It does, but it's pretty hard to know for each swing whether or not there was an instant attack as well in the last three swings.

Ok, version 0.42 released. The major change is to the flurry calcs. I think this represents it much better. Here's the notes and link:

Added DPS field with no execute
Added Sentinel's Plate Legguards
Added -crit from high mob level to hit mods, was adding +skill to +hit twice
Changed flurry calcs; instead of the swings per yellow attack, I take the average amount of yellow swing per 3 white swings
Fixed haste enchants, added separate field for item haste
Added more consumables and buffs, cleaned up the buff section

http://www.savefile.com/files/7924030

Cooldown trinkets are at there best in short fights or fights with length = A(X-1) +B where A=cooldown time and B = effect length and X = number of times it can be used.

STR does not directly effect execute of course, but it does increase your rage gen so you can do more executes more consistantly with more rage. 2* 1.2 * (150AP/8.6/30) = 1.395 rage per second with recklessness + deathwish (bit less with ac of course). Someone needs to write an addon to show how fast you really spam execute. It certainly isn't every 1.5s due to miss/glancing blow/dodge streaks that prevent consistantly generating 10 rage in 1.5s. Execute is especially sensitive to std dev on rage gen due to it getting 60 damage per rage for the first 10 rage and 15 after that, but the same effect can be seen with very slow weapons dual wield.

Also there is 1%haste on helm and legs which (I believe) at high gear levels is the best dps option.

As I mentioned, you can get around quite a bit of messy math if you use an in game addon that ties into a damage parser.

Just saw you added a new version, will take a look...

Looking at your execute dps it seems off.
DPR for execute should be
IF rageper1.5s < execute rage cost, then 600 * yellow mod/execute rage cost, else (600 + 15*(rageper1.5s-execute rage cost) * yellow mod / rageper1.5s
aka.

[top]IF(B48 * 1.5 < 15-2.5*R32, 600*B43/(15-2.5*R32), (600+15*(B48*1.5-15+2.5*R32))*B43/(B48 * 1.5)
DPS for execute should be
DPR * RagePS
or

B56 * B48
RPS should be RagePS
or

[top]B48

This brings DPR for my character to 54.44, DPS to 546 and makes execute obviously the best for DPS, and 2nd to overpower for DPR (as expected)

Edit: minor issue with one part of the if statement
Edit2: for the uber default gear, execute DPR is 46.6 and DPS is 707 which makes sense. More RPS

less DPR for execute but more DPS.

Celandro - "As I mentioned, you can get around quite a bit of messy math if you use an in game addon that ties into a damage parser."

Links to any of those addons?

I'll have to update the DPS and RPS cells for execute. I had originally used those back when I had an excruciatingly number of IF statements as my DPS calcs. For now, ignore those. The DPR is worst case scenario(full rage). If you look in my final DPS cell, I don't use any of execute's cells in that equation. Primarily because you have to account for the full bar execute you get when the boss hits 20%, then you can start spamming it.

Otherwise, I believe my algorithm matches your's.

Perhaps add a row for full bar execute and use the current one for execute spam and then use references to those cells in your dps calcs to clean up the total dps calc?

Yojimboo: The addon Im talking about is one in my head I have yet to write. I don't really have enough time to write it myself but it should be fairly straight forward, especially if you steal the excellent tankpoints UI as a starting point and take the calcs from this excel spreadsheet. Basically it would have the damge per attack/dps/dpr calcs from the excel sheet as well as a method of adding/subtracting stats that would calculate out what the change in damage is. If the optional dependency to a damage parser was available, the actual % distributions of damage per attack would be used to give a predicted real dps. It would initially cover fury warriors but would be expanded to arms warriors and rogues next. If done well I fully believe it would be more popular than tankpoints.

Here is how you really model flurry uptime and why 1-(1-c)^x is only a gross over simplification of what statistics really are.

