rokshan

Is there any problem known with simming totems?
I believe since I have reinstalled my pc i get errors when I try to sim with totems in the priority list.
Everything works well without the totems in the list.

 

Andunie

Is EnhSim currently able to understand the way armor penetration works? The reason I ask is: when I sim Grim Toll/Grim Toll (no Mjolnir yet, so this is how I simulate having Mjolnir/Grim Toll), I get higher DPS output than Grim Toll/Darkmoon Card -- which is not what I was expecting. In the real world, with boss debuffs up and the 100% armor penetration cap, there are situations where you would have both trinkets up simultaneously and a lot of the benefit would be wasted. So my question: is EnhSim compensating for this?

 

Levva

If you are using version 1.7.0.3 of EnhSim then yes its using the latest Armour Pen calcs, as described by Ghostcrawler and tested by the community in the Combat Stats at 80 thread here on EJ. Mjolnir should be in v1.7.0.3 are you using an old version of the sim?

Author of ShockAndAwe Enhancement Shaman max dps addon
Author of Rawr.Enhance an automated gear checking program that can generate config files for EnhSim.
Please use the EnhSim by Tukez, Sylvand & others to simulate what gear, priorities etc are the best dps.
FAQ: Hit cap 342 Draenei, 368 Horde, Expertise rating cap 140 with 3/3 Unleashed Rage. Cap those before worrying about other stats.

 

Quaglia

According to Verges:

Anyway I also found this in Cat DPS Guide for Dummies

but I cannot find any reference to the 665 proc. Anybody could help on this? Should be this item be updated again in the sim?

 

Malan

Yes the proc is definitely higher than 612, one of the warriors in my raid was talking to me about it last week. You can see the actual effect right here Mjolnir Runestone - Spell - World of Warcraft

Shitting up every single thread on EJ since '06

 

Andunie

Neat, thanks Levva. Grim toll/Mjolnir comes out 50 dps above grim toll/darkmoon card!

 

rastotem

I've got something I need your input on, with pyrite infuser\fury of the five flights I'm getting around 6650DPS with pyrite infuser\wrathstone I'm getting around 6550DPS, but when comparing the EP wrathstone comes up higher, 216EP for the crit+142.7 for the AP buff(856*0.1667,16.67%up time)=358.7EP while the fury of five flights is 16*20=320Ap=320EP, what exactly am I doing wrong here to get these strange results?

 

Kraakku

The fact that your simulated DPS drops when you equip Wrathstone seems a bit strange, especially since your crit EP is as high as 2. The only logical explanation I can come up with is that Wrathstone has a 2min cooldown, while wolves have a 3min cooldown, so in a 6min simulation you will be using your trinket and wolves only once simultaneously and after that their cooldowns will be out of sync, while your FotFF will always provide extra ap for your wolves.

 

Peterle

An option similar to the "sync bloodlust with trinkets"-option would be nice. It could be enough to extend the current option by also synchronizing Feral Spirit.

 

Levva

How are you coming up with the EP figures? Are you getting one set of EP figures with one set of gear equipped and calculating the EP for that, then swapping gear calculating new EP values then applying the figures to the second setup?

Why are you using EP values at all?? I hope its just for a sanity check on the sim results? How are you balancing the EP values, do you take durations and cooldowns into account?

It is my sincere hope that with jEnhSim we can banish the notion of thinking of items in terms of their "total EP" and instead think of them as "total dps contribution". To me it makes far far more sense if the sim can say to you Trinket A is worth X dps and Trinket B is worth Y dps. Rather than changing them in to EP values which are random depending on what you change your gear for.

You do know that ANY change of gear will change your EP values don't you? Sometimes not by much admittedly but sometimes by radical amounts. So trying to compare items using EP values is inefficient if you can just compare dps directly.

Author of ShockAndAwe Enhancement Shaman max dps addon
Author of Rawr.Enhance an automated gear checking program that can generate config files for EnhSim.
Please use the EnhSim by Tukez, Sylvand & others to simulate what gear, priorities etc are the best dps.
FAQ: Hit cap 342 Draenei, 368 Horde, Expertise rating cap 140 with 3/3 Unleashed Rage. Cap those before worrying about other stats.

 

Atren

I can think of few --

1. Sim is not handling one of those procs correct. Not likely as others have not reported such problem.

2. You forgot to remove/add Wraithstone crit part to your stats. Or you had typoed crit before or after changing it for Wraithstone.

3. Not sure how Fury handled so this point might be wrong, but maybe you added the 320 AP as well as the trinket meaning it has double 320 AP?

4. You did some typo due to which Wraithstone was discarded.

 

Malan

Since levaa mentioned jEnhSim a few times recently, and because we've been making decent progress lately, I thought I'd give a little more info on this project.

jEnhSim is a complete redesign from the ground up of EnhSim in java. Our primary aim is to provide a platform independent simulation that has the same functionality of the current simulator while improving on usability, flexibility and maintainability. We've done an initial round of data structure design and are working on that implementation right now and I expect that within a couple weeks we'll also get started on the main simulation engine.

