Torpesh

All of the links in the first three posts are for spreadsheets, which as far as I can tell use theorycrafted numbers and models. I don't want to try to model a combat situation, I want to try to simulate it.

There's a tanking simulator on the druid wiki, but I'm looking to write something for DPS.

 

Torpesh

That depends to a large extent to who your tanking and how good your healers are. You can't generate hate if you're dead

Also, what will you have to do to remain crit immune if you lose the 21 defense here?

 

Immortal

Well, i think this item is good for some kind of max-tps set. And you can always use parts of PvP gear to remain crit immune after losing some defense.

 

Torpesh

Absolutely. My point is just that he's asking for a comparison of the TPS stat vs. mitigation stats. You can't really compare the two outside of a specific context.

As an example, when I'm tanking Leotheras I always put on my [Shapeshifter's Signet] instead of my [Shermanar Great-Ring] because the most important thing is picking him up after whirlwinds and you often only have one or two shots at him before he runs off, so you have to make sure they hit. Plus, Leo hits like a sissy. On the other hand, I would never think of making that switch against Morogrim because he hits like a tank and threat is never a problem in that fight since DPS is often taking time to kill murlocs.

Neither ring is necessarily "better" than the other.

 

Allev

I was hoping you'd find Toskk's page, which isn't exactly a simulator. Toskk tried writing it in simulator-form, but ran into problems with web browsers thinking the page crashed (he uses his GetDPS() function on that page to evaluate every item, which means the simulator has to run a ridiculous number of times to get good precision). You may want to inquire with him about whether he has a build of it available.

Why do you want a straight-up simulator for DPS, though? It's not like you have to analyze spikes in personal DPS like you do with incoming damage, and is as vulnerable to mechanics flaws as the theorycraft.

 

Fariseo

You are right, I forgot to mention somethin about my gear. My stats -raid buffed- : 32k armor, 20/21k hp, 50% dodge, 2200 ap, 33% crit and 62 hit rating (still without GoA and BoM). Crit-immunity almost capped (0.4% to reach the cap) ora capped losing somethiing like 10 stamina and 1 dodge.
My idea is Brooch of Deftness + Shadowmoon Insignia (already passed 1 to our warr MT, next shall be mine) gaining enough def for crit immunity, giving away the Moroes pocket watch.

 

Torpesh

There are four reasons why I want to write a combat simulator:

1. Theorycraft models vary and are only as good as their assumptions. Assuming that I get the mechanics right (a big assumption, but I'm an optimist), a simulator should be very accurate. There will still be issues with it being PC driven - it won't account for mis-timed powershifts, etc., but it should be pretty close. Every model out there for feral DPS has to make certain assumptions and in a lot of cases those assumptions can make a big difference. This is obvious from the fact that some models differ from each other by a substantial amount.

2. Some things are just too tricky to model. For example, some models make haste simply increase DPS by a scalable amount. Your attacks increase in speed by 2% so your DPS goes up by 2%. The problem is, feral combat has a LOT of interdependent parts. If I'm attacking more frequently I'm probably critting more, I'm getting more OOC procs, I'm getting more energy back from the 2T4 bonus, etc. Of course, if I manage to get more free attacks I might get to a Rip faster. There's also the case where a Rip may be available back-to-back with another. In that case extra energy or free yellow attacks to get more CPs don't give as great an effect as they would otherwise, because you can't get a Rip up any faster than you already have.

3. It's easier to make changes to DPS strategies or to add new mechanics. This is mostly just IMO, but if I decide that I want to try shred to CP5 instead of shred to CP4 then it's easy to make that change in the code's logic. In a model, depending on assumptions that you've made, you may need to remodel parts of it. This can also be true as Blizzard adds things like expertise. Of course, it depends on the model itself, but a lot of changes to feral combat can mean a complete overhaul for a model.

4. I'm getting tired of database design and programming. I'm itching to get back to some C++ or C#. I've practically forgotten how to program in anything other than SQL.

The models that some people have put out there are fantastic. I just think that something like this to go along with them might be useful.