Let's say that we want to calculate uptime for a fight that will last 10 white swings. If our crit % is 30% that mean that over a long enough period, any 10 swings portion of a fight will have 3 crits, leading to 3 flurry. Sometime less then 3, sometime more then 3, but that's why we take the average.

The best case scneario is that we crit on the first hit, getting flurry for hit 2, 3, 4, then criting again on hit 4, flurry up for 5, 6, 7, crit on 7, flurry up on 8, 9, 10 and our 10 hits are done. Uptime of 100%. That's with 30% crit and this got a very low % of happening. Our worse case scenario is not criting before the 8th hit, then criting on the 9th and 10th hit, flurry uptime of 20% since the last hit is "out of bound". This also got a low chance of happening.

You have to calculate the probability of each and all permutations of 3 crits in a 10 hit period and add the uptime for each value. You will probably need to change the model if you want an "unlimited" fight somehow at the end of the serie. I'm not saying that the 1-(1-c)^x formula cannot be a decent enough estimate but if you want a model that really take into account overlap and give you the real number, you will have to program something like this above in mathlab and add the hundreds of numbers this will give you.

Statistics are alot more complex then what people generally give them credit for.

Ignoring the first two swings and instants...

Are you trying to claim that the chance that a single swing is flurried is not 1-(1-c)^3?

Or are you trying to claim that if each swing has a 50% chance of being flurried, flurry will not average 50% uptime?

Including instants does make it more complicated, though.

Thorb, that equation DOES t0ake into account overlap. I think we can at least agree that (1 - C)^3 represents any possible sort of 3 swings without a crit. Thus, subtracting that from the whole collection of swings(which is represented by 1), gives you the said permutations for possible combination of crits.

Thorb, your problem is you're trying to model a 10 hit period, not a fight. This isn't about a 10 hit period. This is about sustain. You're trying to model edge effects, when one of the base assumptions we're making is that edge effects are going to be ignored.

Flurry uptime is essentially modelable as (again, ignoring edge effects):

1 - ((1-c)^3 +n*c)

Where n is the average number of instant attacks in the time occupied by three swings. The only hard part is getting a good value for n. :)

Kalman, I would think the equation would look like:

1 - (1 - C)^(3 + n) where n is the number of instant attack during 3 melee swings.

So I can't seem to find a way to be able to not input a 2h weapon and still have it able to calculate the dps, but due to my 2h being the OEB it throws off the total +skill calculation. Any suggestions? (I'm using OpenOffice, it might be something wrong with that)

Select something like axes for the 2h spot with no attributes.

Yeah, what Yojimboo said. I believe I put that in the notes section of the spreadsheet.

As for +skill, I assume you're a human? No, that won't mess up the skill calcs, I thought ahead for once. Your +skill will be +13 for calculations like +crit and +hit, but in regards to glancing blows, I cap it. You can see that near the bottom with "effective weapon skill". Put human and OEB on, with a level 63 enemy it should still say 10 down there.

No I'm a Tauren. When it loads up with the Axes thing in 2h it shows Err:503 in the crusader STR boxes, the white cells in collumns N, O and P, and for about 90% of the result boxes.

The OEB thing is a non issue I only put the weapon in there because it made the Err:503's go away. Once again this is with OpenOffice (version 1.1.4 here home, much newer one on my laptop (the laptop one supports dropdowns but gets the same errors).

Hmmm, it might be an OpenOffice thing, never used that. Are you setting the DW slots to Axe, or something similar, and their enchants to none?

I'm guessing it's OpenOffice then. Just double-checking, you set the DW slots to Axe, and their enchants to None, and you still get an error?

BTW, for anyone, I don't see how Kiss of the Spider is that great. DFT and Slayer's Crest beat it out even with Naxx gear, and HoJ still gives competition. Did I get the cooldown right? 2 minutes?

Hi there, long time lurker. Great model Deathwing i've been using it since .31, however i've noticed some wierd numbers with .42. I get more projected DPS with CoTBD in MH than with Crul, all other variables stay constant. That seems wrong to me or am i just being dim? I've tried it with the default setup and with my own gear.. Surely that can't be right?

