+ Reply to Thread
Page 5 of 5 FirstFirst ... 3 4 5
Results 81 to 91 of 91

Thread: WoD - Protection Warrior Spreadsheet

  1. #81
    Join Date
    Jul 2014
    Posts
    64
    http://wod.wowhead.com/spell=158298
    effect number 3.

    i'll retest the dummy when i have a chance.
    i remember just taking the damage taken from my buff, and calculating the resolve in my spreadsheet and checking to see if it matched the buff again. it was dead on if i recall correctly. - but maybe i muddled something somewhere along the way.

  2. #82
    Join Date
    Jul 2014
    Posts
    64
    Quote Originally Posted by Frugos View Post
    Edit: Or has the damage of the tank dummies been adjusted, too?
    i'll try to verify tonight.

  3. #83
    Join Date
    Sep 2014
    Posts
    5
    I'm really confused by the dummies now. The damage seems to be half as before, 30-45k (I get hit for 14-21k after armor, def stance and versatility). If I include the 2 multiplier, it should result in ~127% resolve on average without the dummy strike. the dummy strike strike is physical - but how can he result in bumping the resolve to 160-170% on average, if he is just doing ~13% of my damage taken?

    Edit: Your expression in some cells like bSummary!B7 are really long. I suggest to use something like HLOOKUP() in the future to simplify your work. That would result in something short like "HLOOKUP(race;race_table;2;0)" for the race_strength_bonus (race and race_table would simply be named ranges for a cell containing the race and a range containing the lookup table)
    Last edited by Frugos; 09-27-2014 at 04:03 PM.

  4. #84
    Join Date
    Jul 2014
    Posts
    64
    k, i'm showing average melee from the dummy as 35k (raw)
    average raw damage from the dummy strike seems to be 17.5k (sample set obviously smaller).

    resolve buff swings between 232k and 340k damage taken in the last 10 seconds - average ~286
    calculated average damage taken should be 251k to 268k (based on one or two dummy strikes in the last 10 seconds)
    i should likely run some more tests to see exactly how much resolve that dummy swing is giving.

    in any case
    when my resolve buff shows 317k damage taken, it shows a resolve percentage of 177%
    if i input 31.7k dps on the boss in my spreadsheet i get a resolve of 177.13%

    (but obviously if i'm counting that damage incorrectly, that doesn't prove anythign)

    edit:
    i'll look into the vlookup style stuff as well.
    actually, i might throw some time at named ranges, see if i can't make everything a lot more readable that way.
    Last edited by buoy; 09-27-2014 at 09:42 PM.

  5. #85
    Join Date
    Jul 2014
    Posts
    64
    alright, i'm pretty happy with my resolve calculation

    here's the dummy modelled at 35k melees 17.5k dummies (this seems to match in-game quite well)



    then i:
    removed the dummy swing.
    calculated resolve 0.1 seconds after a melee swing (138.52%)
    calculated resolve 1.4 seconds after a melee swing (90.36%)

    averaging those yields 114.44% resolve
    my calculation for a 35k swing every 1.5 seconds yields 114.25%


  6. #86
    Join Date
    Sep 2014
    Posts
    5
    In your bMarkov tab, I18 and and J19 should be "0.7 * N13", not "MIN(N13,0.7)" because "Revenge is usable" and "Sword and Board procs" are not mutually exclusive. The case, that Devastate procs SnB and revenge gets usable the same time is handled by your probability r2 then

  7. #87
    Join Date
    Jul 2014
    Posts
    64
    Quote Originally Posted by Frugos View Post
    In your bMarkov tab, I18 and and J19 should be "0.7 * N13", not "MIN(N13,0.7)" because "Revenge is usable" and "Sword and Board procs" are not mutually exclusive. The case, that Devastate procs SnB and revenge gets usable the same time is handled by your probability r2 then
    Okay, trying to get my head around what you're saying.

    The D1 row corresponds to the first GCD after a shield slam as a devastate.
    There is a 30% chance that when I use that devastate, I will proceed to a SS'
    There is an r1 chance that revenge will proc during this gcd (or be off cd) - pushing me to R2.
    The remainder is what pushes me to D2

    0.7*r1.... because if i proc a revenge, 30% of the time i'm still going to proceed to shield slam.
    i can get on board with that. i'll make sure it's in the next pass.

    just waiting on servers to come back up, trying to verify all the damage quantities again - blizzard snuck a few prot scaling changes by me on the hidden protection aura.