 

Malazaar

I don't get what you mean by DPS simulator in contrast to a DPS calculator (like Toskks). The only thing i can assume that you mean a simulator takes actual dice rolls instead of modelling statistically. The consequence would be that it is just more inaccurate than a DPS calculator.

I wrote a DPS calculator but it is just code atm (no interface). I pretty much gave up on it since i realised Toskk has effectively done the same and already a good interface standing. Just look over the code (it's simple javascript so you won't have problems if you know C) and you'll see that it takes quite some factors into account.

All simple combatstats (like haste as you mentioned) are calculated correctly and are as accurately as it gets in a model. The one thing that is hard to model are synergy effects with procs and haste and lag / human error.

 

anathor

I have a little question re: +hit, just to see if I've missed something or whether there are just some dark times ahead. As of 2.3 (which is to say, as of tomorrow or so) my Shapeshifter's signet will go back to the bank and I will be losing, like most of us, 3% hit right there. According to what I've read so far, +hit for ferals (talking about dps here) was quite good but not good enough to gear for it instead of agi (point for point). I still managed, grabbing things here and there, to get about 80 hit rating, which coupled with feral weapon skill would bring me reasonably close to the hit cap.

However this situation will change soon, and as much as missing 0.5% hit didn't look like a big deal for a kitty, missing 3.5% seems like another thing entirely. I have browsed the list of gear currently in game and I've noticed that most of the dps gear with +hit came from either Karazhan or t6 content, but few from t5. I think the fact that there is so much +hit on t6 gear is that there's about no itemization for ferals there, so we must share the loot of rogues, and rogue items typically have a lot of +hit.

Anyhow, what are the best options right now to close the huge +hit gap that we will experience soon? (and which won't go away easily since e.g. the Stranglestaff's replacement from ZA doesn't have +hit). I'm almost considering going to Romulo's Vial for the +35 hit (but not quite). And more particularly, what should a cat wear in t6 content, when not able to get t6 gear yet? The only competitive items I can think of with +hit on them are Edgewalker Longboots, Garona's Signet Ring, Worgen Claw Necklace and the Stranglestaff. There's also Midnight Legguards but they are not so good for druids anyway.

Or should I just forget about +hit altogether, accept my ~3% miss rate and just keep on stacking agi?

 

Fariseo

I suggest you to keep your +hit as close to the cap as you can. Can't access T6 uh? Well these new badges (off-tier slots) rewards will help you :
Nyn'jah's Tabi Boots - Items - World of Warcraft (better than Edgewalker)
Waistguard of the Great Beast - Items - World of Warcraft

+40 base hitrating (+3 socket bonus) and decent stats, total cost 120 badges :|

Almost forgot, try to get an arena staff, even the merciless'one after the cost reduction.

 

Allev

There's plenty of items with +hit out there that are pretty competitive with the AP buff/STR nerf.

First, the S3 gear has lots of scattered hit rating.

Other select pieces:

Back: [Drape of the Dark Reavers]
Legs: T5 pants, Skulker's
Waist: [Belt of Deep Shadow]
Chest: [Bloodsea Brigand's Vest], crafted Kara chest among others
Feet: Lots of options-- Edgewalkers, BoNG, Nyn'jah's, etc.
Neck: Look at [Brooch of Deftness] if you REALLY want to maximize +hit in a slot, with expertise having a similar effect.
Trinket: [Tsunami Talisman] is pretty hot here.

You might even want to use yellow gem slots even for lackluster bonuses.

 

Allev

Very accurate given a very large number of cycles. Again, what is the practical use? Especially Toskk's work has been consistent and accurate enough to both ballpark my DPS limitations to a fairly consistent degree. Given the ballpark is right, all the practical use cases are covered in gear upgrades and so on, and there's simply too much gear to run a simulator for each piece.

I guess where the simulator would shine is in evaluating DPS cycles, but luckily Toskk has already done that legwork with his calculator.

First-- the testing available on OoC procs isn't at all conclusive as to the mechanism. The last I'd heard, the most extensive testing of OoC procs is that it has a high likelihood of proccing at the very beginning of a fight, and less over extended periods; but has no cooldown, as you can get back-to-back procs or refreshes of the OoC effect. Similarly, there's been no testing at all as to whether this proc is significantly affected by haste-- probably because no druids have ever been in a position to acquire a large amount of haste gear.

I could take a guess at mechanics (my first guess would be two independent timers able to proc OoC, one with a higher chance with cooldown and the other with a lower chance), but that's really all it is-- a guess.

Finally, while the scenarios you point out are relevant and show a weakness of a calculator, a simulator would still have to make assumptions of the perfect things to do in every single scenario (or at least perfectly typical). What do you do when you have 3CPs, you get feared, and you come back with no mangle up? OK, you mangle, and it crits-- then what? Rip, shred as long as you can? And you have every one of these odd situations to look at, because they'll occur on a long simulation (even on a sequence of misses). If you add to the REAL benefits of simulation, i.e. modeling actual bossfights, you get a lot more decisions that you have to account for.

I know that Toskk has rewritten his model from scratch at least twice and has made countless tweaks to it, in addition to evaluating several different DPS cycles to find the more effective ones. In addition, since we're talking code, while it will be true that YOU can make the changes, the typical end-user won't.

I will say that its most compelling use in for modeling multiple types of combat, i.e. different cycles on a static boss, or bosses where you have to run out every 40 seconds (thus needing very precise cycles), or bosses with interrupt/mind control procs, and if you do start on this project, that is where I believe the most useful information would lie.

 

seminarca

Brooch of Deftness will be the de facto trash neck for most tanks, imo. The amount of avoidance negation you get from the one item is huge, and it has a gob of Stamina too. Personally, I'll be using it on quite a few bosses as well (Hydross, Leotheras, Void Reaver, Solarian) since I'm currently using Necklace of the Deep for tanking, not needing the defense on Juggernaut.

As people have said, it'll depend largely on what you're tanking.

 

• Voldin

Feral Druid DPS 2.3.0

- Fixed Rip damage to match the Wowwiki Data
- Fixed Double Dipping on Mangle Buff
- Improved the Import/Export Functionality
- Fixed (I hope) Dodge and expertise Calculations
- Updated AEP Values for Armor Pen and Expertise
- Added Pawn Export Strings
- Cosmetic Changes

 

Fenris

alright, so i am absolutely terrible at math. 100% horrid at it. so perhaps someone can help me here.

if i get the 3 peices of s3 honor loot (boots,bracers (with socket bonus),belt) adding up to 71 resilience, along with the 23 defense of shemanar, 19 defense of violet signet, and the 22 defense of juggernaut, that should be put over the crit-immune cap, right?

as i said, im really bad at math, but im pretty sure that would put me far enough over the cap to lose the vindicators belt, and get Waistguard of the Great Beast - Items - World of Warcraft and still remain capped, or just below it? that would be 64 defense rating, and 45 resilience without the belt.

that would be a bit over 1% in +defense, and about 1.4% in resilience according to Combat rating system - WoWWiki - Your guide to the World of Warcraft ,i think.

 

seminarca

Using the EU WoW Forum post in the OP, 39.4 resilience rating grants 1% crit reduction, so your 45 resilience gives 1.14%. 2.4 defense rating grants 1 defense skill, so with 64 defense rating you'll have 26 defense skill (floored) which gives a crit reduction of 1.04 (0.04% crit reduction per defense skill). This is a combined total of 2.18% which is still 0.42% short of the 2.6% crit reduction you need from gear to be crit immune vs bosses provided you have 3/3 Survival of the Fittest.

 

Tuhalu

You have the [Ashtongue Talisman of Lethality] in your data when actually druids can only use [Ashtongue Talisman of Equilibrium].

Also, you have [Berserker's Call] set to Pre-BC in your Trinket page. I didn't even realize you had it in there until I dug around to figure out why it wasn't showing up. You should probably add an "Items from Instances: Zul'Aman" option.

