pf

If you have combat potency, the next-stat sheet will almost always come to the conclusion that abacus is better. If you dont have CP, it depends on how buffed you are. Abacus obviously favors buffs better since it scales with all other stats where BB only scales with hit/crit. Anyway, the next-stat data is quite accurate except for those times that are mentioned in the known issues like not counting extra mongoose procs gained from +hit, or cycle changes gained when gaining crit with mutilate. Of course, if the next-stat sheet shows something like 1 hit = 30aep, then something is obviously wrong (this happened a bit in the past mostly problems with weapon/poison procs, but I eliminated all known issues a while back).

The problem is not that netherblade bonus decreases DPS, the problem lies with how the sheet interprets slack energy. The reason why its showing the set bonus to lower dps is that it lengthens your snd uptime (ie increasing time between ruptures) while not increasing your snd % uptime (since both are 100%). To get the best results, one needs to force the SnD uptime for the required cycle to have almost 0 slack energy (or base surplus as the sheet refers to it). Another way of putting this problem is that the sheet assumes you waste combo points instead of time left on SnD.

Anyway, after playing around with this for a few minutes, I am finding that 1s/5r is the optimal cycle if you have 2p NB bonus. More rupture damage is outweighs the lost instant (SS damage). However, I also find that it is not worth it to do less than a 5point rupture. Anyway, a 1s/5r cycle takes about 17 seconds to perform. 5 point rupture lasts 16 seconds. However, despite all of this, there is a chance that you will have crappy cpot/relentless/ruthless procs and not have the required energy to perform all of this and keeping snd up 100% of the time. 2s/5r is probably a safer cycle to upkeep and unless you have a mangle up, there is very little dps difference. In fact, without mangle, there is only a couple dps difference between 5s/5r and 1s/5r.

 

• Aldriana

There's two reasons I question the next-stat sheet.

1) When I hand-calculate the values myself, I get different numbers - in particular, I find that haste rating and crit rating are worth more than the spreadsheet indicates. Of course, I could have made a mistake myself, so this isn't conclusive.

2) The values in the spreadsheet seem inconsistant. For instance, lets take a look at the aforementioned Abacus/Bloodlust Brooch comparison. With the gear I have punched into the spreadsheet at the moment (which is a projection of the gear I'd like to have by sometime in late SSC) and Abacus as one of my trinkets, I get a total buffed dps number of 1549.14, and Haste Rating is listed as being worth 2.06 AP on the buffed dps spreadsheet. Hence, the total value of the trinket should be 21.67 * 2.06 + 64 = 108.6 AP. So if I replace it with a trinket with 109 AP, my DPS should go up, right? So I modify Bloodlust Brooch to be 109 and stick it in, and my damage goes *down* to 1547.62. Fiddling with the numbers a bit, we find that the "real" value of Abacus is between 113 AP (1549.03 dps) and 114 AP (1549.39 dps)

Now, a partial answer to this would be the nonlinearity of the value of stats. But with the 114 AP trinket equipped, the incremental value of Haste is listed at 2.11 AP, which would list Abacus at 109.7 AP-equivalent - hence the "true" value would be expected to be between the above 108.6 and 109.7, rather than ~113.3 as in seems to be. Based on this analysis, the next-stat value for haste rating ought to be more in the ballpark of 2.25-2.3 rather than 2.06-2.11. Hence it seems to me that the next-stat value of haste is about 10% lower than it should be.

 

pf

My guess is the difference you are seeing is the added benefit of more mongoose uptime. Also, deadly poison does not work correctly with next-stat now either after the change I made. There is something that is very slightly off in combat potency/nextstat, but I could never quite nail it down, but the type of numbers you are giving me suggest that you are using mongoose in those numbers. So see how much closer it is (and if you like, figure out the minor problem with cpot / nextstat)

There are 2 solutions to this, 1 is to calculate mongoose for each +stat calculation and then determine the change. This is the solution I used pre-TBC with crusader, but since crusader was a linear equation, it was a much smaller formula. The 2nd solution would be to create a macro that would iterate through all the next-stat values using the cycle sheet as well. This is a better solution, however I do not think you could retain OpenOffice compatibility with this solution. I never wanted to break OO capability, as many users of the sheet do not have excel.

 

• Aldriana

Closer, still off. I changed the mongooses to 15 agi and the deadly poisons to instant, next-stat is still valuing Abacus at between 109 and 110, while looking at actual dps puts it between 111 and 112

 

pf

If your cycle is complete, then I would guess that the problem lies somewhere in my Cpot conversion on next-stat(as I said, its slightly off). I'm looking at it now and the numbers arent adding up for me.