    EDIT:
    thanks for catching that by the way, pretty obscure - i wasn't even sure anyone was looking at that rubbish. glad to have the second set of eyes.

  8. #88
    Join Date
    Jul 2014
    Posts
    64
    v.6.01.024
    • 2014-10-05 armour food gives 50, not 75
    • 2014-10-05 broke down blockBarrier tab calculations a bit
    • 2014-10-05 added tooltip column to abilities list, still needs some work with rounding/floors
    • 2014-10-04 fixed bug with weapon damage normalization calc
    • 2014-10-04 fixed a bug in the markov matrix that proc'd too many revenges (thanks frugos)
    • 2014-10-04 removed old 20% modifier to thunder clap damage
    • 2014-10-04 protection multiplier for execute is 2.0 not 1.4
    • 2014-10-04 fixed bladestorm coefficient to be 1.6*1.4
    • 2014-10-04 ravager is now affected by bloodbath
    • 2014-10-04 fixed bug with unyielding strikes when there are no devastate casts
    • 2014-10-04 removed resolve diminishing returns for multiple mob tanking (build 18988)
    • 2014-10-04 removed 60% healing reduction on barrier, changed resolve factor to 3.4 from 8.5 (build 18988)

  9. #89
    SimCraft is showing a different Statranking for DTPS:
    Armor >Str > Haste > Mastery > Versatility > Crit (T17N)
    Haste >> Crit > Mastery > Str (T16H)

    In your sheet, Haste is a valid option but not for expected DTPS.
    Why this difference?

  10. #90
    Join Date
    Sep 2014
    Posts
    5
    I don't trust simcraft for tanks at all.

    - Shield Block is used together with other defensive abilities, shield barrier else. That doesn't make any sense to me, since stacking percentwise reduction as a tank reduces the effectiveness of every involved skill

    - the scaling factors from simcraft are for "damage reduction" and the effect of shield barrier (and other heals like blood craze) doesn't seem to be involved. more crit means more rage generation means more barrier.

    - there may be other issues like those

  11. #91
    Join Date
    Jul 2014
    Posts
    64
    Quote Originally Posted by Fenris Gul'Dan View Post
    SimCraft is showing a different Statranking for DTPS:
    Armor >Str > Haste > Mastery > Versatility > Crit (T17N)
    Haste >> Crit > Mastery > Str (T16H)

    In your sheet, Haste is a valid option but not for expected DTPS.
    Why this difference?
    The biggest difference between DTPS metrics of simC and my spreadsheet is the incorporation of self healing. simC won't take into account leech/blood craze etc heals, while I try to incorporate that in the spreadsheet. Also my survivability stat weights don't take into account casting spells in place of SS, R, or D. This is unrealistic in the sense that when you cast these other spells your rage reduction goes down. However, if you are truly in a survivability squeeze and need all the rage you can get, you're likely to push back the priority of these other spells until you know that casting them won't kill you.

    The best talent for getting closer results would be to use heavy repercussions as it does not come at a survivability cost.

    Quote Originally Posted by Frugos View Post
    I don't trust simcraft for tanks at all.

    - Shield Block is used together with other defensive abilities, shield barrier else. That doesn't make any sense to me, since stacking percentwise reduction as a tank reduces the effectiveness of every involved skill

    - the scaling factors from simcraft are for "damage reduction" and the effect of shield barrier (and other heals like blood craze) doesn't seem to be involved. more crit means more rage generation means more barrier.

    - there may be other issues like those
    In simC I had originally prevented casting shield block while on another CD - as I was trying to closer model the chain shield blocks then ride other CDs while they recharge. I modified this to appease the heavy repercussions crowd who were complaining that the shield block uptime was dropping and devaluing the talent. With the change to a 12 second cooldown on shield block, I could likely swap it back without too much noise.

    But you're absolutely right about the limitations, which is why I try to maintain both. It allows me to catch bugs on the other platform by trying to account for inconsistencies between them.

    That being said, if the stat weights are drastically different, there's likely a bug in one or the other.


    edit:
    also if you're comparing simC to that spreadsheet, make sure that the character and boss profiles are identical between the platforms. It looks like someone was trying to ram a level 90 profile into my spreadsheet, no level 90 support is intended.
    Last edited by buoy; 10-11-2014 at 01:24 PM.

+ Reply to Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts