Jump to content


Photo

Prot [5.4] - EF You


  • Please log in to reply
69 replies to this topic

#41 Konke

Konke

    Glass Joe

  • Members
  • 14 posts

Posted 30 September 2013 - 03:11 PM

Wondering one thing about the ef dicussion: Ef benefits from haste, we all know that. Now im not having that much haste on gear as its mine offspec and just taking anything which i can. So is there point where certain levels of haste gives that edge to ef and if you are too low on haste you should stick with ss?
"He's got to make his own mistakes and learn to mend the mess he makes. He's old enough to know what's right, but young enough not to choose it. He's noble enough to win the world, but weak enough to lose it. He's a New World Man..."

Rush - New world man

#42 Theck

Theck

    Don Flamenco

  • Members
  • 296 posts

Posted 30 September 2013 - 03:58 PM

Wrathblood: what are you basing the 25H results on? The TMI bosses in 540-3 aren't fine-tuned, they're just rough placeholders based on normal-mode logs. 540-4 should have them correctly calibrated for Garrosh melee on each difficulty level.

Wondering one thing about the ef dicussion: Ef benefits from haste, we all know that. Now im not having that much haste on gear as its mine offspec and just taking anything which i can. So is there point where certain levels of haste gives that edge to ef and if you are too low on haste you should stick with ss?


No, both EF and SS scale almost identically with haste.

#43 Wrathblood

Wrathblood

    Don Flamenco

  • • Guide Author
  • 351 posts

Posted 30 September 2013 - 06:06 PM

Theck, I'm just using 540-3. While the numbers may not be fine tuned I'm actually ok with using them as they are. IIRC, 25H produces rough DTPS numbers in the 320k range (and TMIs expressed in scientific notation) which seems adequately hard-hitting to me to reflect 25H.

In part, I think that's because you're thinking of "25H" as being "Garrosh on 25H" whereas I'm thinking of 25H as being "generic 25H, ranging from the not overly scary 25H Immerseus to the 'freight train that dual-wields Lich Kings' that is 25H Garrosh". As long as I'm clear on my descriptions I figure that's ok.

#44 Theck

Theck

    Don Flamenco

  • Members
  • 296 posts

Posted 30 September 2013 - 06:33 PM

Sure, but the point of a TMI boss is to emulate the hardest-hitting boss in the tier, which is pretty clearly Garrosh. And the 540-3 TMI bosses don't nearly do him justice. For example, here are the 540-3 raw swing sizes:
T16N10: 1200k
T16N25: 1350k
T16H10: 1500k
T16H25: 1650k

Here's what I tentatively have them set at in 540-4, based on updated log browsing (and information that Garrosh 25H auto-attacks for 800k-850k per swing):
T16N10: 1250k
T16N25: 1550k
T16H10: 1900k
T16H25: 2300k

#45 Wrathblood

Wrathblood

    Don Flamenco

  • • Guide Author
  • 351 posts

Posted 01 October 2013 - 02:28 AM

That raises an interesting philosophical issue, though. My general rule of thumb is that EF>SS unless you're in danger of being 2-shot. If this is correct, EF will be (as I understand the fights to be) the talent to use for the majority of even 25H. But on the 25H setting in 540-4, if I'm correct, EF will likely be clearly inferior to SS in terms of TMI (the numbers are going to be astronomically gigantic) steering people wrongly.

Simcraft (Prot paladin module) is your baby and your vision for it is the one that matters. What is that vision?

#46 Theck

Theck

    Don Flamenco

  • Members
  • 296 posts

Posted 03 October 2013 - 03:24 PM

The numbers won't be as gigantic as you think. With the T16H profile, that T16H25 boss that auto-attacks for 2.3M damage still only gives a TMI around 300. Of course, it's much higher for the T16N profile, but by the time you're pulling 25H Garrosh you should be far ahead of the T16N profile (which has 0/2 upgrades on everything).

Also, EF should dominate in TMI under all circumstances in the sim. Even for the heavy-hitter, the fact that we're ignoring external healing means that raw absorb or heal is what matters, and EF wins that battle.

It would be interesting, however, to compare EF vs. SS in a sim where you add a healer with a simple "keep up HoTs and cast Heal/Greater Heal based on tank health" priority queue. I think in that case, SS might fare significantly better, which might be a better comparison for the "in danger of being one-shot" consideration.

#47 Wrathblood

Wrathblood

    Don Flamenco

  • • Guide Author
  • 351 posts

Posted 04 October 2013 - 09:19 PM

Ah, good point about EF dominating.

I'm a little shocked that 25H only comes in at 300 against the T16H profile, as I'm ilevel 552ish (not terribly far off the T16N profile though admittedly missing set bonuses and with poor itemization) and my gear against the 25N gives the gigantic numbers I mentioned before.

This also raises the question about vision I mentioned before. Someone who is working on T16H25 would actually want to use the T16N25 boss profile most of the time as only someone working on Garrosh himself would be interested in T16H25. So the profile really is only relevant for a single fight. On the other hand, its *THE* fight of the tier (and expansion) so if any fight deserves it, this is the fight.

