Theck's MATLAB thread - Cataclysm/4.x

Had nothing to do so wrote up the list for you. Sorted it out into four categories as it says at the very top. Hope this is what you were after.

Nothing short of awesome still, Theck. As always, thank you for providing hard numbers to back what we've been seeing and doing, especially with the big release less than 24 hours away.

Thanks for your hard work Theck, I'm sure it's much appreciated by the entire community.

I must assume I didn't manage to check all the 43 pages of the thread so please ignore if it was already requested.

First of all great work here. Besides being usefull for the game might be a good way for some to understand how usefull simulation tools can be and eventually learn matlab .

Besides the overall data for the "rotation" analysis, could be interesting to see TPS over time from a initial pull. Both as 80 and as 85 in beta, the only issues I could see on paladin aggro was the initial pull itself (yep those easy trigger happy dps, etc etc).

One good example is the use of Inquisition. I can see that overall doesn't make much diference but I empiricly seen some diference if I used it before multimob pulls (after using DP). Could be nice to have some burst TPS graphs of the first minute(s) of fights.

Call it the pull rotation against sustained rotation.

Thanks again for all the work put into this.

tlitp wrote:Presenting only the values that correspond to (veng=1.0, default=low hit/exp) is merely offering a skewed perspective.

Perhaps, but I didn't have time to modify calc_abilitydmg to give me pretty-print outputs at different vengeance levels. And looking at your data, the prio_sim results seem to hold quite well. #3 is still the highest single-target rotation (<10 DPS is not a significant difference for these rotations), and #20 seems to keep the title for execute range. That said, #20 pulls farther ahead of #21, so my comment about them being fairly equivalent might be out of place at low Veng.

I also think that it's far more likely we'll see high% Vengeance while tanking in Cataclysm than Wrath. Plummeting avoidance and the change in damage intake patterns should help keep Vengeance more consistent.

Either way, I'll try and find some time today to add your extra data to the first page. If you have the full output of those sims handy for the two extra cases you ran, I'll just use those.

Theck, first post ever here, and I'd like to thank you for your work on these simulations, providing the data and generally being awesome...

I have a query concerning Inquisition, and excuse me if my ability with mathematics and understanding the code are failing me, but I cannot interpret from the data what the uptime on Inquisition was in the simulations.

If I have worked it out correctly, and using a real life rotation rather than a mathematical priority, let me put to you the following scenario. I'm maintanking a cata raid boss, assuming best case scenario that I am hard hit capped and soft Exp capped, I use AS to pull, immediately follow it up with a Judgment for the buffs/debuffs, then pop AW and DPlea together and SotR the boss. This puts my Holy Shield up and a whole bunch of buffs and debuffs from the start. Then I start my regular rotation generating HoPow and subbing in consecrate by necessity since I used AS and Judge on the pull (CS>HW>CS>Cons>CS) - Now I have 3 HoPow again and 11 secs left on Holy Shield. Imagine I then use Inquisition (12 secs of 30% buff), and repeat the process, alternating each rotation with SotR and Inquisition, giving me approximately 66% uptime on Inquisition. Would this still be a tps loss? Should we be dropping Inquisition altogether? And if so, do you tihnk Blizz will buff it or something to make it more useful?

I sincerely apologize if this is a dumb question which has been answered already (I'm at the office and don't have time to read all 600+ posts), or I should post this somewhere else that better reflects my noobiness. Please go easy on me /sheepish

Humbly, Aash.

Edit: I guess my question can be condensed into; which rotation provides the most single target tps out of these two...

CS>AS>CS>Judge>CS>SotR>CS>HW>CS>Judge>CS>Inq...repeat ad infinitum (66% uptime on Inq if I'm not mistaken, and 1 3*SotR per rotation)

or

CS>AS>CS>Judge>CS>SotR>CS>HW>CS>Judge>CS>SotR...repeat ad infinitum (No Inquisition uptime at all, and 2 3*SotR per rotation)

That's exactly what the Inq>SotR>CS>J>AS>Cons>HW queue is calculating. As I said earlier, it isn't inherently obvious, but if you work out what spells you cast GCD by GCD, you get:

Inq-CS-J-CS-X-CS-SotR-CS-J-CS-X-CS-

The X's get filled with AS, Cons, or HW depending on availability, 66% Inq uptime. This is exactly the rotation you're looking for. Unfortunately (for your question, anyway), it's a net DPS loss compared to just using 2 SotR's. Earlier, preliminary data suggested this wasn't the case, but it was due to a bug in the Sacred Duty implementation.

pep wrote:Besides the overall data for the "rotation" analysis, could be interesting to see TPS over time from a initial pull. Both as 80 and as 85 in beta, the only issues I could see on paladin aggro was the initial pull itself (yep those easy trigger happy dps, etc etc).

This is something I'd be interested in calculating as well, but right now there are more pressing matters. Plus, the change to Divine Plea has made this more or less irrelevant.

theckhd wrote:That's exactly what the Inq>SotR>CS>J>AS>Cons>HW queue is calculating. As I said earlier, it isn't inherently obvious, but if you work out what spells you cast GCD by GCD, you get:

Inq-CS-J-CS-X-CS-SotR-CS-J-CS-X-CS-

The X's get filled with AS, Cons, or HW depending on availability, 66% Inq uptime. This is exactly the rotation you're looking for. Unfortunately (for your question, anyway), it's a net DPS loss compared to just using 2 SotR's. Earlier, preliminary data suggested this wasn't the case, but it was due to a bug in the Sacred Duty implementation.

