I've been working with the C++ simulator a lot in the last two days (and ported it to Java, for my own sake) and I'm really confused as to what the purpose of this is:

If you look at all the sims people have been running for the last few pages, the lowest DPS sample from all 100,000 simulations is always 6600. Why are we hard-coding the minimum simulation to be 6600?

Furial, it's not even just the 100% bleed and SR uptime. What Rawr does it calculated the average amount of CPs generated over the duration of a 5 min fight. It then calculates how many of those are needed to keep rip at 100% uptime and SR at exactly 4/5 CPs, depending on gear level, at 100% uptime, ie you only SR at 4 CPs at the T9 gear level. The amount of CPs needed to keep rip and SR running is subtracted from the total amount of CPs generated, which is then divided by 5 to calculate how many FBs are used. Pretty much there is a max of 4 CPs wasted in an entire 5 min fight with Rawr's model for dps.

Ok, just to confirm what has been said above on Deft Ametrines vs Deadly Ametrines, I've rerun the "4xT9 + ilvl258 Offset Gear+Mjolnir & softcap" case with Deft Ametrines in all 5 Yellow sockets and the 4 Blue sockets populated with 1 Nightmare Tear and 3 Delicate Scarlet Rubies. The rest (10 red sockets + 3 Prismatic) with JC ArP and epic ArP gems to reach the 740 ArP rating mark. Here's the result:

Aggressive Stats after all the buffs (full stacked raid + Bloodlust used 1 time
in a 5 min fight)
_______________________

Cat (crit% vs. a boss (-4.8% added)): 66.0587%
Bear (crit% vs. a boss (-4.8% added)): 62.0587%
Cat (AP): 12708
Bear (AP): 9028
Agility: 2395
Strengh: 417
Intellect: 377
Spirit: 354
Mana: 7200
Average MP5 In combat: 181, Average MP5 out of combat: 372
Expertise: 20
Hit: 8.50869%
Armor Penetration: 52.8733%
Haste (from equipment): 21.6469%
+weapon damage: 0
Shield value = 2257


Boss base Armor (before penetration and debuff): 10643
Boss modified Armor (after penetration and debuff): 3978.46
Boss Damage Reduction (after debuff and penetration): 20.7093%
Weapon Speed Reduction % (Total without Bloodlust): 33%
Missed % (Total): 0%
Dodged %: 1.37725%
Not normalized attacks parried by the boss: 0%
_____________________________________

DPS: 9950.48 +/- 1.21541, (Min-Max variance: 47.2758%, Lower dps sample: 6600, Upper dps sample: 11304.2) , Cat: 100%, Bear: 0%, TPS: 7064.84

**************************************
CAT

White cat: 3615.52, Yellow cat: 6334.96
White: 36.3351% (SR: 0%)
Mangle: 0%, Avg dmg: -1.#IND, #/cycle: 0), Shred: 31.9873%, Avg dmg: 11440.6, #/cycle: 7.60308), RIP: 17.2323%, Avg dmg: 49432.1, #/cycle: 0.947965)
Rake: 8.49914%, Avg dmg: 12414.1, #/cycle: 1.86173), Ferocious Bite: 5.94616%, Avg dmg: 21559.7, #/cycle: 0.749986, Average Energy Usage: 43)
Savage Roar uptime: 99.166%

**************************************


For the record, the proper DPS value for Deadly Ametrines instead of Deft Ametrines and the rest as described above is 9941.77 DPS points. The "9919.15" value in my previous post is slightly off due to non-optimal use of the gem sockets (i.e. Nightmare in a Yellow socket and use of Purple gems).

I haven't changed any settings for FBN other than hiding panes I wasn't using. It used to work for me, but has been acting up since it was updated for patch 3.2

Disable lag correction, then. You'll need to do this every time you login or reset your UI, as it's not permanent (unless you edit the LUA itself).

Thank you for a great addon, I look forward to checking out the simulator once the interface is done.

Are there any plans for the FbN addon to do timer bars instead of icon countdowns, and to be able to track non-Feral-applied short-term buffs and debuffs that affects us greatly - i.e., IFF, Sunders, any -Armour debuff, Hysteria, Shattering Throw, Horn of Winter falling off etc?

Also, can someone point me to a discussion on how best to set up the Ability Weightings numbers?

Aye, if you want some timer bars, get classtimer, you can add as many as you like if they aren't already there.

