Well bought first the levels of advanced BF

[Caave Monster]

I have gotten all the double attack AAs and pet flurry AAs so far.  Just did Keldovan and parsed 480dps (me=378, pet=112 not counting procs) just using disc and chaining growl. I use Yarusha in main and Anguish mace in offhand.  DPS seems to be a bit higher than normal IMO for Keldovan.  There is a very noticeable difference between 15% chance and 30%.

[Tastian]

Fugging blanks.  8(   

I understand most people don't know the data or the numbers or some of the finer details of parsing, but for the love of god the phrasing makes a world of diff.  It's one thing to post:

"Has anyone done any parsing yet to verify the benefits of Improved BF on live servers?"

verse

"Hey I got no idea or data or a hunch or a fairy told me or nothin, but I figured I'd drop a very negative sounding post on people and maybe get some information out of it while making our parse junky/corr want to pound his head through another new computer desk."

Understand what you are talking about is a % of a % and in some case a % of a % of a %.  Right now beastlords with BF5 have a 15% double attack rate.  If you get improved BF and it's like it was in beta you have a 30% double attack rate(for both hands).  That means if you are talking about approx 13% more swings if it were just a matter of swinging a big 2h.  People aren't actually capable of seeing that diff and with how random EQ is you'd need a sizeable parse to actually clear the margin of error anyway.  Understand though that you are talking about a % boost to your melee damage, which is a % boost of your total weapon damage(procs matter), which is a % boost of your overall damage(pet/spells/etc).  Buying improved BF5 isn't going to add 3 mobs to your solo'n rotation or anything.  It will raise your dps and it will show up quickly in an actual parse.  To judge it by eye though is seriously impossible unless it was bugged and removed our original DA ability. 

[Nusa]

Consider yourself lucky, Tastian....At point blank range, blank cartridges are nearly as dangerous as the real thing. Some of you may remember the story of Jon-Erik Hexum (Phineas Bogg in in the Voyagers TV series and Mac Harper in the Cover Up TV series) fatally shooting himself with a blank gun in 1984.

[iamweaver]

OK I ran baselines for 4 hours today.  Unlike DPS parses, you need a much smaller sample size, as there is only one variable to be considered here; we actually don't care about number of hits or DPS or anything, only the number of swings.  I should have 5 AAs ready by tomorrow to see how much it is for rank 1.

[iamweaver]

One important note:  remember that "baseline" tests are already run with Dual Wield at 15%, Benefits gained are from 1.15 to the new percentage (touted at 1.18, 1.21, 1.24, 1.27, 1.3).  The actual gain is a little less than 3% per rank; for each test I will list the expected gain.

Tests are run with a level 70 BST, max weapon proficiencies, attacking the Katta banker Bella Helsin from behind. For those that don't know, she is lvl 1 with raid-level regen, and doesn't attack.

Baseline testing was run for 3 hours using no buffs with Bestial Fury 5. To maximize the number of swings, Lupic equipped his worn haste item [Shroud of Survival] (46% haste), and dual wielded.  The actual weapons used for testing are relatively unimportant, as I care not at all for DPS, only for the percent change in number of swings. Used [Fist of Five Blades] in MH, [Bloodforge Hammer] in OH.  This also let me do some MH/OH comparisons, as 1 is crush damage, the other punch.

BF5 Baseline (including crits):
MH swings/hour: 3160.5
OH swings/hour: 2563
This resulted in dual wield/double attack of 81.8% for OH.  All OH results are shakier, statistically, than the MH results, since you have 2 calls to the RNG for each swing..


Improved Bestial Fury 1 results (5 hour test)
5 hour test.  Expected gain: 2.61%
MH swings/hour: 3221.2
OH swings/hour: 2649
DW+DA: 82.24%

MN net gain: 1.92%; OH net gain: 3.35%
Parsed Double attack estimate of 18.03% (up from 15%)

One note:  hour-by-hour results for the IBF test consistently showed a lower net gain in MH over OH.  Very odd.  Perhaps someone has an idea why this might be? (*** edit:  It could simply be due to the small sample sizes for the baseline test.  I was going to run it all night, but I got an offer to to a Tipt run, and Lupic didn't have the KT flag yet, so... ***)

[Bengali]

I always found it simpler to run DA attack rate parses using a 2hb, so I wouldn't have to worry about the offhand or account for failed dual wield checks.

Anyway, I bought all 5 levels of it at once, and compared it to a pre-purchase test that showed my DA rate at 16%.  After the second test, I showed a DA rate of about 32%, and both parses include a Fero 2 item that I had forgotten I had.  :-P  The parses weren't particularly long, about 1.5 hours each, but they involved thousands upon thousands of swings.

So far my tests show that it has the same gain as the original BF had.  It costs more, but BF came out in PoP where AAs were much slower to gain than they are now, so the cost of everything has gone up due to "inflation" so to speak.   Anyway, if you (the original poster) don't think doubling your DA rate (i.e., a 100% gain) is worth 15 aa then I don't know what to tell you.  Especially if you think a 30% additional proc rate would be worth your time for the same price.

EDIT:  Oops, I forgot it's 5 per level so it's 25 to max, sorry :)

[Gunzak]

Well it would be 25aa to max not 15aa which is what it cost me for level 3.  I guess I was wrong.  What I should have doen is what Tastian said.  Post if anyone has done any parses post patch to see if there was a big difference not say I don't see a difference.  In the future if I have any questions I'll be more thoughtful before I post.

