Simcraft Thread - Legion/7.x

Warning: Theorycraft inside.

Moderators: Fridmarr, Worldie, Aergis, theckhd

Re: Simcraft Thread - Legion/7.x

Postby jere » Sun Aug 28, 2016 12:07 pm

It doesn't build on GCC for windows (mingw or tdm-gcc)?
Image
User avatar
jere
 
Posts: 2412
Joined: Fri May 11, 2007 5:12 pm

Re: Simcraft Thread - Legion/7.x

Postby Baeldin » Sun Aug 28, 2016 6:02 pm

I have no experience in building things on Windows, so I just followed the build instructions in the Wiki. Could I build the GUI with those too?

That aside, I reran my sims with some bugs fixed, but not everything seems to be working yet. I still have the following problems:

1. Stamina and DPS increase even after 34 traits, but since Unrelenting Light is always the same at the end of any progression path I simply took it out of my sims.
2. The first point (Truthguard's Light) gives an insane DPS and survivability boost which seems way out of place. These first two points are irrelevant for this comparison because we pick them up almost instantly and they are the same for any progression path anyway. However, it might still be worth checking whether these two points are too low or the other 33 points are too high in DPS and survivability.
3. Hammer Time does not work in my sims (more below)
4. Painful Truths is too strong (more below)
5. Bulwar of Truth is still at 20%
6. Light of the Protector is still at 25%

New results for aftifact progression. SimulationCraft 703-01 build 22423, 10000 iterations
Talents: Bastion of Light, Blessed Hammer, Blinding Light, Blessing of Spellwarding, Hand of the Protector, Judgment of Light, Righteous Protector

Before we continue, a few words on the chart. Generally, the higher the number the higher the increase/decrease. This means that the same trait gives more DPS but less TMI the later we pick it up. This makes DPS-traits look stronger and defensive traits look weaker the farther we progress. Keep this in mind when evaluating something like Bastion of Truth, which is a very good defensive trait even though it lowers the TMI only moderately by the time we pick it up.
This is what things look like for the recommended path:
Image
Note: The error bars show the DPS-Range as given by SimCraft and the min and max TMI, representing the maximum variability seen, including outliers. The error is around 0.1% for DPS and 0.03% for TMI and would not be visible because the error bars would remain inside the dots. I chose to add the entire variation to illustrate how little one artifact trait contributs compared to fight-by-fight variability of even a static fight and a perfectly identical rotation.

These sims confirm the best DPS-Traits: Consecration in Flame, Tyr's Enforcer, Stern Judgment and Painul Truths. The best TMI-Traits in this sim are Faith's Armor, Resolve of Truth, Scatter the Shadows and Bastion of Truth, with Bulwark and Light of the Titans being ahead of all minor traits. However, as we move through Hammer Time it seems to do nothing at all except for the basic damage and stamina increase. Here are the values:
Code: Select all
rank   BH hit   crit   avg   Delta_h   Delta_c   Delta_avg
HT 0   20962   42752   23176         
HT 0   21059   42977   23289    97,00   225,00   113,00
------------------------------------------------------
HT 1   21155   43165   23393    96,00   188,00   104,00
HT 2   21262   43350   23510   107,00   185,00   117,00
HT 3   21352   43533   23605    90,00   183,00    95,00
------------------------------------------------------
HT 3   21453   43745   23727   101,00   212,00   122,00
HT 3   21561   43971   23837   108,00   226,00   110,00

These are the average hits by Blessed Hammer for each of the sims just before, during and after HT is raised. Each rank of HT should increase its damage by 5%, but as we can see here, this is clearly not happening. The effect of the trait should add around 1000 damage to each hit on top of what we see here. Interestingly, the rise in damage seems lower while Hammer Time is increased compared to what it is for the lines where HT is not increased. For normal hits it looks like random variation, but for crits, something seems odd, as the rise of crit damage seems consistently lower by about 10%. I have yet to test this using a different talent setup to see if it happens when BH is not chosen, but at least with BH active, HT is broken in 22423. It alsmost looks as though Hammer Time lowers BH damage very slightly but the effect is offset by the passive damage increase.
//edit: Switching from Blessed Hammer to Holy Shield removes the problem and Hammer Time appears to work normally, increasing HotR damage by more than 1000 damage for each rank.

Now let's take a closer look at Painful Truths. The sims are telling us that it's by far the strongest DPS trait of the entire artifact. Seeing how it works, I understand that it can be quite powerful, but in a single target boss fight it should not perfom this well. Here are the numbers:
Code: Select all
AD uses   PT hits   hits/use   boss_hits/AD
4,0       41,8      10,45      5,33

So we have the number of times Ardent Defender is used on average per fight, the number of total hits of Painful Truths on average, the number of average hits per use of AD and the average number of boss melee hits we would expect during an 8 second window (the duration of AD). Painful Truths reflects holy damage on melee attacks and there is no way that our 1.5s main hand attack hits more than 10x within 8 seconds. The boss also has a nuke spell every 24 seconds (would not explain the discrepancy) and a dot ticking once every second (might well explain it). So it appears that PT reflects damage on more events than it should. Realistically, it should be a bit less than half as strong as it is in these sims. Assuming that we have the luxury of using it as an offensive CD, we can get out a lot more damage if we time it with our offensive CDs. If not, it is still a good DPS trait, but not as spectacular as the above chart makes it look like.

