be HIGHLY appriciated if someone else would post your comparison of this setup and combat builds.
Because I do believe that that mutilate spec is the best spec as of now, but I still haven't seen any confirmation yet :S

Narumi · 02/14/09, 1:07 PM

Why do I lose like 4 dps when I take away 3 talents from Ruthlessness. It makes no sense to me.

Also I modified spreadsheet, and is it me or dual instant blows away usual Mut setup.

• Vulajin · 02/14/09, 5:09 PM

Finally got around to whipping up version 0.4.5 of the spreadsheet. As always, it's available via the download link in the first post. Changes:

- Mutilate now only receives a 20% bonus on poisoned targets, and always assumes targets are poisoned.
- Slice and Dice now provides a 40% haste bonus.
- Hunger for Blood now provides 15% increased damage.
- Changed cycle calcs to assume 100% Slice and Dice uptime initially to allow the model to reach equilibrium more quickly.
- Removed one each of the "DPS Calc" and "Cycles" sheets.

Basically this version brings the sheet up to speed with patch 3.0.9. In addition, the last couple changes in that list are basically step 1 in a major effort to streamline the calculation paths in the sheet. This is going to be necessary if I ever want to expand the cycle model, because as it was, adding any new calculation paths (for example, for a more dynamic CttC model, or for HAT) was going to be a gigantic pain in the ass. My goal is to have two DPS Calc sheets and one Cycles sheet, though that may or may not be a pipe dream. At worst I'd have two of each. We'll see what I can manage.

If I've inadvertently broken something while removing those two sheets, please let me know. A cursory examination showed no problems, so I'm reasonably confident, but we'll see.

onkl · 02/14/09, 6:38 PM

Are you going to change expertise truncation too or don't you consider this as proven yet?

• Vulajin · 02/14/09, 6:50 PM

Expertise truncation was changed, I just forgot to include a note in the version history.

(edit) Or maybe it wasn't, because I'm stupid and can't keep the various versions I'm working on straight in my head. It'll be in the next version, and there are instructions on an earlier page to make the change yourself for now.

Tunus · 02/14/09, 7:50 PM

The highest rank of Skinning bonus, Master of Anatomy, was changed a while ago to 32 crit rating, instead of 25.

Valustria · 02/14/09, 8:01 PM

Originally Posted by Vulajin

Finally got around to whipping up version 0.4.5 of the spreadsheet. As always, it's available via the download link in the first post. Changes:

*SNIP*

Thanks, naturally... one quick question: any chance you added [Enchanted Tear] as a gem option?

Grunge · 02/14/09, 8:32 PM

Originally Posted by Vulajin

Finally got around to whipping up version 0.4.5 of the spreadsheet. As always, it's available via the download link in the first post. Changes:

- Mutilate now only receives a 20% bonus on poisoned targets, and always assumes targets are poisoned.
- Slice and Dice now provides a 40% haste bonus.
- Hunger for Blood now provides 15% increased damage.
- Changed cycle calcs to assume 100% Slice and Dice uptime initially to allow the model to reach equilibrium more quickly.
- Removed one each of the "DPS Calc" and "Cycles" sheets.

Basically this version brings the sheet up to speed with patch 3.0.9. In addition, the last couple changes in that list are basically step 1 in a major effort to streamline the calculation paths in the sheet. This is going to be necessary if I ever want to expand the cycle model, because as it was, adding any new calculation paths (for example, for a more dynamic CttC model, or for HAT) was going to be a gigantic pain in the ass. My goal is to have two DPS Calc sheets and one Cycles sheet, though that may or may not be a pipe dream. At worst I'd have two of each. We'll see what I can manage.

If I've inadvertently broken something while removing those two sheets, please let me know. A cursory examination showed no problems, so I'm reasonably confident, but we'll see.

Is Dual IP "fix" included in this one?

Genre · 02/14/09, 9:24 PM

I did some testing on the PPM of mongoose. You currently list it at 1 however, some tests using Calamity Grasp in the MH and a caster OH indicates otherwise.

Before I proceed, I just want to make sure I follow properly how PPM works. Technically the chance of a weapon to proc is equal to PPMx(weapon speed)/60. Example, my Grasp was at 2.37 delay thus giving me a chance to proc of 3.95% per swing if the PPM is at 1.

Thus a preliminary test (that I will update when I have more time) and with the above set up, I had 40 procs out of 611. My initial assumption is that PPM0 = 3.95%. What I found is 6.55% which we will call PPM1.

p_PPM1= 6.55%
q_PPM1= 93.45%
Std(PPM1)= 0.0100 who is SQRT(pq/n)

Assume that:
1. we ignore the fact mongoose increases you attack speed for PPM calculations.
2. we are computing the std using the estimators of p and q.
3. each swing is totally independant.

