Remove Advertisements

A Call to Arms - MoP Mechanics Testing

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Fri Mar 23, 2012 10:20 am

Mechanics changes are about all we can test right now. Things like seal behavior and triggers are unlikely to change with tuning. Damage values are much more volatile (and frankly, I'm not even interested in them until lvl 90 in the first place).
"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: 7803
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby tlitp » Fri Mar 23, 2012 1:35 pm

1. Weapon normalization.
Grab a DS and a BA, use the default gear set, cast 2 SoT and 2 HotR on a low-level dummy (one for each weapon, respectively). Report if the damage values are dissimilar or not.

2. EF and Cons as triggers for dynamic effects.
Enchant any weapon with HS, find any isolated low-level dummy (no clustering, you must hit only one target), proceed to pound it with a sequence of HotR>Eternal Flame (1 HoPow). Use SoI and autoattacks freely. Stop after 30 EF casts (or after the first proc, whichever happens first) and upload the log.
Repeat the procedure using only AAs and Cons. Stop at 50 Cons casts, irregardless of the occurring procs (if any).

3. GrCr.
Being naked (0 hit/exp), using SoI, cast CS on a boss-level dummy. Stop after 30 CS casts and upload the log.

4. AW.
Stack Censure on a low-level dummy, cast a Cons, interrupt any attacks, wait ~3s, cast AW. Report if Cens/Cons update their damage values on-the-fly or not.

5. SoR.
Being naked (0 hit/exp), using SoR, grab a DS and pound a boss-level dummy from the front. Use only autoattacks. Position yourself so that you're able to cleave SoR (a second target in proximity). Stop after 500 AAs (~10 mins) and upload the log.
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A Call to Arms - MoP Mechanics Testing

Postby Jaannaa » Fri Mar 23, 2012 3:39 pm

Rhiannon wrote:Using a weapon with a random haste proc is probably not a great idea for testing proc rates.


My bad >.<

I tested again...

Test #7: Seal of Insight proc rate

Build: 5.0.1 (15464)
AP: 7859
SP: 3908
Weapon: Venerable Dal'Rend's Sacred Charge (2.8) (http://www.wowhead.com/item=42945)
Glyphs: None
Toon level: 85
Dummy level: 60 (attacked from behind)

WoL: http://www.worldoflogs.com/reports/eku7 ... details/0/
Jaannaa
 
Posts: 7
Joined: Thu Mar 22, 2012 12:46 pm

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Fri Mar 23, 2012 6:51 pm

Jaannaa wrote:
Rhiannon wrote:Using a weapon with a random haste proc is probably not a great idea for testing proc rates.


My bad >.<

I tested again...

Test #7: Seal of Insight proc rate

Build: 5.0.1 (15464)
AP: 7859
SP: 3908
Weapon: Venerable Dal'Rend's Sacred Charge (2.8) (http://www.wowhead.com/item=42945)
Glyphs: None
Toon level: 85
Dummy level: 60 (attacked from behind)

WoL: http://www.worldoflogs.com/reports/eku7 ... details/0/

278 procs out of 420 attacks.
Expected proc rate for 15 PPM is 15*2.8/60=0.7000
Observed proc rate is p=0.6619, 95% CI is [0.6167 0.7071]
"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: 7803
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby tlitp » Sat Mar 24, 2012 12:57 am

tlitp wrote:2. EF, HW, Cons as triggers for dynamic effects.
Enchant any weapon with HS, find any isolated low-level dummy (no clustering, you must hit only one target), proceed to pound it with a sequence of HotR>Eternal Flame (1 HoPow). Use SoI and autoattacks freely. Stop after 30 EF casts (or after the first proc, whichever happens first) and upload the log.
Repeat the procedure using only AAs and Cons. Stop at 50 Cons casts, irregardless of the occurring procs (if any).

Repeat the procedure for HW too. Stop at 30 casts or the first proc.

tlitp wrote:3. GrCr.
Being naked (0 hit/exp), using SoI, cast CS on a boss-level dummy. Stop after 30 CS casts and upload the log.

