nonmagical

Interesting starting rotation? You suggest quickly building a 5 combo point rupture prior to starting up SnD? My own judgement is telling me that a 3 or so point SnD quickly at the start would be a smarter way to go... But I guess not?

 

• Aldriana

Please note that I said "steady-state cycle". The question of how to bootstrap it is quite separate from what you want to be doing once you get going.

 

Fae

Hello Aldriana. First of all - thanks for the new features in beta 3 version of the spreadsheet.

But one thing makes me wonder. With my gear (armory should be up to date) it shows me that I should use "low rupture" cycle. Not sure what causes it, probably that I do not have the BiS gear from hardmodes and such.

Anyway, funny thing that I noticed. When I try to switch my headpiece for T9.25 (iLvl 245) I see ~100 dps decrease. Thats right, because of breaking T8 4-set bonus I guess. But when I try the same with gloves/chestpiece/legpiece my dps actually increases by ~40. I do not understand why there is such a difference in behaviour. In all 4 slots T9 gear is listead like an upgrade with ~100EP value so they should be pretty much similar upgrade stat-wise.

 

nonmagical

Fae: You are over expertise cap already. Switching your helm is giving you an even bigger boost in expertise, something that is clearly wasted in your current gear. That is why you are seeing the disparities between swapping tier pieces.

 

Fae

Aha! I see it now. I should have check the stats more carefully.

Thank you.

And it's good to know that it's working correctly and thus I can freely break T8 set bonus. Yay *likes T9 look*.

EDIT: My last statement is nonsense, I did not notice that beta 3 has evis-only talents/glyphs on by default.

 

Tofuu

Given the eviscerate only cycle, is it not better to use 2/3 imp eviscerate and 3/3 ruthlessness? The default gear setup of 1.3 beta yielded a 4 dps increase. Does this tiny dps increase happen for every eviscerate only gear sets? Intuitively, the more crit you have, the more finisher you will use in a given fight and this in turns translates to a higher value in ruthlessness as it will generate more CP per fight.

 

• Aldriana

2/3 Imp Evis and 3/3 Ruthlessness appears to be marginally higher spreadsheet DPS in most circumstances - always a bit ahead, never by more than about 5 DPS. However, from a practical perspective, any fight with target switching or requiring a short-term burn on a given target (and there are several of that sort in ToC) is going to favor larger Eviscerates over extra combo points, so from a practical perspective I think going 3/3 Imp Evis 2/3 Ruthlessness is a very strong option.

 

Bryt

I think I found a bug. If I change nothing to the default spreadsheet available, except change the GoSS for GoSnD I get #VALUE errors.

EDIT 1: After further inspection, I think I am breaking a cycle calculation instead of a bug in the spreadsheet. The default spreadsheet with a eviscerate only cycle does not like GoSS removed. With 4/5 T8 gear and a high rupture cycle, removing GoR also causes a #VALUE error.

EDIT 2: Ah sorry, I assumed this was fixed.

 

Tinwhisker

Aldriana has already addressed this and is presumably working on a fix. The point being that it was never anticipated that anyone would not use the GoSS; as such, any calculations attempted without it are likely to produce errors or be wrong. Realistically, there's no reason to think that GoSS would not be a "top 3" glyph for combat so I'm not sure it really matters.

You're a lot of DPS, you know that? You wanna' earn 14 achievement points the hard way?

 

Wytryszek

If the spreadsheet calculations are wrong, how are we going to know how good the glyph really is?
Anyhow, I am using the Tricks glyph, and I was just about to replace the SS glyph, because the spreadsheet says it is weaker than KS and Rupture ones. I do not get any errors when I remove SS, but now you got me worried if the results are accurate.

 

Tinwhisker

The error was first noted here in the beta2 sheet, I had also assumed that this had been corrected in the beta3 but since you got the same error it would appear not or at least it has appeared in another portion of the cycle calculation.

Until we get some clarification from Aldriana, it's probably safe to assume that all calculations with GoSS are correct and calculations without it are unmodeled due to the previously stated reason that such low CP generation cannot support the quantity/quality of finishers the spreadsheet is trying to model.

Edit

After I saw your edit I think it's safe to say that Aldriana did correct whatever error was occurring but Evisc only without GoSS and High Rupture without GoR are very likely to be outside what anyone would reasonably want to model.

You're a lot of DPS, you know that? You wanna' earn 14 achievement points the hard way?

 

• Aldriana

The specific problem is that when you do Glyph of SnD instead of Glyph of SS in an evis-only based setup, it breaks the high-rupture calculations as it can't cope with CP generation that low. The eviscerate damage calculations are still just fine, just masked by errors in a cycle that isn't even being used. I'm working on a fix. However, rest assured that Glyph of SS is worth 150-200 DPS in most cases, and thus a critical part of any evis-only cycle. If you're going to drop any glyph in Evis-only, drop Glyph of Eviscerate - it's by far the least powerful.