#48 Theck

Theck

    Don Flamenco

  • Members
  • 296 posts

Posted 05 October 2013 - 06:00 AM

Another option is just to have even increments of bosses, so that you can just keep climbing the ladder from boss to boss. The downsides are that a long list of bosses may be more confusing and you'd have to remember that e.g. Garrosh is T16-X, Thok is T16-Y, etc.

#49 Capstone

Capstone

    Piston Honda

  • Members
  • 189 posts

Posted 05 October 2013 - 08:38 AM

At that point, why wouldn't you just implement the bosses themselves as profiles? ^.^

#50 Theck

Theck

    Don Flamenco

  • Members
  • 296 posts

Posted 05 October 2013 - 02:34 PM

It's certainly one way to do it. Though it does sort of break down the normalization system. In other words, as it is now, I can be reasonably certain that I'll get a TMI of ~1k-10k for a character that's appropriately geared for the hardest-hitting boss in the instance. If we start introducing more variation in the bosses, that breaks down a bit. For example, if we implemented a Lei Shi boss, it would give us radically different results from Garrosh or Siegecrafter.

I'm not opposed to that route, and maybe when I overhaul TMI for v2.0 (probably with the next expansion) that's one of the changes we'll make. I'm also likely to overhaul the normalization system so Wrathblood stops pestering me about it. :P

#51 Wrathblood

Wrathblood

    Don Flamenco

  • • Guide Author
  • 351 posts

Posted 07 October 2013 - 01:39 AM

Hey, its been almost 72 hours since I brought it up last. I figured you'd enjoy a break from the harassment.

I like the concept of TMI, I just fear people won't embrace a stat that can produce results ranging from 10 to 8.7x10^12.

Edit - Just saw 540-4 is up. Thanks for all of your work on it!

Updated T16H25 boss damage? I'm curious how it looks.

#52 Theck

Theck

    Don Flamenco

  • Members
  • 296 posts

Posted 07 October 2013 - 02:05 AM

Yeah, T16H bosses were updated a bit. They should be a decent bit more challenging now. Unfortunately I didn't finish tweaking their DoT damage in time - I made some more changes to that today, but those won't go in until 540-5 as a result. Not major changes, just boosting the tick size by about 20%-30%. Basically, I fixed the (raw, unmitigated) dot tick size at 5% of the raw unmitigated melee swing damage for consistency. That puts it at about 15% of overall damage, which seems reasonable.

e: also, regarding the TMI range, I'm not planning on tweaking it again until the next expansion. But tentatively, I'm thinking of trying a log_x(sum(x^y)) setup so that it's more constrained. Basically the log_x() of the current result. I need to do some numerical testing with sample data sets to see exactly how well that will work and whether it's has a detrimental effect on stat weight generation.

#53 Sacri

Sacri

    Glass Joe

  • Members
  • 2 posts

Posted 09 October 2013 - 10:32 PM

I apologize for the random derail here. I returned back to WoW after a long period of time and I'm just settling into the new mechanics (this thread has been very very helpful!) but was curious if there is a favorite set of mods (or more helpfully a mod/addon compilation) that you might suggest as a starting point for a new prot paladin? Theres just a lot more 'stuff' going on it seems and was curious what a nice method is to monitor that visually.

#54 Nobleshield

Nobleshield

    Glass Joe

  • Members
  • 1 posts

Posted 12 October 2013 - 12:01 PM

A question on the beginning part: Do you think it is wise to use EF at 3 stacks at the start of the fight? I mean if you are at full health the ticks won't do anything, and EF without any BoG stacks is going to be a pretty weak tick. Wouldn't it be more beneficial to build up those BoG stacks and then use EF, and only try to refresh it with BoG stacks instead of just keeping it rolling as though it was Sacred Shield?

#55 Theck

Theck

    Don Flamenco

  • Members
  • 296 posts

Posted 12 October 2013 - 08:30 PM

It really depends on the boss, and whether you're tanking first or not. I tend to build up 5 stacks of BoG first if I'm not actively tanking on the pull, obviously. If I am tanking, it depends on whether the boss presents any real danger in the first 20 seconds; if not, then I'll probably wait for 5 stacks; otherwise I'll use it the first time I can effectively make use of the WoG.

#56 Wrathblood

Wrathblood

    Don Flamenco

  • • Guide Author
  • 351 posts

Posted 24 October 2013 - 09:36 PM

So, I've been advising prot pallies tha tyou always want to go for ilevel in your gear.  The thinking was that if you've got a 540 flex haste/mastery piece and a 553 parry/dodge piece drops, you want to grab it and use it, itemization be damned.  You can reforge some of that avoidance into Haste, plus you get additional STR/STA/Armor on top of the avoidance and my math back in 5.1 or whenever it was said you can out ahead with the ilevels.