[Tastian]

Thanks gun.  You just have to understand how much schit I've had to deal with even before this expansion went live.  Even with real data out there people are still making false claims based on speculation and the like. 

The way to test double attack is to either use a 2h or just take 2 different weapons(blunt/hth) and swing away UH-HASTED.  Getting your haste too low or having your DW rate and other things factor in just adds more variables.

Long as you keep your delay high enough then eat melee round will be independent and you won't have triple attacks and such skewing your results.  Every test I've ran has taken DA rate for both MH and OH from ~15% to ~30% without fero factored in.  I always welcome more parses and data, but you have to be careful how you parse and what you are actually testing. 

[iamweaver]

I always found it simpler to run DA attack rate parses using a 2hb, so I wouldn't have to worry about the offhand or account for failed dual wield checks

If you dual-wield, you can just ignore the OH results. I usually ran my parses that way, just to see a number of other things outside the scope of IBF.  Besides, it lets you see if something odd is going on with OH  double-attacks, in a general fashion.  Nonetheless, it appears that 3 hours was just too short a sample size. I can't go back and rerun at IBF0, of course, but I will probably rerun Lupic for a day in here somewhere, or at least 8-10 hours.

Getting your haste too low or having your DW rate and other things factor in just adds more variables.

Rats. I was unaware that haste did anything other than lower your delay by a set amount. I know that there was an urban myth going around for a while about getting weapons below 10, but I had assumed that only 46% haste on a delay 19 weapon shouldn't skew anything.  I can rerun the results without haste, tho obv. that should cut down on the number of samples by 46%, and means that I should really run for 12+ hours to lower the sigma.

[Tastian]

The reason it works good to have either 2h or a slow dw(unhasted) is well a few reasons.

First using a slow DW combo helps because BF got bugged at one point to where it wasnt' effecting the OH.  Every since then i've been sure to include a DW parse when checking DA to make sure that the bug doesn't show up again.  Also since the DA is applied evenly to each hand this basically gives you double the sample size over the same period of time.

The 2h works good to remove the potential for rounds interfering with things.  EQ timestamps to 1 sec(10 del) and if you are under that then you run the risk of multiple rounds showing up in the same time stamp.  If I have my moss-twig offhand and I'm hasted I can show triple attacks with my OH and even quad attacks, but there is no way for the OH to quad, what is happening is that multiple rounds are being tied together and since the random nature doesn't let us know for sure if it was a double and a single or a single and a miss or whatever that just cases problems.

Also when using multiple weapons of the same type you then have you factor in DW rate and work backwards to account for the triple/quad hits(even if you aren't hasted).

Finally, there is no need to parse BF5 because it's been parsed to death for years now.  Your base DA rate with BF5 is ~15%.  Run a parse at IBF5 and see what you get for a rate(~30%).  The return per rank is the same (15%/5ranks) so ~3% per rank.  That's the same boost that original BF gave. 

Also haste works such that your delay becomes del/100+haste.  Haste in EQ is a function of swings, not a function of delay.  Minor detail, but worth mentioning every now and then.

If you want to parse Improved BF yourself just have a 2h weapon or 2 1h weapons of different types with a 20+delay and remove all haste effects.  Swing away on the katta banker or whatever for a couple of hours and you'll see a decently accurate sample based upon what is being tested and margin of error. 

[iamweaver]

EQ timestamps to 1 sec(10 del) and if you are under that then you run the risk of multiple rounds showing up in the same time stamp.

Ahh, I see what you are saying.  but honestly, I am not looking to see how many attacks are double-attacks per se.  If I know how many attacks total I did, then I can calculate exactly how many of those are double attacks. My real problem is two-fold.  1) I forgot that I don't HAVE to parse for baseline, since baseline can be calculated. 2) I don't know the granularity of the SoE combat engine.

Wielding a 19-delay weapon with 46% haste,  1.9/1.46=1.3014 seconds per swing, but the engine most likely runs in 0.1-second increments.  But does it round? Truncate?  Step-function?  Cutting out haste eliminates this question.  As a nebbish kinda guy, I will point out however that 3160.5 swings/hour at BF5 (from above) turns into 2948.3 single-attack swings/hour which would be 1.31 swings/second, which points to step-function. This means that *perhaps*, my parse shows a 17.2% Double-Attack from IBF1.

But the real answer is just to rerun the tests unhasted, so that I can compare the swings/hour with the expected swings/hour (1894.7) :)  I will kick that off before I head out for the evening.

[Tastian]

EQ truncates in almost all examples I can think of atm.

Also BF is a constant boost and so is improved BF that's from testing and talking with the dev that coded it.

Your way of testing just looking at total swings works ok and especially if you remove OH or just use two different weapon types it can be easy to get a clear picture of what is going on.

For people that don't like math or parsing or anything like that though all they have to do is put on the right weapons, type /log on, smack the katta banker overnight, then in the morning run the log through EQcompanion and they'll see 30% or 18% or whatever in the double attack field.

[Kanan]

thanks for the parsing advice and explanation tast :)

[Gunzak]

I think what I will do then is finish IBF5 then finish SCF3 then do Ambi and WA5.  Should be about 50aa total.  Gonna need more MPG,  DoN and RSS groups I guess.

[Tastian]

That's a pretty solid idea gun.  Those are some of the best offensive AAs we've got.  Once you get ambi don't forget about sinister too.  That's good for ~3 more dps on its own.  8)