Usability - WoW simulations to date have had their UI mostly thrown together as a bit of an afterthought. They were often confusing, had inputs taking multiple formats (ratings or percentages, not often clear which one was needed) and had no sanity/validity checks on the inputs. We're going to ensure that as many inputs as possible use the same format, that they are labeled in a meaningful manner, that the intent of each option is well communicated, and most importantly we're trying to design the UI so that it assists in the main use cases of the community. It should be easy and straight forward to open jEnhSim and figure out if Gear A or Gear B is better for your dps, not an excercise in adding and subtracting stats and conducting multiple runs.

Flexibility - One of the major gripes with the simulations is that they can't keep pace with the often rapid changes to classes, spells and items on the live and PTR servers. We are constructing jEnhSim to be as data-driven as possible. This means that things like spell coefficients, min/max damage ranges, trinket/aura/enchant effects, etc, will all be represented in an XML data file that can be modified at any time by the user community. This will include base assumptions for the sim for things like the Windfury Cooldown rate and any sort of constants or assumptions that are used in dps calculations. This approach will do 2 things. 1 - it allows the community to alter the behavior of the sim immediately without waiting for a developer to edit code. 2 - it allows those who are interested in improving the accuracy of the sim, or in testing assumptions about the game mechanics, to adjust and fine-tune those mechanics so that the sim matches in-game observations as much as possible.

Shitting up every single thread on EJ since '06

 

Kraakku

I agree that such an option would be nice, but then again wouldn't solve all simulation problems. The only way to really get an accurate result from enhsim would be adding an option that allows you to decide when to burn your cooldowns, so you could maximise your damage potential.

Here are three different scenarios that require different cooldown uses:
1) A 6min combat, in which you would use your trinket at 0.00, 3.00 and 5.00 to gain maximum uptime for both cooldowns.
2) A 5min match, in which you'd need to use your trinket on every cooldown and save your second wolves untill you get your trinket up a third time.
3) A 7min match. In a 7 minute timeframe it would be impossible to maximize the uptime of both wolves and Wrathstone, so you'll have to decide whether you'll maximise your spirit wolves output, thus missing one use from your ap trinket, or if you'll just blow your cooldowns when they are up.

Even tho scenario #3 might be rare, some kind of changes to the sim might be in place since simulating use trinkets is obviously inaccurate. These issues with use trinkets might become an even bigger question with the release of 3.2 and the new badge trinket, which currently has a "1024 ap for 20sec" -use in it.

 

Falcon213

As nice as that would be, EP values are still very useful for anyone who raids without DKP and doesn't want to sim every piece of gear before it drops. It's a quick way to check a dropped piece vs an equipped piece.

 

Levva

Agreed, but IF and ONLY IF you have calculated new EP values for your current gear before you start the raid. This is indeed what I do myself using EquivalencePoints to display a total value on the tooltip. It does at least give a quick and dirty guide to if an item might be better.

However the problem with EP values is they aren't static, the very act of removing one item and equipping another will change your EP values, often radically especially for stats like hit & expertise with their caps, but also with stats like haste that don't appear to be uniform in their effects.

If you try to use static EP values from websites (even our very own TTT) you are more than likely not optimising your gear properly. This is why the TTT contains a health warning about using the values and a plea to sim it for your own gear levels.

Author of ShockAndAwe Enhancement Shaman max dps addon
Author of Rawr.Enhance an automated gear checking program that can generate config files for EnhSim.
Please use the EnhSim by Tukez, Sylvand & others to simulate what gear, priorities etc are the best dps.
FAQ: Hit cap 342 Draenei, 368 Horde, Expertise rating cap 140 with 3/3 Unleashed Rage. Cap those before worrying about other stats.

 

Kraakku

I was just about to type the exact same thing as you Levva. I myself sim my own EP values after every single new item, so that I'm able to get a fast estimate of an item when it drops, but this only works if you sim regularly and with high enough EP ranges. Trinkets however are a bit tricky to estimate compared to other items, especially if the proc/use is something else than ap, since huge leaps in stats can also produce dramatic changes in EP values, making the procc more or less valuable than it seems at first glance.

 

Malan

Hey now, I removed any EP values from the article completely back when 3.0 hit.

Shitting up every single thread on EJ since '06

 

• Raut

Care to explain this in more detail? Showing items X and Y as A and B "EP" shouldn't be any different from C and D "DPS". EP translates into DPS after all. Swapping items would ruin your fundament for reaching that figure, it being EP or DPS. Or are you trying to incorporate some subtraction of items into the equation to get this "DPS" number?