On the next-stat sheet, I think "uohhit/ohspd*3*0.2*cpot*(1+$R$4)*($Q$26-1)" should be exactly the same as "('Buffed Cycles'!K25-'Unbuffed Cycles'!K25)/'Buffed Cycles'!M25" assuming you are forcing a 5s/5e cycle by disabling rupture. Both of these numbers should be the energy/second gain with +1 haste rating. I am showing them to be .000031 different which is causing the haste rating gain to be slightly off.

Also, when searching for next-stat problems, its best to disable all buffs and add +1 stat to your buffed dps so you can check the values. I'm showing +1 haste rating to be about .0001 dps off, which is not enough to be 2-3 ap difference on abacus as you are showing. Without looking at your sheet, I can only assume what is happening, and at this point, it could be haste rating, but as I said, without seeing the inputs on your sheet, I'm at a loss to explain it.

Edit: changed paragraph order to have more flow in the post, but there is still very little anyway.

 

Crumpet

Possible Bug:

With a combat swords/fists build, the first point you put in Ruthlessness nets you roughly 4 dps (ub and buffed), and the second point another 4, but when you put the 3rd point in your dps doesn't increase at all.

However using a combat daggers build the 3rd point still has a big impact on the dps.

Why does putting a 3rd point in Ruthlessness with combat swords/fists not change your dps output at all? Unbuffed or Buffed.

 

Zoro

I noticed that 1s/5r worked out best for 2pc NB as well.
As I've mentioned before, Blizz are only going in one way regards to energy management and that's cycle compression. Where do you go after 1s/5r? 1s/5r/1e?
This 'slack energy' issue is only going to get worse as time goes on. The essential question is at what point in an energy cycle is it most dps efficient to hit your finisher. The obvious answer, for rupture, is every 16s provided S&D uptime is maintained and the energy in/out is stable. Previously constraints meant you had to shift to 5s/5r or 4s/5r before you maintain it, leaving 10-15s without rupture ticking.
Working all that out is a bit more complex than the current absolute system that merely calculates total energy in/out over the S&D duration. You'd need to arrange something like a energy flow over time comparison with variable finisher locations. That'll be fun.

Regarding next stat, if you're not worried about bloat the Open Office solution is just to make a seperate hidden dps/cycle sheet for each stat+1 for both unbuffed/buffed and compare with the original. That's an extra 12 sheets that are partially redundant but it would seem the foolproof way of handling it.


Crumpet:- That's going to be an issue with your particular setup. On mine, that last point in Ruthlessness shifts me from a 4s/5r to a 3s/5r rotation. Granted, the sheets with Xs/5r aren't publically released yet and are just something those of us with Excel knowledge are messing with, but it does show the implementation is there and working.
I suspect in your case the extra point isn't effecting dps as it's not changing a rotation pattern and numbers are just working out for you that the percentage loss of combo point is exactly balanced by the percentage gain of a SS (can't actually see that happening, I'd have thought the two would be clearly different).

Just for giggles, I've uploaded a sheet with Xs/5r capability to:

http://www.savefile.com/files/653624

It's also got the latest itesm from the PTR added to it. Including Choker of Serrated Blades, with the - 154 armour included. In case people hadn't spotted that's a really powerful facet to have. Blizz seem to be having a lot of fun messing with items that effect armour, potentially there's a lot of dps gain there.

Only item I didnt add was the Ashtongue trinket - 20% chance per combo point on finisher to add 162 crit rating for 12s. Still trying to figure a way to get that working correctly.

 

Shime

Call me a noob or whatever, but i didn't manage to use that rogue dps sheet. I never used Excel before in anyway, and its quite confusing to me. After a hour of trying i just gave up.

What to do? How to use this?

 

Zoro

At the bottom of your Excel window you'll see a row of Tabs called...
'Intro''Talents''Gear_Buffs''Unbuffed DPS''Unbuffed Cycles''Buffed DPS''Buffed Cycles'

As a newb, all you're interested in is 'Talents' and 'Gear_Buffs'.
Click on the 'Talents' tab to take you to the talents page.

This sheet allows you to select what your talents currently are.
Broken into the 3 trees, you can select your talent spec. Next to teh talent you wish to change, under 'Points' is a box. Clicking here will show a grey box with a downward pointing arrow next to it. Click the box and a drag down menu will appear showing you the valid choices for points in that talent - select one.

That way, working through the talents you'll set your talent spec up.

Next, go to the Gear_Buffs page.
The first part is where you set what your current gear is. Just like before, you use menu boxes to select your gear. Clicking inthe box which has a weapon/armour name will show you the drag down menu grey box for that slot.

Once you've selected all your gear, the spreadsheet is displaying your predicted dps in the yellow fields at the top of either page.

Below the gear selection boxes are the buff selection tick boxes. Clicking on the white boxes to the left of a buff toggles it on/off. The 'buffed dps' fields show you what you dps is when you have the selected buffs active. 'Unbuffed' always returns the value for when you have no buffs active.

