02/14/09, 11:49 PM
|
#1
|
|
Vula'jin the Void, blessed by the loa
|
Warning for Mavanas: 2. All opinions should be stated as succinctly as possible.
Post: Free Hat: Deep Subtlety and You
User: Mavanas
Infraction: 2. All opinions should be stated as succinctly as possible.
Points: 0
Administrative Note:
Message to User:
|
Please don't post back-to-back.
|
Original Post:

I can finally share a beta version of a simulation spreadsheet that I have been using to test various hypotheses regarding rogue HaT dps. Even though it is similar to two simulations I have seen on these forums, to the best of my knowldge, it is unique in modeling the stochastic processes that govern combo point generation through HaT, while giving the user control over the group setup.
The spreadsheet simulates a combat of a given length by choosing at any given point which next ability to use out of a set of predefined abilities, such as hemo, backstab, sinister strike, eviscerate, rupture, etc. The user has an ability to choose the rules defined around energy and combo points that determine the order in which these moves occur. Autoattacks occur automatically at intervals dependent on haste at the time. All hit table rolls as well as poison application rolls are implemented using Bernoulli random variables. The combat is simulated several times and average values are calculated for each setup. The spreadsheet is equiped with several easy to use macros and statistics that allow for comparison of dps and calculation of EP values while changing rogue stats, buffs, talents, dps rules, and many other variables that affect HaT dps.
I have uploaded the file to filefront at Rogue_Simulation_3.0.8a.xls Download File on FileFront. Do not forget to allow macros when you open the excel file since some of the code was written in VB for Excel. First time you use the file, I strongly recommend reading "How to Use" tab to make sure you understand how to make best use of the file. The Inputs tab is your control center from which you can run the macros to calculate dps. Other tabs allow you to adjust talents, buffs, etc.
In a separate post below I will list the major findings from using the spreadsheet, which can answer some of the questions that have been asked above. Keep in mind that the spreadsheet is still in beta version and may contain errors. Please do not hesitate to let me know if you find any bugs.
I have extensively examined HaT dps using the simulation file above, which is loaded with a preset group that I personally ran in Naxx before. The group consists of a survival hunter, a combat rogue, and two fury warriors. WWS report was used to determine crit/sec rates and frequency of special attacks. I am aware that there definitely are better setups for HaT, and I have given the user of the file a choice of his group, so you can put in the appropriate numbers that describe your group best.
Based on the group setup, I found the following regarding HaT dps:
1. Highest DPS (while allowing for shadowstep for utility, which is common among HaT users) is attained using Talent Calculator - World of Warcraft (13/14/44). The DPS with that spec was around 5,950 on average. Note that I tested it in best dps gear currently available (4 piece t7.25 + Malygos chest, DMC:Greatness, FotFF, quest neck, thrusting bands, stalk skin belt, KT cloak, surge needle ring, strong-handed ring, and footwraps of vile deceit.)
2. At this rate of combo point generation (around 1.3 per second not counting HaT procs from own abilities), rupture is definitely worth using in your rotation.
3. Hemo is used quite extensively (about 8% of your damage) to burn energy. Testing showed that hemo was best used only at 0-2 combo ppints when energy was above 60.
4. Around 30% of the damage comes from eviscerate that is best used above 90 energy and at 4-5 combo points.
5. Rupture provides about 10% of the damage, and is best used at 5 combo points.
6. SND (glyphed and talented) is best used at 3 combo points.
7. Substituting double WB and switching to backstabbing did not improve the DPS I already achieved using Calamity's and WB offhand and using hemo. It is possible that higher combo point generation can change these results, but as it stands, the gap between backstab build and hemo build was of the order of 200 dps. I tried different talent specs, including some that involved aggression and puncturing wounds in search of higher dps with daggers, but I could not improve upon my hemo HaT dps.
8. Using WP/WP is superior to DP/WP even for Calamity's in MH (note that I already implemented the new 40% SND in the latest version). It's definitely the best choice for dagger HaT.
9. Sinister strike HaT, even with glyphs and appropriately changed talents did not produce good results.
10. EP values (following pocket guide format) were as follows: ag 2, haste 1.2, apen 1.5, crit 1.3, AP 2.0, expertise 2.1, hit 2.3 below yellow cap, 1.7 below poison cap, and 1.1 above poison cap. Note that precision of EP calculation suffers most from the variability that's inherent in this type of simulation.
I will continue my analysis as well as testing the spreadsheet for robustness, and will share the results.
|
|
|
|
|
|