I definitely prefer the icons with the clear numbers over bars--however, I find that I'm missing the little "clock hand" indicating the duration that I had using Debuff Filter. I prefer to have both a visual indicator and a number. I actually caught myself squinting at the little icons on my target frame trying to see the "clock hand." (Debuff filter is otherwise inferior because it doesn't count shreds and doesn't understand that Mangle/Trauma are the same debuff, for example.)

Also, for me, the FbN preference panes don't scroll. All the checkboxes just continue down past the bottom of the preferences window and even down off the screen on tabs that have enough of them.

Edit: coldbear, for tracking other non-feral buffs and debuffs, I have found Debuff Filter extremely useful. It has icons, though, not bars. I still use it to track Faerie Fire (which actually would be a good addition to FbN's icons, as might Infected Wounds).

A work around for the settings scrolling out the bottom of the window is to get ackis recipe list addon (it has some code that lets you resize the entire addon options UI by dragging the lower right corner) although I'm sure there's other ways to do it besides just that addon.

FBN tracks Hysteria btw. I´m always slapping my DK if Hysteria is missed :p

Sadly, since Tun got banned, he could no longer argue his point about validity of the final DPS numbers and the merits of FB versus not. As far as merits of FB go (which he didn't dispute necessarily), I think at his gear level, he may have been right, to a degree.

After defending the great majority of TC models, it may not be worth it to FB. FbN is a little bit of a rut right now in terms of modifications of cycle and fact-checking: Nightcrowler has been out of the picture for a few months, and he hastily included T9 stats/gear (see the various modifications since). There hasn't been much testing done with cycle modifications since 3.2 came out.

Knowing that SC and FbN have been close in the past (and not at all wanting to sort through FbN code), I've been playing with SC a lot. It's a lot easier to modify than FbN, the math is approximately right (despite a few recent fixes), and the cycle is user-editable. So, I went about editing-- with the default gear set.

After adjusting it to properly energy pool instead of shredding non-stop, every attack cycle I could model, from bites every time SR/Rip were up and I had 5 CP, to no bites at all, landed in a range of about 100 DPS. The maximum DPS I could muster only had a bite 3 times in a 5-minute fight, with fully talented FA; one of those bites happens in the last 10 seconds of the fight. Without FA, any FBs were typically worse than the Rip cycle.

I'm not saying I did a perfect job in modeling the attack cycle (I'm still not 100% convinced in what I've done), and I encourage others to duplicate that work (it's only a text file, to edit, after all.) But it's questions like this that we can't really answer in FbN right now-- I haven't seen anyone but Nightcrowler modify cycles in his code, and he's not around right now. Is it worth trying to execute a perfect FB-based cycle and losing 100 DPS? Is the utility of doing that not worth 50 DPS?

The recent discussion here has also all taken place under the assumption that the cycle variable has been constant. Without FBs, can you achieve the same results simply by focusing on Agility instead? (SC seems to still like the soft-cap, but I haven't messed with the gear there.)

We have to clearly distinguish between these two approaches. When you have too many CPs shortly after a Rip refresh and still 15s or so left on SR, then we need FB to get rid of those CPs. But that wasn't actually the point. The question is, if we should still include FB as a static element in our rotation, even if this means less Rip uptime? This has been the agreement over the past couple of months. Right now, like Tun was pointing out, it looks like FB might not be worth it anymore and 100% Rip uptime should get the higher priority again.

Why? I think there are two major influences. First, 4t8 had its greatest effect at SR1. Because of this the loss of Rip uptime due to messy FB timing was always minimal, even in a worst case scenario. Second, haste has become better than crit, so that in higher gear levels the overall crit rate is pretty low and therefore we can't afford to FB in between Rip and SR refreshes. So, if using FB at a cost of Rip uptime isn't worth it, we could very well end up not using FB at all, because we never get these lucky streaks anymore, which is what Tun was describing.

Now, it would be interesting to see if a setup using ArPen soft cap + Delicate/Deadly in order to maintain a FB cycle can result in better DPS or if FB2-4 are options. Actually I doubt it since this would have to include 5FA and most cats that do heroic modes prefer the utility of SI or even NI but this is totally subjective.

That's kind of silly. Nothing about our rotation should be static except the first few opening abilities. Throwing FB's as a static addition into our "rotation" is wrong, quite simply because that's an error in play style, not a matter of numbers.