 

Cel

Little guide:

Start by making sure the Talent and Gear Buff spreadsheets are set to what you are currently. This is done via drop down boxes accessed by selecting the cell then clicking the arrow that appears. Also make sure you select buffs you have in a raid on a general basis in the lower portion of the Gear Buffs spreadsheet via the check boxes on the left.

From here, you use a few resources to see what different things are worth:

You can make changes to your gear to see how the dps is adjusted and what dps cycle would work out best for you. For example, you can see how Windfury measures up against having Grace of Air + Instant Poison based on your spec and gear, or you can see whether having a shaman in your group would benefit you more than a warrior. It's also nice to see how different talent specs benefit different gear choices. For instance when I equip my Gladiator's Shanker and Malchazeen, Mutilate is clearly the best choice for single target DPS against mobs that can be poisoned. This will help you make gear choices such as "Is Latro's a better choice for my Combat Swords offhand than a Spiteblade, because of the speed on the weapon even though it's significantly lower dps?"

How to use the cycle box
--The number in the box is how many combo points you should have when you use the skill associate with the letter, and the order of the number letter combination indicates what order you should use the skills in. Take a 3s/5s/5r cycle. s = slice and dice r = rupture. If this cycle is calculated to be your optimal DPS cycle, you get 3 combo points via your spam move (backstab, sinister strike, etc.) then use Slice and Dice, then you work up 5 combo points and use Slice and Dice again, finally gather another 5 combo points and use Rupture, then start the cycle over again.

You can use the AEP calculations on the Talent sheet to find out how much you'll gain from a given stat.
--The higher the number, the more the stat is worth, using Agility as the base stat holding at a value of 1. If the number is higher than 1, you will gain more than you would gain from Agility. If the number is lower, then, naturally, you would gain less. This is useful in comparing gear in that if you're trying to compare an item with 25 agility 10 hit and an item with 40 attack power and 15 crit, you can see which you will be gaining more from in long term dps. ie: for my spec and gear 25 agility = 25 AEP 10 hit = 8.7 AEP... add those together and you get 33.7 AEP for that piece of gear. 40 AP = 19.2 AEP and 15 crit = 11.55 AEP, bringing the overall item value to 30.75 AEP, showing that I would gain more from the first item than the second. Another very useful tool is the link to Sp00n's AEP site where it will take the AEP values the Spreadsheet calculates and inputs the values into his list of gear a rogue might consider wearing. The PHP script will sort the table of gear into what would benefit you most, according to this gear ranking convention.

There are more ways that you can use the information provided in the spreadsheet to help optimize your rogue, but here's the basics. If I misrepresented something, or got something wrong, please point it out. :]

"There is much pleasure to be gained from useless knowledge." - Bertrand Russell

Alpha is recruiting... go go.

 

Siq

I tried to change the 4s/5r cycle, so that it calculates the dps for a shortened cycle, as if you start your cycle again befor the snd time ran off.
The point were I am stuck is the inst/sec field (V17 in the cycle tabs).
I don't really understand the indirect command, so i couldn't really understand the calculation of this field.
I only know: with 2 nb bonus and my 4s/5r changes, this value increased quite a bit too much (from 0,33 to 0,39 which makes +80-100dps).

If somebody could describe how the inst/sec field is calculated I would be very grateful

*edit* think i understand the indirect command, gonna try to understand the rest too
*edit2* ok i think i managed to fix this problem, still something is wrong, dunno what, have to check this out later

 

Crumpet

Thanks a lot :>

 

Zoro

Right...
Inst/sec is telling you how many combo point attacks per second you are making. Obviously, if this goes above 1 your cycle is bust as you're breaking global cooldown. It's used elsewhere in the sheet for calculating your damage from spam moves and proc rates.

V17 is a bit complex to describe.
The first bit is IF(Mutilate) - basically the mutilate stuff is completely seperate form the normal cycle stuff - don't worry about it.

THe next is a set of gobbldygook with indirect commands referencing X14.
X14 is the cell that tells the sheet for reference purposes which row corresponds to the cycle that is your optimal.
All that INDIRECT nonsense is about pulling values out of the row number created in X14
So, spelling it out, it does the following...

(Take the value under column "i" of the row number X14)/(time for cycle of the row number X14) + If the value in the column "l" of the row number X14 >0 then (That number)/Energy cost of your chosen spam move/(time for cycle of the row number X14)

Make sense?
"i" is the column telling you the number of instant attacks inthe cycle, "l" is surplus energy for the cycle.

So...
(Number of instants in cycle)/(Time for cycle)+If surplus energy(Surplus energy)/(Energy cost of move)/(Time for cycle)

It's that last bit that's behind pf's statements that at present the sheet when faced with slack energy wastes combo points, not S&D time.