 

Beatus

On the topic of +hit, a much forgotten trinket is the [Icon of Unyielding Courage] from heroic Blood Furnace. I am currently using this in my tank and dps set, to fill the gap. (Nice on use too)

 

Mijae

Dodge calculation is still wrong (it should be expertise * 0.25%).
Miss calc is still different between melee and specials (effectively only for non-boss fights).
For crit calculation, you add a constant 21.2 agility. Why?
Specials crit percent does not reflect 2 roll system. This was correct in previous version.

 

Rannasha

I've written a simple combat simulator for Cat Druids about 2 months back, to investigate the actual usefulness of +hit (it turns out that getting to the hitcap is overrated and point for point, hitrating is not as useful as plain stats like STR or AGI). It's rather basic, has no user-interaction (all DPS stats are hardcoded) and the coding is far from pretty (I'm more of a "I'm happy it compiles" programmer), but you can use it as a basis for your own program if you want. With the settings that it currently has, the deviation of the outcome of one run from the long term average is about 0.2%.

The program is written in C and compiled with the OpenWatcom compiler under Windows. I'm not entirely sure whether it will immediately compile on other platforms, but no more than minor tweaking should be required.

http://www.wormeester.nl/WoW/catdps.c

 

Vaccine

The problem with that is that there are generally much better options for the trinket slot, whereas you can pick up hit all over on other gear.

For the neck I'll use it on all trash and any boss where you can't really afford misses, Leotheras being given as a good example. Hydross another one for picking up adds. Its a nice neck for any fight where either high TPS is needed or fast pickups where a missed mangle is going to mean someone else gets aggro.

 

Carlos

Going some pages back due to a discussion elswhere, I found the last sentence of Boevis confusing...

My Swipes going up to 450 on crit. So if the average Swipe is about 210 with 30% crit in tacking gear that would mean an average damage of 273 damage from Swipe which is just above your value.

Did I calculate something wrong or was that allready taken into account?

 

Vaccine

No you haven't calculated anything wrong. I don't know what Boevis was ranting about there, his post is confusing.

I think hes trying to justify keeping a max stack of Lacerate ticking in between swipe spam.

This is 201.5 per tick of lacerate with mangle up.
The 4 is how many ticks he has before renewing.
The .2 is the threat modifier on lacerates bleed effects so the 806 bleed damage only works out as 161.2 damage worth of threat. Then he adds in the threat of the lacerate that hes using to renew the stack and adding the previous 4 (or next 4) mangle ticks to that lacerate's threat.

Lacerate threat is:
So his 295 flat threat figure includes the upfront damage of 50.

The bit that confuses me is the leap from a 456.2 threat lacerate stack refresh to a 272 swipe. I can't remember if swipe has a threat modifier but as far as I'm aware the old 175% one was removed n 2.0.10 and it was just pure damage done from then.

So the 272 figure is strange. Unless I'm totally missing something 456.2 is the exact damage swipe would need to do to beat even keeping lacerate up at all.

Including bearform and bearfiance the 272 is 394.4 threat whereas if you include that in the lacerate + 4 bleeds one it comes out at 661.49.

So to sum up as far as I'm aware, your lacerate is doing a flat 295 aggro on a hit. You also get maybe 1 tick per application for another 155/0.2 = 31 threat so 326 threat. So you'd need a 326 damage average on your swipe to beat this, less on bleed immune.

 

WhitehornsRT

nice post dukes,

just a quick question tho, it's about haste rating, i've seen no post about it, not for druids anyway, if there is i can't find it, anyway, i just woundered if it's worth having for a druid, maybe not on the gear as such, but in trinkets etc, are we capped at the haste rating or can we get more if it's worth getting without sacrificing too much +hit/crit/AP?

 

Allev

More Voldin's spreadsheet stuff:
- Wildfury Greatstaff feral AP is wrong
- adding/removing GoA is changing my unbuffed DPS, for unknown reasons

Feature request:
- for non-bleedable, have a non-finisher cycle, i.e. mangle->shred till 12s->regen->mangle->shred till 12s->....