Sorry, but increasing Damage Per Second is always a matter of numbers. You can't "feelings" your dps higher. You're not a Jedi.

Last night, after getting 2t9 and dropping 4t8, my feelings from playing with the sim were more reinforced. There isn't a lot of extra rotation time at that gear level-- really only biting after berserk streaks.

I am not sure if I can agree there, I have been running 2t9 since there have been enough badges in the game to do so ( 245 ver 2t9 i took some of the first idols in our guild due to nothing else useful dropping... ). I find that 2t9's rake duration increase lets me get in as many if not more FB's then I used to be able to. Though the thing I definitely note is that when FB does not crit and I introduce Rip downtime it can lower my dps over that time frame if I overestimated my energy gain / cp gen ( tf didn't come up soon enough after the fb and rip was down too long etc ).

However I think getting a couple upgrades / 4t9 would help alleviate this phenomenon. Plus we have to consider the fact that there are times that FB is very useful and serves a greater purpose then Rip or refreshing SR would. Both northrend beasts ( the jormungers ) and Anub drop our combo points when they go under ground, and we are able to pool energy on jormungers, and most of us probably swap to swipe for anub minions assuming 3+ targets lined up. FB right before either of these bosses go under is a substantial damage increase, vs either doing nothing or reapplying either of our other two finishers.

The other thing that has been pointed out, and I would like to add a bit too is for those of us running Arp soft-capped. During Runestone procs, Rip can "afford" to have more downtime introduced due to its lack of gain from a Runestone proc. The Dpe of FB goes up the higher our Arp is, even if only temporarily during an Arp trinket proc, allowing for a greater downtime of Rip before using FB is a dps loss, even in a period of lower CP gen.

Maybe what I am seeing is a product of my current gear, maybe not, I would be interested if anyone else is seeing the same kind of effect on their dps of greater percentage Rip up-time vs weaving FB's.

Reduction in energy on Rip is going to free up more energy for Shred/FB, and increased ilevel is going to allow for more ArP stacking than before.

Allev, can you post the SC action priorities you used? I think it would be valuable to use as a starting point, since it sounds like it's a bit different from the default cat profile.

Edit: I found his profiles; here's a link for anyone else who is looking:
Issue 142 - simulationcraft - feral druid: add "energy>=N" condition for "berserk" action - Project Hosting on Google Code

Hrm, I did some napkin math on the new set bonuses. The 2 piece seems very lackluster as it's barely worth 0.45 energy/second whereas our 2T9 is worth ~1energy/second with 100% rake uptime. 4PC on the other hand is really good. Assuming rake is ~10% of our total damage, assuming a 70% crit rate, averaging procs which is probably pretty likely in T10 gear, [.3+(.7*2*1.1*1.03)]*10%=18.862%, so ~8.8% dps increase if my math is right. It's a really solid set bonus.

Yes, but Shred / FB also benefits from agility. I get the impression that even now, after the 3.2.2 arp nerf, the difference between agility and arp is almost negligible. With those set bonuses, you're given an incentive to prioritize your bleeds over Shred / FBing again. and I'm thinking it will push the question over in favor of agility.

It's just a theory, though, and we'll obviously have to wait for the gear stats before any real math can be done.

I wonder how it would look if rip was calculated with an agi set, 100% uptime and glyphed berserk instead of shred?

Considering that with the best in slot gear that exists right now, arpen is worth heavily stacking - it's hard to imagine another 20 ilvls of gear making this untrue. 5% more energy and 5% more damage from abilities you were using anyway almost whenever you could isn't going to make you use other abilities less or make them less damaging via arpen, and there should be enough arpen on gear (one way or another) to hit the hard cap.

If anything, the death knell of arpen stacking may be that the set bonuses are so good relative to the tier pieces not actually having arpen on them.

I have to wonder if these are the real bonuses or just place-holders. All of the four-piece bonuses seem severally overpowered. As an example, consider the prot-warrior 4piece bonus:

Not costing health and giving an almost on-demand 8K damage shield (assuming a 40K health for prot warriors) seems strongly game modifying. Unless Icecrown bosses are going to hit that hard (but Blizzard has stated that they wanted to get away from hard-hitting bosses and use other mechanisms to make the encounter harder).

I would suggest waiting a bit to start theory-crafting these bonuses till we are sure they are the `real' ones.