I find that 19 times out of 20, the PPM will be between 1.15 and 2.15. It means we can, with quasi certainty, reject our basic assumption that it is equal to 1.

Anyways, maybe it should be higher.

• Aldriana · 02/14/09, 10:48 PM

I'd be hesitant to draw any conclusions from 611 attacks. Assuming it's 1 PPM, your data is 3 standard deviations from the mean - unlikely, certainly, but not so unlikely as to rule out a statistical fluke.

In the early days of the expansion, I recall testing the proc rate using my Webbed Death, which, at the time, had Mongoose on it. And I recall concluding that the PPM was probably around 1. Regrettably, I have mislaid the original parameters of the test, so I can't swear as to the details, but that's my recollection. And, in point of fact, my best guess at the parameters of the test puts the proc rate even lower than 1 PPM - I'm getting .8 or so, though I make no claims that I haven't misremembered the details of the test. So I'd suggest that we perform slightly more systematic and lengthy testing - 5 or 10 thousand attacks at least - before jumping to any conclusions.

Rilias · 02/15/09, 3:43 AM

Saving with Open Office (in original format not ODF) breaks the the Envenom DPS Cell B48 on DPS-Sheet (the one that is never hidden if there are more than one) for Microsoft Excel 2007. Other DPS numbers seem to work and in Open Office the Sheet works as normal.

The formula in the cell itself looks the same and Copy pasting it over from the original sheet does not fix it. My guess is some underlying variables formula is formatted by Open Office and Excel does not get it.

Problem is the same in new and old version of sheet.

Edit: Yes the Mutilate damagebonus seems to be enabled by default. In other words the sheet assumes the target to be poisoned at all times.

Rilias · 02/15/09, 4:26 AM

@Vulajin: Did you change anything with the cycles? I tried realy hard to enter the exact same information in the two sheets and somehow the new one comes up with a thighter cycle 18.35s down from 18.80s and thus more DPS.
Stats in the DPS sheet are almost identical (I somehow have 2 AP more) but all in all the difference seems to come from the shorter cycle and nothing else.

I did edit the old sheet to account for all the 3.0.9 changes in what I presume is the correct way.

PS: The True/False trigger next to Bloodlust has no effect and thus should be removed to not confuse people.

kwinto · 02/15/09, 5:48 AM

Originally Posted by Rilias

@Vulajin: Did you change anything with the cycles? I tried realy hard to enter the exact same information in the two sheets and somehow the new one comes up with a thighter cycle 18.35s down from 18.80s and thus more DPS.

In my case (Combat) my 5s/5r/5e cycle build went from 28.17s to 28.24s (lower DPS) and of course a bit lower DPS (same settings, same gear, same buffs). I investigated it to find that the only cell responsible is Cycles!B1 (Energy per second) - inserting exact value from 0.4.4 resulted in reverting back to 28.17 and almost same old DPS. Further investigation shown that previous value of this cell referred to cycleslice2 cell which had value of 100% while now it refers to cycleslice1 cell with value of 96%. I've lost myself there and have no idea which calculations are correct and which are not, so until somebody else will explain it better, let's assume that new version is simply more accurate.

Edit: I can bring more details on my specific gear and settings if needed

Genre · 02/15/09, 9:06 AM

Originally Posted by Aldriana

I'd be hesitant to draw any conclusions from 611 attacks. Assuming it's 1 PPM, your data is 3 standard deviations from the mean - unlikely, certainly, but not so unlikely as to rule out a statistical fluke.

In the early days of the expansion, I recall testing the proc rate using my Webbed Death, which, at the time, had Mongoose on it. And I recall concluding that the PPM was probably around 1. Regrettably, I have mislaid the original parameters of the test, so I can't swear as to the details, but that's my recollection. And, in point of fact, my best guess at the parameters of the test puts the proc rate even lower than 1 PPM - I'm getting .8 or so, though I make no claims that I haven't misremembered the details of the test. So I'd suggest that we perform slightly more systematic and lengthy testing - 5 or 10 thousand attacks at least - before jumping to any conclusions.

I performed further testings and found some interesting tid bits.

After 3660 HITS, it procced 258 times. This means the proc rate should be between 1.57 and 2.00 -- 19 times out of 20. However, this is not the reason I am posting this. Could it be that the proc mechanic concerns itself with the level of the mob you are fighting or that glancings are unable to proc?

I was initially performing my tests on a level 60 dummy in order to avoid misses since I was just interested in the hits. When I turned to the boss dummy for the heck of it, I found out my proc rate to be drastically reduced in the range of 1 PPM. The sample for that test was very low (11 procs out of 300), but I was curious to know if anyone was aware of such an effect before wasting another 3-4 hours. Thank you.