Partly a typo, partly not noticing that GrCr is back to 20% in the latest beta build. The target is 120 casts (~9 mins). Sorry for the confusion.
Last edited by tlitp on Sat Mar 24, 2012 11:34 pm, edited 1 time in total.
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A Call to Arms - MoP Mechanics Testing

Postby Sabindeus » Sat Mar 24, 2012 11:09 pm

http://dl.dropbox.com/u/294821/rotation-test.log.txt

Test #1: General Rotation

Build: 5.0.1 (15464)
AP: 563
SP: 260
Weapon: Longsword
Glyphs: None
My level: 85
Dummy level: Boss (attacked from behind)

Notes:
- My character sheet has the following to say:
Hit Chance: 0%, Miss chance against 88/Skull: 7.50%
Expertise: 1% (from Human racial) Dodge and Parry chance against 88/Skull: 6.50%
- Shield of the Righteous is off the global cooldown (Apparently you all were aware of that but I wasn't so I thought that was significant)
- Holy Wrath has a 9s cooldown. I didn't see it in your list of things to cast so I only started casting it later on.


where can I find 3 dummies close enough to test HotR?
Image
Turn In, an NPC interaction automator - http://wow.curse.com/downloads/wow-addo ... rn-in.aspx
User avatar
Sabindeus
Moderator
 
Posts: 10472
Joined: Mon May 14, 2007 9:24 am

Re: A Call to Arms - MoP Mechanics Testing

Postby tlitp » Sat Mar 24, 2012 11:38 pm

For the moment we're interested in #2 and #8-13. Do note the updated #10-11. As the boss-level dummies seem to be hardcoded at 93 on the beta, #11 is particularly attractive.

We're also interested in Sanctity of Battle. Stuff haste/attack speed wherever possible (spell haste, melee attack speed, haste rating, but no Bloodlust-type effects), and then cast a few CS/HoTR on cooldown; I assume the tooltip won't be "working as intended". Report latency, haste rating, upload the log.
User avatar
tlitp
 
Posts: 556
Joined: Mon Jul 27, 2009 3:25 pm

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Sun Mar 25, 2012 12:14 am

Sabin's log: http://www.worldoflogs.com/reports/hcx2h9038nf3xkk8/

I'm about to go to bed, but I'll look at it in the morning.
"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: 7803
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby Mirawen » Sun Mar 25, 2012 10:46 am

I'm going to have a look at some stuff, probably tonight. Do you want WoL parses, or are raw logs ok?
Image
User avatar
Mirawen
 
Posts: 15
Joined: Tue Jul 12, 2011 9:38 pm
Location: Kolding, Denmark

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Sun Mar 25, 2012 11:26 am

Something seems funny with Sabin's log. Unfortunately, there are no parries - that's my fault, I should have specified attacking from the front (the whole point of this test is to determine what types of avoidance apply to each ability). However, even beyond that, there are some oddities.

Here are the hit/dodge/miss rates for the different abilities:
Code: Select all
  Percentage
Abil  Total  Hit    Dodge  Miss
SotR   92    72.83  14.13  13.04
CS    267    70.41  19.48  10.11
AS    100    31.00   0.00  69.00
J     183    27.87   0.00  72.13
HW     42    33.33   0.00  66.67
Cons  909    30.69   0.00  69.31
Melee 579     0.00  19.34  13.99  (66.67% glancing)


There are a few things of note here. The formula for miss/dodge/parry chance is 3% + 1.5% * (npc_level - player_level). For a 93, that should be 15% dodge/parry/block. Spell miss is 6% + 3% * (npc_level - player_level), which should be 30%. Since Sabin has 1% expertise from racials, we'd expect 14% dodge/parry, 15% melee miss, and 29% spell miss.

The low sample sizes make it tough to draw too many conclusions, because the 95% CI is about +/- 5%. CS's dodge and miss rates fall just outside the CI, but on different sides. However, we're seeing similar dodge results from the melee data. It may be that there's an additional bump for bosses 4+ levels higher.

This is more apparent from the Cons data, which should have a 29% miss rate. From the data, the 95% confidence interval is [66.31% 72.31%] - still a wide range, but way off. Again, we're probably seeing some sort of discontinuity in the formula for level differences greater than 4.

The most curious aspect, however, is the J and AS data. Despite the low sample size, the CI for J misses is [55.13% 82.87%], way over anything we're observing for melee. The fact that J and AS both match the Consecration data strongly suggests that these have been changed from melee attacks that cannot be dodged/parried to spells.