I'm all for you guys rebuilding EnhSim in Java to make it more portable and what Malan wrote about making fixed variables exist outside of the binary for easy editing sound good. However, I don't think swapping EP for DPS would be a good choice if you're not planning to subtract the item you are replacing in favor of this new item and running the sim again with the new stats. If nothing else, EP is abstract enough to allow us to give it certain rules, like "EP is highly situational and will vary when you change items or change ability priority". If you say item X has A DPS and item Y has B DPS where B > A, and swapping the two results in a DPS loss, I'm sure the jEnhSim thread will be consumed in flames. (Imagine the calculation where your findings gave hit and exp low to zero "DPS" because caps were reached. You swap for a crit trinket and DPS goes down.)

I may be missing some vital part of the puzzle here and if so please enlighten me.

(Hm. Is calculating every piece of gear to find a BiS list an NP complete problem?)

Baby, you can hold my balls.

13:17 < Kalroth> gays on men tv? I love that channel

 

Synkronos

Every now and then, I consider running sims for my gear, removing one piece for each sim. This would give EP values for each slot without the item equipped, and hence presumably more accurate weightings. It's rather a massive undertaking, what with continual alt-tabbing and typing or exporting a billion times tho, so I never do.

The sim doesn't build up your stats from actual items and gems and enchants tho, it relies on being fed those values pre-calculated by WoW, so in its current format, it could never do the kind of 'item subtraction' you're talking about. It would be hella neat, but would also make the sim run 17 times longer, and produce 17 different sets of EP weights. That would be far too much information to be useful, unless the sim were to give gear upgrade lists instead of weightings.

Even with either of the methods described about, you still won't be accounting for items that 'cross' hard or soft caps - 40 hit rating when you need 20 to soft cap, or 25 expertise when you're 20 away from hard cap. The only way to get truly accurate results is to sim a full gear set before and after any items are changed (single or multiple) - which would rely on you either taking the item before simming it, or calculating how it would affect your stats.

On that note, are there any reliable Paper Doll item selection-based modeling websites around, a la Warcrafter.net's Sandbox mode? Last time I tried that it was so very very horribly wrong. Who knows, this weekend might be the weekend I'm bored enough to run those billion sims >_>

 

Malan

The new jEnhSim will do this. We are modeling gear on a per-slot basis, to include gems.

Our plan is also to setup the GUI to make very convenient to take 2 sets of gear and compare them against each other. That way you don't have to run a baseline check, then remove or add an item and run again. You'll just have to set up your current gear, indicate what items to swap out for a different item, and then sim it. The report will indicate the DPS difference between the two setups.

Shitting up every single thread on EJ since '06

 

Kyro

I'm wondering if it would be possible to have jEnhSim source all your character data from the WoW armory site. We would enter our name and realm, hit "get character data" (or something like that), and all the info would be downloaded.

Would make it easier to get most of the stats pre-loaded.

11 million WoW subscribers, 99% of whom are retarded and should probably delete their accounts.

 

Malan

We discussed it, and for now the decision is no. We have a lot to do just to get the sim engine constructed for basic functionality without also trying to build an armory scraper. It would be nice for the future, but honestly we'd need a few more very active developers to get that ball rolling.

Shitting up every single thread on EJ since '06

 

Kyro

If I were any good at developing, I'd gladly help with that. I'll ask around and see if anybody I know would be willing to help out with this, since it would be a very helpful addition.

11 million WoW subscribers, 99% of whom are retarded and should probably delete their accounts.

 

Malan

Hm. Actually just took a look at the armory's xml and it's a hell of a lot easier to figure out than wowhead's. I can see if I can throw something together, I'm not too familiar with using http services in java though, but I found a couple of open source projects working on it so I'll try using their code as a starting point.

Might be able to do something like reading the item ID of equipped gear and then looking that up in our local gear representations, then loading that onto the Player object. I'd like to keep the actual data representations of items, enchants, etc local just because of things like Blizzard having incorrect tooltip values, or so that people can change values for testing as they see fit.

Although this would create a dependency on the developers to keep maintaining the product. If at any time Blizzard changed the schema for their xml, we'd need to analyze the changes, adjust the code, and the push out an update. One of our major goals is to make this tool capable of lasting beyond the attention span of the initial developers.

[e] yah thinking on that more, it's definitely a 'nice to have' sort of feature, but I'm not sure that the value added outweighs the man hours for implementation or the dependency it might create on having an active development team.

Shitting up every single thread on EJ since '06

 

Sylvand

The computation would be exponential on the number of item slots (assuming a multiple options for each slot). However the number of slots is fixed for practical applications. If you have some total number of items N with m slots, you maximize the number of combinations at (N/m) items per slot, with the number of combinations being (N/m)^m, or N^m / m^m. Of course with m being constant (17? maybe more if you count glyphs etc.) this will be O(N^m) which would be clearly in P.