 

• Aldriana

TL;DR version: Mongoose and Berserking are both 1 PPM enchants; however, contrary to our current theories, both appear to benefit from haste, yielding significantly higher uptime for both than is currently modeled.

Long Version: I'd like to take a few minutes to return to a previous discussion, namely, the relative merits of the various weapon enchants.

Now, it became apparent some time ago that at end-ToC type gear levels, Mongoose was rapidly catching up to Berserking in value. Which is not to say that Berserking wasn't doing more damage, but it's lead was down to single-digits of DPS, and an argument could readily be made that the advantages of having a proc that increases your defensive stats rather than lowering them is worth a tenth of a percent DPS loss - particularly on a fight like Faction Champions. However, this notion is predicated on the assumption that Mongoose and Berserking have the same proc rate. If they don't, this analysis breaks down. As such, it became important to verify the proc rates of Mongoose and Berserking. So Latito and I set about doing so. There's a lot of backstory and the like I could at this point be provided, but the important bit - at least for the moment - is these two data sets:

2.7 speed weapon, 500 haste rating, no haste procs, 3/3 Lightning Reflexes:
12265 swings, 544 Berserking Procs - 4.44% proc rate

2.7 speed weapon. 32 haste rating, no haste procs, 0/3 Lightning Reflexes:
7572, 320 Berserking Procs - 4.23% proc rate

So, under the standard assumptions governing Berserking - that it's a PPM enchant based on hasted weapon speed (that is, haste increases your number of attacks but proportionately lowers the proc rate so your uptime remains about the same) - we can estimate the PPM based on each test. The first gives an estimated PPM of 1.25, with a 95% confidence interval of 1.15 to 1.35; the second gives an estimated PPM of .95, with a 95% confidence interval of .85 to 1.05. Note that these ranges are completely disjoint - if our assumption is correct, at least one of our data sets is a full 3 standard deviations from the real value. And this seems tremendously improbable. We thus apply a Fisher Exact Test to these data sets, and with a 2-tailed p-vaue of .496, we conclude that there is no evidence of any difference in proc rate between the two data sets.

For those of you that aren't math/statistics savvy, let me put that in simpler terms: a difference of roughly 25% haste resulted in no measurable change in proc rate. Thus, we can only assume that haste does not reduce the proc chance of Berserking; thus, increasing one's haste similarly increases the uptime of Berserking.

You may be wondering why I'm making a big deal of this. The short version is that it's a big deal - it means that the expected uptime of Berserking will be significantly higher than previously believed - between SnD, WF, and other sources, we tend to have a *lot* of haste rating, so the uptime of the enchant will be markedly higher. And in particular it means that Accuracy is no longer even remotely viable as an OH enchant.

None of this, of course, helps with the original question of Mongoose vs Berserking. Well, combining these data sets and a third, shorter one, I estimate the proc rate of Berserking as .98 PPM, with a 95% confidence interval of .92 to 1.04. From this, I feel reasonably comfortable saying that Berserking is 1 PPM.

We have taken a fair amount of data on Mongoose as well; however, it's not as directly conclusive:

2.6 speed weapon, 222 haste, 3/3 Lightning Reflexes:
13371 attacks, 570 procs with Mongoose not procced
3950 attacks, 156 procs during Mongoose Uptime.

2.6 speed weapon, 0 haste, 0/3 Lightning Reflexes:
11901 attacks, 541 procs with Mongoose no procced
2997 attacks, 132 procs during Mongoose Uptime

Now, applying the same analysis we did for Berserking in inconclusive; if we assume that Mongoose does not benefit from haste and use only the non-procced swings for both data sets, we get a 95% confidence interval for the first data set of 1.06 to 1.25, and .96 to 1.14 for the second; since these ranges are not disjoint, we can't 100% swear that Mongoose isn't, for instance, a 1.1 PPM enchant that does benefit from haste; and the Fisher test is similarly somewhat inconclusive.

However, if we use the fact that haste should not affect buff *uptime*, we can use the Fisher Exact Test to see if there's any difference in uptime between the two sets. And doing so gives a 2-tailed p-value of 4.74 * 10^-9 - that is, these two data sets have decidedly different uptimes, which is strong evidence that mongoose does, in fact, benefit from haste. We thus combine all the data together, and estimate the proc rate of Mongoose as 1 PPM, with a 95% confidence interval of .95 PPM to 1.05 PPM. Hence, Mongoose appears to benefit from haste and be 1 PPM, same as Berserking.

