Originally Posted by
Zoopercat
@
Bigbazz @
Hashtronaut - I'm from AMR. We spend a lot of time comparing our sims to actual logs to ensure they are accurate. I recently
posted the results pre-7.2 - where we were able to match rotations and damage very closely to logs. I know that a lot of things changed, but I wanted to illustrate that we are able to match logs really well (demonstrating accuracy). After this week's logs start getting uploaded, I'll be able to do more comparisons like this for 7.2. SimC can't be compare to actual logs like this, but you could compare it to AMR on the target dummy script to see how accurate it is.
We also take into account more things than SimC, which might explain some differences. We can simulate
krosus for every dps (and tank) spec, as well as
spellblade. We also have an advanced spell queuing system that matches the game's actual mechanics.
As Bigbazz said, a lot of things could be happening with your specific setup. It could also be rotation differences. If you think you found a bug however, let me know.
What would be interesting would be to be able to import an actual log, using actual player data for the rotations to use in generating stat weights, because that would allow you to more closely tailor gearing choices to personal fight strategies on specific fights.
That said, player data is almost always going to be inferior to a well made priority list executed flawlessly as the sims do, when looking at a generic situation. As I mentioned I've not used AMR so I can't have any educated opinion on how accurate it is, but one thing we are sure of is that for Arms specifically Mastery is incredibly powerful to the degree that people are sacrificing a lot of ilvl for it and still seeing a dps increase. Simcraft (which I do have a lot of experience using) has put out some really whacky results for Arms in Legion but one thing for sure is it's definitely very consistent with regards to Mastery being super strong, just as we see with in-game results.
- - - Updated - - -
@Zoopercat - I decided to give the sim a little quick look. AMR sim has a Krosus HC sim so I found the last kill I did, fight lasting 3min 14s and so I set the sim to that number, dropped the player lag/latency down from the default and set the spell queue to 100ms as I have it in game. The sim is based on my current gearset in 7.2, which is slightly superior to the gear I was using on the kill. https://www.warcraftlogs.com/reports...pe=damage-done
Code:
This is AMR simulation of my ilvl898 7.2 Arms gearset, which in Simulationcraft currently sims at roughly 830k dps.
This is a screencap of the actions page for the log, where I did 790k dps on Krosus for the same duration, at ilvl895.
Mortal Strike usage
ARM = 44 uses, 40.6mil damage
Log = 35 uses, 40.6mil damage
Draught of Souls
AMR = 2.7 uses? 23.8mil damage
Log = 3 uses, 29.98mil damage
Execute
AMR = 19.5 uses? 16.6mil damage
Log = 19 uses, 20.6mil damage
Slam
AMR = 45.9 uses? 12.2mil damage
Log = 42 uses, 12.14mil damage
Corrupted Blood
AMR = 7.9mil damage
Log = 9.65mil damage.
So yes we are close in some aspects but we are way off in other aspects, and the simulated numbers are a long way below the kind of dps being achieved by my character, my log is 120k dps ahead of the sim despite being in inferior gear, where as Simulationcraft is giving me closer to what my log is showing.
Edit : It appears I forgot to turn on Old War potions. Though I only used 1 in the log, it's quite significant.
These are my results with Old War, overall dps is closer but the big differences in overall spell damage still remain.