Thanks for confirming.

I realise this next question is outside the scope of your thread, but... Do you think Blizz are likely to buff it? Or do you think that they have a different purpose in mind for Inquisition (AoE only)?

Also, is my understanding correct that the single target tps loss overall is *approximately* 5% with the Inquisition rotation?

Aashenfox wrote:I realise this next question is outside the scope of your thread, but... Do you think Blizz are likely to buff it? Or do you think that they have a different purpose in mind for Inquisition (AoE only)?

Also, is my understanding correct that the single target tps loss overall is *approximately* 5% with the Inquisition rotation?

Blizz has said in the past they want Inquisition to be our AoE finisher, and not something we use on single targets. So I don't expect them to buff it. If anything, the fact that it comes within 5% of our highest DPS rotation might be an indicator that Inq is too strong for Prot.

Ugh. While updating the rotation module, I was getting disagreement with the sims. Checking them carefully, I found that in fixing the SD bug I mentioned earlier, I introduced another one. I'm re-running the sims now and will be updating the data.

Just wondering, has the HotR glyph bug been unbugged yet? Since it's still in the guide it "will be a nice glyph IF the bug is fixed"?

Sorry if this was stated somehwere in other comments already.

In await of one's answer,

thanks in advance.

Amonet

On Theck's other guide it states it's already been fixed.

I haven't updated the Glyph comparison yet, but the HotR Glyph is indeed working properly in 4.0.3a.

Single-target Priority sims are updated, hopefully for the last time. Now we can all waste loads of time speculating whether it's worth using Inquisition again. /sigh

I'm hoping to get the talent and glyph updates applied in an hour or two here. AoE rotations may have to wait a bit, I want to clean up the list of queues and that will likely take some time.

The_Marsh: Thanks for the excel spreadsheet, though that's not quite what I meant to ask for. I was a little too vague with my request. I was hoping for a list of items sorted into three categories the way I did in my weapon sims. I.e.:

Tank, sorted by ilvl:
12345
13234
34223

DPS weapons, sorted by ilvl:
31415
66266
88343

Spellpower weapons, sorted by ilvl
32134
13241
14526

If anyone doesn't mind doing that, it would save me a lot of time.

You want them listed by their item id#, right? Also, do you want agi weapons included or only those with Str?

Thanks for the fast and clear reply.

Did not include agi weapons.

Sorted by spec. Also with notations to differentiate heroics and epics from their regular versions.

There are decidedly few pure tank weapons. The vast majority of one handers I saw had a combination of hit/expertise + mastery. I put them with the DPS group but marked them as threat/tank weapons.

IF you require the ratings, I'll get working on that right now.

I find the enthusiasm in here utterly disgusting. Yeah.


@Theck :

• Let's settle for veng=0.6 across the board (till t11 is actually relevant). Deal ?
• I've come to the same "conclusion" regarding HoW. (cs.j.cs.how.cs.sotr) is easily executed and pretty close to optimality. As such, I'll implement it in [RM].
• forget the fucking listing, I'll compile it till tomorrow. Hell, I may as well update the whole [CW].

uh, what!?

Updated spreadsheet -- included each and every single rating for each and every single weapon. Again, did not consider AGI weapons but if needed be, shouldn't be too hard to add them at this point.

tlitp wrote:@Theck :

• Let's settle for veng=0.6 across the board (till t11 is actually relevant). Deal ?
• I've come to the same "conclusion" regarding HoW. (cs.j.cs.how.cs.sotr) is easily executed and pretty close to optimality. As such, I'll implement it in [RM].
• forget the fucking listing, I'll compile it till tomorrow. Hell, I may as well update the whole [CW].

I don't think 0.6 is going to be representative. Even in Wrath content, there were fights where I sat at 95%+ vengeance for most of the fight, and other fights where I might fluctuate between 10% and 50%. I like your idea of showing multiple values, one for the high end and one for the low end, at least in cases where there's a significant variation.

Admittedly, just running at 0.6 would be a lot less work...

Give me another hour or so, I've done a massive overhaul to rotation_model and I'm about done with calc_talent.

If you want to update [CW] feel free, but to be honest it's pretty fast - all that needs to be done is to replace the old list of weapon item ids with the new list. Hence why I asked if someone else would mind sorting them for me, so I can copy/paste and be done. I don't need a zip file either, just listing them in a post is fine.

Klaudandus wrote:Updated spreadsheet -- included each and every single rating for each and every single weapon. Again, did not consider AGI weapons but if needed be, shouldn't be too hard to add them at this point.

You're going to hate me for this, but you didn't need to put all of the stats in there. The stats for all of the items should already be in the gear database. All I needed was the item ids, sorted into lists. Basically, your column B.

However, including agi weapons would probably be a good idea, as they'll be relatively effective tank weapons.

As I compiled the lists and filled the ratings manually, it helped me notice something. There are few crit weapons and fewer expertise weapons. In fact, there are no combination of Expertise + Avoidance stat weapon and only two ilvl 333 weapons that combine expertise and mastery. Mastery + Hit however is a popular combination.

Here it is, with AGI weapons included. I kept them separate from the STR weapons and did not include the combat ratings for them, that cut the time needed to update the list by a lot.

Talent comparison updated. Also, I noticed a bug that probably meant the Focused Shield and Consecration glyphs weren't properly being applied in the priority sims. I'll have to go back and re-run them to update the results.