Hence, both have the same proc rate, but that uptime is significantly higher than previously believed, and Mongoose gets even an extra bit of advantage as the haste given by the proc makes it more likely to refresh. This appears to have very little effect on their relative DPS at high gear levels - the best double berserking setup I can come up with does 9679.4 DPS, while Mongoose tops out at 9669.6. However, for those 10 DPS, you get +defense procs rather than -defense procs, and use Wound Poison instead of Deadly in the OH, both of which can be seen as significant advantages. As such, I do think we're going to reach a gear level sometime in late ToC where Mongoose takes over as the enchant of choice.

 

Valustria

I presume all of this will eventually be accounted for in your next full release of the combat spreadsheet?

I actually jumped over here now to grab the 1.3 version (since I'd taken a break from the game for two weeks), and saw this post, which makes me wonder if I should hold off for now. ^_^;;

"In emptiness there is good, but no evil. Wisdom exists, logic exists, the way exists, and the mind is empty."
~ Miyamoto Musashi, The Scroll of Emptiness, The Book of Five Rings

 

• Aldriana

Beta 4. Minor graphical options, a few new options on the settings page, assorted bugfixes, and the Mongoose/Berserking change noted above. I also updated recommendations to allow it to recommend both the normal and heroic version of an item, as that appears to be, for the moment, possible.

This is getting close to ready for official (non-beta) release, so if you have any other adjustments you'd like to see before it's made official, now's the time to suggest them.

 

Istaril

Out of curiosity, how much of an impact on haste EP do the mongoose/berserking finds have?

Two beta spreadsheet questions; first what exactly does "Use Agi Value for ArPen?" do (and why?). Secondly, when I switched to the settings sheet and back to the equipment sheet, I appear to have lost all non-greyscale colours, all drop-down menus (item slots are blank, enchants/gems are still listed). This appears to be cosmetic, as it's still displaying the same DPS it was... and it doesn't appear to be something I did (undoing everything prior to switching sheets didn't reverse it). I wasn't able to re-create this a second time re-downloading the sheet, but that first sheet remains messed up even after closing/restarting.

 

Tinwhisker

This locks the EP value for ArP to be the same as the one calculated for Agility. Toggle it on and off while looking at the EP values on the equipment sheet to see what I mean. The reason it does this is more about convenience when looking at gear choices and upgrades, most notably when you're at either the soft or hard ArP cap.

You're a lot of DPS, you know that? You wanna' earn 14 achievement points the hard way?

 

Valustria

I'm gonna grab this one, but quick question (or suggestion, if it's not yet implemented):

Is there an option to switch between Alliance and Horde versions of the item names? Or, more intelligently, does the sheet do this based on your race? If not, could it in the final release?

"In emptiness there is good, but no evil. Wisdom exists, logic exists, the way exists, and the mind is empty."
~ Miyamoto Musashi, The Scroll of Emptiness, The Book of Five Rings

 

Gekautes

Just a minor bug for people who don't deal with all the calculations in the sheet: Bracers of Dark Determination got 50 Haste instead of 50 Hit in the last beta release of the sheet.

 

• Feist-Mok

I would wager that somebody will take the time to do this/upload it with a released version of the sheet - it's kind of pointless to do when a new release is being dropped every week or so.

 

Driphter

If you're using excel you can use my modified version of Dinesh's Combat Sheet Companion to toggle between Horde and Alliiance item names. It's posted here: Combat Sheet Companion

 

Jaron

Would it be possible to add the expertise bonus for Humans, Orcs, and Dwarves? I know it is possible, I keep doing so on my sheet every new release... just wondering if this is worth your time to get it added into the core of the sheet.

Also, can you add Treads of Dismal Fortune? These boots are, in almost all cases, better than Treads of the Invader which are still in the sheet. Thanks.

 

Loot

Guys, think it is possible to model the dps loss of the bug that precision is not working on axes using the 1.2/1.3 spreadsheet? I got the [Lionhead Slasher] heroic and would love to see if it will still be an upgrade considering precision will work on MH only.

 

Daarky

There would be no point, this bug was already hotfixed on live.

 

• Aldriana

Precision on axes is fixed, as of about 2 days after it was discovered.

Re: Expertise racials. Um, it'd be a fair amount of work, and it's not at all clear to me how much difference it actually makes. So no, I have no immediate plans to do so. I mean, you say you've added it, but how do you account for the fact that the spreadsheet doesn't currently distinguish between axes and swords? That it doesn't track expertise for the hands separately, so if you're using a sword in one hand and an axe in the other, your hands are not being dodged at the same rate? And so on.

As a side note, for those wondering about the "Allow DP" option: an argument can be made that for short-lived targets and target switching, wound poison has advantages over Deadly in terms of getting a burst of damage that requires no stacking. Hence, it is reasonable to optimize your gear around using double wound rather than optimizing for wound/deadly. So I've provided an option for performing that optimization. "Allow DP" disabled, you'll use double wound (or double instant, depending on which is more damage), allowing you to optimize your gear around that poison option. Doing so generally boosts the value of crit, hit, expertise, haste, and mongoose by a significant amount.