 

This has been a surprisingly tricky thing to continue to validate as the stat change on a single piece of gear is pretty small, less than the variance on a single simcraft run, particularly with our increasing focus on spike damage vs overall damage taken.  Its been a nagging concern of mine, whether I was right about that.

 

Finally it occurred to me (like, literally today) that while swapping a single piece of gear would produce a below-margin-of-error result, I could fix that by swapping multiple pieces of gear at the same time.  So I picked 4 slots (legs, shoulders, hands and head) and put together pairs of gear, one at 540 with haste/mastery and one at 553 with dodge/parry (ok, to be honest I had to invent a pair of plate legs with dodge/parry but it wasn't very hard.  The rest of the slots i just used available SoO drops) all with the same number of gem sockets.  I took the 553 pieces and reforged them for Haste.  I then compared the results of the two different gear profiles via simcraft (T16N25).

 

It was actually pretty easy to do.  In a nutshell, it just comes down to stat differences.  At those 4 slots, 540 well-itemized gear vs 553 poorly-itemized gear comes down to comparing relative stat values.

 

The battle works out to:

 

Well-itemized gear

2036 Haste + 3816 Mastery

 

vs

 

Poorly-itemized gear

1242 Stamina + 908 STR + 714 Armor + 3119 Parry + 3877 Dodge

 

Turns out, the Well-itemized gear actually won.  I ended up running multiple sims for both gear sets and averaged them together because the margin of victory was pretty small, but it was clear.

 

The Poorly-itemized (but higher ilevel) set results in approximately a 4% decrease in overall damage taken, plus you have the extra ~1250 Stamina to help survivability, but the Well-itemized gear-set was sufficiently superior at damage smoothing to more than make up for it.  The Well-itemized set never had a higher TMI than the Poorly-itemized set (though the gap for both sets was over a pretty tight range), and the Well-itemized set also delivered 1.5-2% more tank dps which is a plus.

 

Edit - What the hell happened to my text?  I'm sure that's not annoying to read at all.

 

Edit2 - Ok, manually fixed.


Edited by Wrathblood, 25 October 2013 - 05:13 PM.

  • Akawa likes this

#57 MysticalOS

MysticalOS

    French Hacker

  • ♦ Administrators
  • 778 posts
  • LocationGeorgia

Posted 24 October 2013 - 10:12 PM

If you copy and paste text into this box, it inherits the rich text values of that text unless it's in raw text edit mode. Or maybe you just clicked a weird combination of style and font buttons :)

#58 Wrathblood

Wrathblood

    Don Flamenco

  • • Guide Author
  • 351 posts

Posted 25 October 2013 - 04:08 PM

If you copy and paste text into this box, it inherits the rich text values of that text unless it's in raw text edit mode. Or maybe you just clicked a weird combination of style and font buttons :)

Thanks, I'm just going to manually fix it in a little bit.



#59 Theck

Theck

    Don Flamenco

  • Members
  • 296 posts

Posted 26 October 2013 - 04:56 PM

Turns out, the Well-itemized gear actually won.  I ended up running multiple sims for both gear sets and averaged them together because the margin of victory was pretty small, but it was clear.

 

The Poorly-itemized (but higher ilevel) set results in approximately a 4% decrease in overall damage taken, plus you have the extra ~1250 Stamina to help survivability, but the Well-itemized gear-set was sufficiently superior at damage smoothing to more than make up for it.  The Well-itemized set never had a higher TMI than the Poorly-itemized set (though the gap for both sets was over a pretty tight range), and the Well-itemized set also delivered 1.5-2% more tank dps which is a plus.

 

I had been thinking about this as well, but hadn't had time to do any simulations on it.  A bit surprising as well, I had guessed the opposite result (that the stamina would win out).  Nice job.



#60 Daravon

Daravon

    Glass Joe

  • Members
  • 1 posts

Posted 22 November 2013 - 09:57 AM

Hi guys, didnt know where to post this or if to make a new thread or so (mostly used to read instead of posting).

 

But i would definately appreciate your help on the following. I am a prot paladin still gearing up and our guild is still busy with flex to gear up. I am noticing i am taking up a lot of damage on a boss fight, but also my dps is to cry on. Before i continue i want to say i have read the guides on icy veines, this forum, other places as well (e.g. Ask Mr Robot). Have tried different builds and spend more gold on changing things instead of enjoying tanking nowadays.

 

 

In a bossfight i keep up sacred shield, shield of the righteous as much as i can, using Light's hammer. so i believe i am doing what i should rotation wise and what this forum advises to do. Also when it concerns to gearing i think i have lost the tracks.

When it concerns to dps i pull up to around 40-60k dps and when i squeeze some CD's i might take it up to 70k, but got no aggro issues mostly.

 

This is my paladin tank:

http://eu.battle.net...nereth/advanced

 

Could someone please help me out?

If i posted incorrectly, my apologies...

 

kind regards,

Venereth


Edited by Daravon, 22 November 2013 - 09:59 AM.





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users