Unfortunately, to narrow down the percentages more finely, we'll need sample sizes of ~5k casts, which is a bit excessive, but possible to do in batches (i.e. if multiple people submit logs of several hundred casts). However, it's not necessary (or useful) to collect this data against a 93 target, since we're rarely going to care about bosses 8 levels higher. Ideally we'd want to test against a level 88 (or a level 82 paladin against an 85 dummy). So let's put this on hold unless someone finds one of those two special circumstances.

I'm going to update test #1 to make it clearer and more specific, and add a test for conclusively determining the nature of J/AS.
"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: 7803
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Sun Mar 25, 2012 11:28 am

Mirawen wrote:I'm going to have a look at some stuff, probably tonight. Do you want WoL parses, or are raw logs ok?

Either is fine.
"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: 7803
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby rodos » Sun Mar 25, 2012 5:06 pm

theckhd wrote:The most curious aspect, however, is the J and AS data. Despite the low sample size, the CI for J misses is [55.13% 82.87%], way over anything we're observing for melee. The fact that J and AS both match the Consecration data strongly suggests that these have been changed from melee attacks that cannot be dodged/parried to spells.

Given that Expertise is now supposed to reduce spell miss chance, I guess this is understandable.
User avatar
rodos
 
Posts: 1120
Joined: Mon Sep 24, 2007 8:20 pm

Re:

Postby Mirawen » Sun Mar 25, 2012 9:30 pm

theckhd wrote:Current Parse Requests:

Test #1: Determine J, AS, CS, Cons, SotR, HW ability types
  • Gear: None (0% hit/exp) except for a fast weapon (Dal Sword, Dal Axe, or Dal Cudgel, pick one that doesn't give you racial expertise). Any shield w/o hit/exp.
  • Seal: Truth
  • Glyphs: None
  • Target: any isolated level+3 dummy, attacking from the front.
  • Attack Sequence: Full rotation: CS>J>AS>Cons>HW with SotR as HP dump
  • Goal: several hundred J, AS, CS, Cons, SotR casts, the more the better. Overall goal is to collect >5k casts cumulatively from multiple sources
  • Report: AP, SP, hit/exp (should both be 0%!), weapon details, upload the combat log


Tested. Results here; disregard the first 196 lines.


Seal used is Truth. I chose to use GoFS to avoid my shield bouncing to the other two dummies nearby (take this into account). Holy Wraths hit the second dummy immediately to my right as well (there are no completely isolated dummies in Orgrimmar, but at least I kept my Consecrations on only one dummy).

MH: Dalaran Sword
Shield: Blackhorn's Mighty Bulwark (HC), reforged dodge => parry, gemmed +25 Mastery, 37 Stamina and +75 Stamina. Otherwise naked.

Level 85 Blood Elf Paladin, giving no additional expertise.

AP: 1,089
SP: 526
Hit: 0%
Exp: 0%
Target: ?? Raid Dummy

Talents:

Tier 1 - Pursuit of Justice
Tier 2 - Fist of Justice
Tier 3 - Sacred Shield
Tier 4 - Hand of Purity
Tier 5 - Sanctified Wrath
Tier 6 - N/A

I think that covers it.
Image
User avatar
Mirawen
 
Posts: 15
Joined: Tue Jul 12, 2011 9:38 pm
Location: Kolding, Denmark

Re: A Call to Arms - MoP Mechanics Testing

Postby theckhd » Sun Mar 25, 2012 10:08 pm

Mirawen's log. Will look at it in detail tomorrow morning.
"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: 7803
Joined: Thu Jul 31, 2008 3:06 pm
Location: Harrisburg, PA

Re: A Call to Arms - MoP Mechanics Testing

Postby Sabindeus » Mon Mar 26, 2012 8:38 am

tlitp wrote:As the boss-level dummies seem to be hardcoded at 93 on the beta,


ohhhhhhh That's why I was missing so damn much.
Image
Turn In, an NPC interaction automator - http://wow.curse.com/downloads/wow-addo ... rn-in.aspx
User avatar
Sabindeus
Moderator
 
Posts: 10472
Joined: Mon May 14, 2007 9:24 am

PreviousNext

Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: No registered users and 1 guest


Remove Advertisements

Who is online

In total there is 1 user online :: 0 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: No registered users and 1 guest
?php } else { ?