Just for comparison, here's a very offensive path through the artifact:
Image
So this is basically the same thing as above up to Tyr's Enforcer, then we pick up Painful Truths and Stern Judgment, the next best DPS traits that are just nearby. After that, we go for Bulwark and then straight for Light of the Titans. After that, we get the rest. It's kind of an experimental setup and it might not make a lot of sense, considering that this doesn't pick up a lot of defense until the raids are released. If you just care about damage and don't plan to do raiding or mythic+, this might still be the way to go.

--------------------------------------------------------------------------------
Results for the recommended path:
https://drive.google.com/open?id=0BzdcW ... 0FGMHQ4azQ
Results for the offensive path:
https://drive.google.com/open?id=0BzdcW ... FZibDJyU3M
Numbers from the results in a more condensed form (also the basis of the above charts):
https://drive.google.com/open?id=128u3g ... GPZSoRJtv0
Baeldin
 
Posts: 7
Joined: Mon Aug 22, 2016 5:52 am

Re: Simcraft Thread - Legion/7.x

Postby Dion » Sat Sep 24, 2016 1:53 am

Pawn weights ( Pawn: v1: "Comb_Pre_raid": Strength=7.4, Stamina=7.94, CritRating=6.9915, HasteRating=8.144, MasteryRating=7.566, Versatility=8.2235, Armor=19.608, Dps=1.7905)

I think this was posted on MMO-Champion forums and was celinamunas stat weights. Now, I would like to compare them a little. Crit is 85% as good as Versatility. But 15% difference is significant, right?

No it isn't. There are few reasons for this. First, it's not 15% difference. Let's take versatility (8,2235) and crit (6,9915) and compare them to armor (19,608). Versatility is 41,9% as effective as armor and crit is 35,6% as effective armor. 41,9% - 35,6% = 6,3%. In that 6,3% difference there is littered every single stat besides armor.
Secondly as most gear has two primary stats, armory and two secondarys. We have no control over two primary stats nor armory but can try to get best secondarys. Here is example, Crown of Silver Hand (810ilevel head)
810 head:
508 armor (9980,5 gear points)
894 strenght (6615,6 gear points)
1340 stamina (10639,6 gear points)
659 1.secondary (5419,3 gear points)
466 2.secondary (3795,1 gear points)
(Total best secondaries 9214,4 gear points)

So we notice that even with best possible secondaries, armor beats them. But not only that, stamina beats them by large margin. Add strenght there too and value of secondaries gets even more lower. But lets take a look if that piece had worst possible secondaries:

659x6,9915+466x7,566=8133,2

Total points best secondaries: 36450,1
Total points worst secondaries: 35368,9
So difference of best secondaries vs worst secondaries is 1081,2 gear points meaning that worst secondaries is 97% as effective gear piece as with best secondaries.

Now, it fast becomes obvious that when there are armor and two primary stats, what secondaries happen to be in gear matter extremely little. So what happens if strenght and armor are not around?

Glinting Quartz Ring 810 ring:
754 stamina
678 1.secondary
904 2.secondary

So with best secondarys 754x7,94 + 678x8,144 + 904x8,2235 =
18​942,436

And worst secondaries 754x7,94 + 678x7,566 + 904x6,9915 = 17​436,824

Difference is still small, ring with worst secondaries is still 92% as effective as best secondaries.

But what if there happens to be only single secondary? Holy grail of (currently possible) secondary stat min-max?
So 754x7,94 + 1582x8,2235 =18​996,337
And 754x7,94 + 1582x6,9915=17​047,313
Ring with worst secondary is 89,7% effective as ring with best secondary. 10,3% difference isn't earth shaking but it's significant. But there is twist. As player character has only three possible slot of accesories compared to nine slots of armor that has two primary stats, armor and two secondary stats. So lets compare best ring vs best helm.

Helm:
508 armor (9980,5 gear points)
894 strenght (6615,6 gear points)
1340 stamina (10639,6 gear points)
1125 secondary stat (9​251,4)
Helm total=36​487,1

And best secondary ring:
754x7,94 + 1582x8,2235 =18​996,337

This means that our best itemized holy grail ring is 52% of value from best secondary helm. To put this perspective, we need two rings to compete with one helm.

And I didn't take account that it's not possible to have best secondaries on all pieces simply because it's shared with other classes and we want tier armor because of set bonuses regardless of secondaries.

So when +5ilevel loot drops, it's best to equip it if it's armor piece instead of accesory piece. Because secondaries are so close each other and strenght, stamina and armor are so good.
Dion
 
Posts: 255
Joined: Tue Jul 28, 2009 11:23 pm

Previous

Return to Advanced Theorycraft and Calculations

Who is online

Users browsing this forum: No registered users and 1 guest

cron

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