Remove Advertisements

A Call to Arms - Cataclysm Mechanics testing

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Re: A Call to Arms - Cataclysm Mechanics testing

Postby tlitp » Tue Mar 08, 2011 8:39 am

Theck asks me to update the BW code, thus I'll request the corresponding data set once more. If anyone manages to find an hour or so to invest in it, we'll be able to compare BW to WW.
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A Call to Arms - Cataclysm Mechanics testing

Postby theckhd » Mon Mar 21, 2011 6:52 pm

PTR requests about the new Grand Crusader implementation:

  1. Grand Crusader Munching
    • build: anything with 2/2 GrCr
    • Gear, Seal, target all irrelevant
    • Goal: Observe the following sequence:
      HPG (GrCr proc) -> Inq -> HPG (GrCr proc) -> AS
      Note whether you gain 1 or 2 Holy Power from the AS.

  2. Holy Power Cap:
    Are the gains restricted by the HoPow cap ? Will AS grant HoPow (either directly, either as a hidden buffer) in any of the above scenarios ?
    • build: anything with 2/2 GrCr
    • Gear, Seal, target all irrelevant
    • Goal: Observe one of the following sequences:
      HPG (GrCr proc; 3 HoPow) -> AS -> delayed finisher
      HPG (GrCr proc; 3 HoPow) -> finisher (EG proc) -> AS -> delayed finisher
      See if you still have 1 Holy Power left over after the delayed finisher

  3. ICD testing
      • build: anything with 2/2 GrCr
      • Gear, Seal, target all irrelevant
      • Goal: Observe one of the following sequences:
        HPG (GrCr proc) -> AS -> HPG (GrCr proc) -> AS
        HPG (GrCr proc) -> AS -> empty/filler -> HPG (GrCr proc) -> AS
        HPG (GrCr proc) -> AS -> HPG (no proc) -> empty/filler -> HPG (GrCr Proc) -> AS
        Note whether you gain a Holy Power from the 2nd AS in any of those sequences.

As usual, upload the log so we can scrutinize.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7800
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - Cataclysm Mechanics testing

Postby Jeremoot » Tue Mar 29, 2011 6:36 am

Windwalk may have been stealth buffed, because I'm really questioning the validity of a 1ppm model. I'm getting back to back procs (3 at a time in some cases), I'm even having it refresh itself. Twice I had 3 back-to-back procs of Windwalk on Nefarian last night. That's 30 straight seconds of 600 dodge rating on multiple occasions.

Maybe I'm just really lucky, but 1ppm seems to be undervaluing this enchant.
User avatar
Jeremoot
 
Posts: 434
Joined: Tue Feb 08, 2011 5:36 pm

Re: A Call to Arms - Cataclysm Mechanics testing

Postby theckhd » Tue Mar 29, 2011 7:13 am

Nothing in your description is inconsistent with existing PPM mechanics. A 1 PPM enchant should give an average of about 2-3 procs per minute, give or take.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7800
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - Cataclysm Mechanics testing

Postby Jeremoot » Tue Mar 29, 2011 7:46 am

I misunderstood then, I thought 1PPM meant an average of one proc per minute.
User avatar
Jeremoot
 
Posts: 434
Joined: Tue Feb 08, 2011 5:36 pm

Re: A Call to Arms - Cataclysm Mechanics testing

Postby matthewseidl » Tue Mar 29, 2011 9:26 am

1PPM is on average 1 proc per minute if all you're doing is auto attacking with a unhasted weapon. With a 2.6 speed 1h weapon, that's a 4.3% chance to proc per attack. In a raid though, you've likely got some haste (10% melee haste), and all your specials have a chance of procing it as well. Since you're throwing a special roughly every 1.5s (CS, ShoR, Judge can all proc it IIRC), that dramatically increases the effective proc rate. Add in reckoning, which doubles your attacks and thus proc chances, and you'll see way more than 1 proc every minute.
matthewseidl
 
Posts: 8
Joined: Fri Nov 06, 2009 11:12 am

Re: A Call to Arms - Cataclysm Mechanics testing

Postby tlitp » Thu Mar 31, 2011 9:18 am

The BW issue has been resolved, thanks to snrfrito's parse. With 71 applications off 883 eligible triggers, one can easily compute the 95% confidence interval (brief theoretical reference) :
Code: Select all
(73/887)±1.96*sqrt((73/887)*(1-73/887)/887) = [0.0641,0.1003]

1 PPM corresponds to 0.0667, being a reasonable estimate of the proc chance.
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A Call to Arms - Cataclysm Mechanics testing

Postby tlitp » Fri Apr 08, 2011 1:54 am

Requesting an additional PTR parse (b13850) :
  • build : anything with 2/2 JotJ
  • gear : anything, having equipped a slow (3.6+) 2H weapon enchanted with Avalanche
  • seal : Righteousness
  • target : any low-lvl dummy (debuffs are irrelevant)
  • attack sequence : only J (avoid autoattacks)
  • goal : cast at least 70 J
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A Call to Arms - Cataclysm Mechanics testing

Postby Xenix » Sat Apr 23, 2011 12:16 pm

theckhd wrote:[*]Seals on 4.0.6 live, part 1
  • build : any
  • gear : any, but equip a slow 2H (3.6+)
  • seal : rotate all of them
  • target : Argent Tournament dummies
  • goal : test Exo as triggers; use the following sequence :
  • attack sequence : cast SoR - one Exo - cast SoJ - one Exo - cast SoT - fully stack Censure - one Exo - cast SoI - five Exo
    (repeat for each ability on the list)
  • note : SoInsight may not proc on the very first cast, therefore each of the two abilities must be cast at least five times before dismissing it as a valid trigger.
[/list]


- Tested with Zin'rokh in Holy gear (11.14% melee haste, 14.48% spell haste) -
SoR: No proc from Exo
SoJ: No proc from Exo
SoT: Stacks from Exo, Stacking from Exo procs extra damage, 5-stack procs from Exo.
SoI: No proc from Exo

Also, when SoT stacks up from Exo, I always get one extra tick at the previous dot damage value (which the combat log shows as from SoT at around the same timestamp as the Exo cast), not just when it's at a 5-stack. Secondly, there were some random ~2-second censure ticks as opposed to the normal 3-second ones. No idea what would have caused that - only procs I can have are Conviction ones. Combat log proving that here and here if it's new behavior.

Edit: After looking closer, the tick time varied anywhere from 2.2 to 3.5 seconds, average 2.64, stdev, 0.22. Average is around where it should be for how much haste I have, so might just be something buggy with the combat log reporting.
Kimurellia - Holy/Protection Paladin - Scions of Destiny - Eredar-US
User avatar
Xenix
 
Posts: 244
Joined: Thu Jun 25, 2009 4:56 am

Re: A Call to Arms - Cataclysm Mechanics testing

Postby theckhd » Thu Apr 28, 2011 9:27 am

The erroneous Censure proc on low-stacks is very strange... Thanks for the data set.
"Theck, Bringer of Numbers and Pounding Headaches," courtesy of Grehn|Skipjack.
MATLAB 5.x, Simcraft 6.x, Call to Arms 6.0, Talent Spec & Glyph Guide 5.x, Blog: Sacred Duty
User avatar
theckhd
Moderator
 
Posts: 7800
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - Cataclysm Mechanics testing

Postby Shathus » Thu Apr 28, 2011 11:22 am

I know it's not the concrete evidence you want (logs) but anecdotal at least, for scenario #2 (testing for HP 'reserve'), I have not noticed this to be the case. Having 3 HP, then hitting AS+ followed by SotR did not leave me at 1 HP. Nor has hitting CS again after AS+ and then using SotR.

If that helps at all, there you go. I'll see if I have time to do any logging tonight before raid.
Shathus
Maintankadonor
 
Posts: 1465
Joined: Sun Apr 06, 2008 5:02 pm

Re: A Call to Arms - Cataclysm Mechanics testing

Postby Xenix » Fri Apr 29, 2011 6:18 pm

theckhd wrote:PTR requests about the new Grand Crusader implementation:

  1. Grand Crusader Munching
    • build: anything with 2/2 GrCr
    • Gear, Seal, target all irrelevant
    • Goal: Observe the following sequence:
      HPG (GrCr proc) -> Inq -> HPG (GrCr proc) -> AS
      Note whether you gain 1 or 2 Holy Power from the AS.



Testing on Live:

Grand Crusader Munching: Doesn't happen (1 HP from the AS, just refreshes the duration of GC) log here

I'll get the other two later tonight.

Edit:
#2: Holy Power Cap: "HPG (GrCr proc; 3 HoPow) -> AS -> delayed finisher" sequence did not result in any HP after the finisher

#3: ICD: "HPG (GrCr proc) -> AS -> empty/filler -> HPG (GrCr proc) -> AS" produced one HP from the AS as expected. Proof in this log that if there is an ICD it's under 6 seconds (I was caught off guard by the second proc so there's 5.7 seconds between them instead of the 4.5 there should have been in that sequence).

With regards to number 3, how would it be possible to observe gaining holy power from "HPG (GrCr proc) -> AS -> HPG (GrCr proc) -> AS"? The last avenger's shield would be your fourth point of holy power, so you couldn't gain it and it wouldn't show up in the combat log as an HP gain.
Kimurellia - Holy/Protection Paladin - Scions of Destiny - Eredar-US
User avatar
Xenix
 
Posts: 244
Joined: Thu Jun 25, 2009 4:56 am

Re: A Call to Arms - Cataclysm Mechanics testing

Postby tlitp » Fri Apr 29, 2011 11:45 pm

Xenix wrote:how would it be possible to observe gaining holy power from "HPG (GrCr proc) -> AS -> HPG (GrCr proc) -> AS"?

By injecting a finisher and working iteratively :
Code: Select all
HPG (GrCr proc) -> finisher -> AS -> HPG (GrCr proc) -> X -> AS

a. Start with no filler/empty between the second AS and its HPG predecessor. If you observe a gain then the (presumed) ICD is lower than 3 seconds.
b. If you don't observe a gain, inject something (say J) into the X slot. If the second AS grants HP then 3.0<ICD<4.5.
c. If you still don't observe a gain, inject another filler (say HW) into X. If you observe a gain then 4.5<ICD<6.0.
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A Call to Arms - Cataclysm Mechanics testing

Postby Xenix » Sat Apr 30, 2011 4:32 am

tlitp wrote:
Xenix wrote:how would it be possible to observe gaining holy power from "HPG (GrCr proc) -> AS -> HPG (GrCr proc) -> AS"?

By injecting a finisher and working iteratively :
Code: Select all
HPG (GrCr proc) -> finisher -> AS -> HPG (GrCr proc) -> X -> AS

a. Start with no filler/empty between the second AS and its HPG predecessor. If you observe a gain then the (presumed) ICD is lower than 3 seconds.
b. If you don't observe a gain, inject something (say J) into the X slot. If the second AS grants HP then 3.0<ICD<4.5.
c. If you still don't observe a gain, inject another filler (say HW) into X. If you observe a gain then 4.5<ICD<6.0.


Got it - so you're looking if the ICD (if any) starts when you gain HP from the Avenger's Shield, not when you gain the GrCr proc, which means you're good to use a finisher to get rid of the HP from the HPG that procs the first GrCr. That makes more sense than the initial rotation that had you generating 4 HP in a row with no dumps. I'll test it out real quick.

Edit: 2 HP gains in a row w/ 1 CS between them - ICD < 3.3 secs w/ latency - log @ here.
Edit 2: Got it down to 2.986 seconds by maxing out the latency compensation slider - if there's an ICD on the GrCr HP gain, it's definitely under 3 seconds. log here.
Kimurellia - Holy/Protection Paladin - Scions of Destiny - Eredar-US
User avatar
Xenix
 
Posts: 244
Joined: Thu Jun 25, 2009 4:56 am

Re: A Call to Arms - Cataclysm Mechanics testing

Postby tlitp » Sat Apr 30, 2011 7:31 am

Thanks for the assistance, Xenix. The only high-prio issue still on the table is the JotJ checkup (seal handling/triggering dynamic effects).
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

PreviousNext

Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: Lawfeer and 1 guest


Remove Advertisements

Who is online

In total there are 2 users online :: 1 registered, 0 hidden and 1 guest (based on users active over the past 5 minutes)
Most users ever online was 380 on Tue Oct 14, 2008 6:28 pm

Users browsing this forum: Lawfeer and 1 guest
?php } else { ?