OK, diffference between the two, from Crul's point of view:

+6.5 White DPS
+10 AP(less with Kings)
+6 STA
-1 Crit
-3 STR from crusader proc


I'm sure you know that listed weapon dps is not that important to a fury warrior, stats are more important. Plus, the 6.5 DPS you gain on one hand is only applied to one weapon, while the +1 crit you get from the claw is applied to both hands. I definitely see CoTBD being more DPS than EoC.

I would look for CoTBD and CTS coming out of BWL for max DPS. EoC offhand is a bit less dps for a bit more sta.

I set the two DW slots to AXe / none and the 2h slot to Arcanite Reaper / None. Still getting some errors. Only way I can get rid of the errors is to fill in both the DW slots and the 2h slot. However for whatever odd reason my laptop seems okay with the crusader enchants. This is just downright confusing. :?: I'm gonna go see if I can pickup another $5 copy of Office.

This is likely due to slower weapons in the MH getting a higher bonus flurry dps + whirlwind damage.


One thing not moddelled in this excel sheet is the bonus that paired speed weapons get from flurry. Basically any time both weapons would swing at the same time, you can get an extra free flurry swing. This will happen rarely for unpaired weapons but dual 2.8 speed or 2.6 or 1.6 speed weapons (looking at early naxx) will get the extra flurry swing 50% of the time. This can be modelled by changing to mod to 3 +ChanceOfPairedSwing/2 + ExtraAttacksPer(3+chanceofpairedswing)Swings.


PS. Juju power is not horde only anymore.
PSS. Crit cap is NOT modelled.
Max white damage crit rate for level 63 mobs from the back is is 29.8 + NetHit + ModSkill*.04 (due to reduced dodge)
I easily got to 44.961 crit with buffs and only +9.28 hit +7skill
Max white damage crit rate in this setup is 29.8 +9.28 + .28 = 39.36
Max yellow damage crit rate from the back is much higher (due to no glancing blows and no misses) 100-5.6 +modskill*.04

Easiest way to test this out in the spreadsheet is by using crit stones in the 3 to 14 range.

Interesting note about DW flurry Cel, that's pretty cool.

Btw checking the sheet atm and the legs of qirag general changed stats, they are a bit better now.

http://www.thottbot.com/?i=52769

2 crit, 20 str instead of 35 str.

Would be nice to have a way to specify if the effective weapon skills apply or something. Maybe add 1-2 columns specific to race? Right now I'm an orc and it applied the +5 to sulfuras by default.

Crusader model seem wrong imo. I remove Windfury and the 61 strenght doesn't change. 61% uptime seem kinda high too especially if it doesn't take into account WF.

The enchant bonus are calculated even if you don't have a weapon, which lead me to some confusion, maybe add a check for that.

Another weak thing like Celandro said, the formulas are so complicated that it's not really possible to confirm them it seem. I tried checking if heroic strike would correctly reduce your rage generation by 1 white hit + the rage cost of HS. It also seem most of the numbers for each type of attacks like slam and heroic strikes aren't really used in the total?

Edit:

I'm playing a bit around with crit and hits atm, and using a 2-h, when I go from 0 to hit to 1 to hit the yellow hit mod doesn't change and always stay constant at 0.9416. I don't think this is normal. It might explain how weak +hit is with a 2-h in your model currently.

Net AC seem to count dodge instead of AC.

Small detail but flurry is 10% for the first lvl and 5% for each one after that, not 6%.

I'm also not too sure how you model flurry but with base stats of 1k atk/24.4 crit with no buffs at all I get to 93% uptime, that's too high. For comparaison using similar calculations I get to less then 80% on my side. The fact it's not affected by WF also doesn't help.

Elemental sharpening stones are 2% per stones, not 4%. I think you assume DW and 2 stones there.