Hope that helps in understanding how the cycle sheet works, and good luck with trying to rejig it for shorter than S&D time calculations.
I've not yet devised a method for doing it to find optimisation that doesn't rely on circutous cell referencing.

 

Shime

thx a lot, think i am going to spend some time on this.

off topic:
I used to use Eviscerate instead of Rupture, cause the non existent crit chance for rupture seemed to drop its dps behind eviscerate. I am aware of the fact, that rupture ignores the armor of my target, but in long terms the given crit chance of eviscerate should out dps rupture, or?

Is there a Topic on this Issue?If yes, thanks for the link ^_^

 

Cel

There's more to it than just the damage done by rupture vs. eviscerate in and of the skills themselves. Consider also the facts that many guilds run feral druids, so you will get bleed damage bonuses and also that rupture costs 10 less energy to use.

"There is much pleasure to be gained from useless knowledge." - Bertrand Russell

Alpha is recruiting... go go.

 

◊ Kalman

Rereading that post, Zoroaster, I wonder how Pf ever understood what the hell I was thinking when I wrote that sheet.

 

Zoro

Well, the inputs/outputs from it are fairly clear. It's only when you delve deeper into exactly HOW it's doing certain things does it become more complex.

If I had the time, I think there's plenty of room to strip it right down and rebuild certain parts as I see a fair few places where redundancy has crept in. Not sure if that's an artifact of two authors or being built over two periods, or being a compilation or what.

Alas, I don't have the time. I've got a house move coming up, and it was enough for me to just look into reworking it for cycle times less than S&D time optimisation to tell me I wasn't going to have the time.

Part of that's the heavy use of INDIRECT in places where it's not strictly needed. That's fine if you're the author, but for subsequent users it makes modification really, really tricky as you're never entirely sure you've covered all the past references to a cell you just moved...
The major part however is it involves a complete rethink in the nature of how S&D time optimisation is calculated. You can't just get away with a series of set number of combo points, you have to try and calculate the correct time left on S&D before finisher, which is something much more fluid.

 

Cel

I would personally love to be able to learn the things required for a project such as this... any suggestions as to a good resource to pick up the necessary knowledge beyond poking around in cells and hoping things work?

"There is much pleasure to be gained from useless knowledge." - Bertrand Russell

Alpha is recruiting... go go.

 

◊ Kalman

On the spreadsheet side, that's pretty much what I do. Poke around, hit F1 and look for a function that'll do what I want, and pray.

 

roq

Wow this thread has taken a new direction all of a sudden, but i was wondering if someone could give me a hand.

We now pretty much believe the base miss rate on a boss is 25.5 thus leading us to believe that 308 with WE and Precision will cause us not to miss. Well right now the sheet seems to undervalue my hit since i am sitting at 303 and I am over the cap of 284, which the sheet shows as the cap. (these are the numbers with our precision).

So what do i have to do to adjust to sheet to make it work under these standards? I thought maybe changing WE would do it, but that didn't work. I tired changing the miss percentage on the unbuffed DPS sheet, but that didn't work. Since my MH miss was still zero when it should be something along the lines of .32. Am i missing something or is this working correctly? Because bosses should have a higher miss rate than we currently have set on the sheet.

 

Cel

Cell q2 on the unbuffed and buffed dps sheets has the miss rate = 24.6%.

"There is much pleasure to be gained from useless knowledge." - Bertrand Russell

Alpha is recruiting... go go.

 

◊ gwystyl

Things like this are how I learned to use Excel in Middle School.

^.^

If you're using a recent version of Excel (not sure about OO.o), clicking on the formula bar on the command (IF, AND, VLOOKUP, etc) will pop a little box up with the command name and its required syntax. For more detail, you can click on the command name in this dialogue box and you'll get a pop up window that will take you through the syntax, purpose and variable examples of how to use that particular command. It's actually a very good way to learn the ropes.

It'll take quite some time, but it's time well invested, IMO.

 

Cel

Yeah, I've done some of that sort of thing to generate common keys between lists of data for matching purposes... I suppose that is the way to go: Figure out what I want done before how to do it, then go from there.

I'm a recent transfer to Ysera, actually... so if you get a random tell from some rogue named Ferrit saying something to the effect of "WTF MY VLOOKUP HATES ME" then remember your post in the rogue dps thread. >_>

--

Back on topic, is there any data on the changes made to glancing blows from the PTR? I haven't had the time to get over there to mess around.

"There is much pleasure to be gained from useless knowledge." - Bertrand Russell

Alpha is recruiting... go go.

 

roq

Yes, i know that, i mentioned in my original post that i changed it and nothing happened. I guess i can ask this question. Where is skhit defined? It is used in cell c17.

 

◊ Kalman

One of the more entertaining things you can do:

Go to a blank, unnamed cell.

In the cell naming field, type in skhit.

Hit enter.

It should pop you over to the skhit cell, wherever it may be.