# MoP Mitigation Stat Weights – Monte Carlo – b16048

The last time we discussed stat weights, I promised to update the Monte-Carlo simulation and test it against the analytical calculations to verify their validity.  I’ve finally gotten around to that, and the results are about as I expected.  There have been no changes that affect these calculations since the posting of the analytical calculation, so both it and these results can be considered current for build 16048.

As usual, the code is available through the matlabadin project.  I’ve changed the file names slightly to be more descriptive than just “montecarlo”:

For information on how the Monte-Carlo works, see the previous posts on this subject.  In short, it tracks the boss swing timer, the global cooldown, the cooldowns of CS, J, and SotR, and the buff uptimes for SotR, BoG, Divine Purpose, and Grand Crusader.  It runs in 10-millisecond timesteps and tracks whether each boss attack is a hit, miss, or block and records the damage (including SotR mitigation).

The cast logic is such that it will prioritize CS>J>AS(only if GC available).  Technically it won’t cast J if there’s <0.2 seconds left on the CS cooldown, but now that everything scales equally with haste this conditional should have no effect.

We run the simulation for a total of 100,000 minutes of combat, divided into 10 “encounters” of 10,000 minutes each.  Since the paladin version of the simulation requires small times steps, this is very slow – about 3-5 minutes per encounter (unlike the warrior version(link), which is 5-6 seconds per encounter).  To determine stat weights, we calculate total damage intake at the base stat values and then again with 1500 points of each stat added (“Remember that thing I said was very slow? Great, now go do that 6 more times!”), and compare the results to determine TDR stat weights.

Our input values are based on a Mists of Pandaria LFR gear set (ilvl 483), with the following relevant stats:

buffed STR: 9208
parry rating: 4834
dodge rating: 4892
mastery rating: 6758
hit rating: 900
exp rating: 1777
haste rating: 621

The boss swing timer is assumed to be 1.5 seconds.  I’ve artificially shifted hit rating into haste rating on the gear set so that adding 1500 points of hit won’t over-cap us.

Simulation Results

The table below summarizes the statistical information about the simulation.  The first three rows represent the mean mitigation scale factor, the standard deviation of that scale factor over all 10 runs, and standard deviation of the mean (basically a measure of the repeatability of the mean value if we ran the simulation again).  The bottom three rows are the same data, but normalized by the mastery value (i.e. 0.0091).  This doesn’t change the results any, it just makes comparisons easier.

N=10, tau=10000, stat=1500

|          |  dodge |  parry |    hit |    exp |  haste | mastery |
|     mean | 0.0067 | 0.0073 | 0.0064 | 0.0043 | 0.0027 |  0.0091 |
|      std | 0.0005 | 0.0005 | 0.0006 | 0.0004 | 0.0004 |  0.0006 |
| std_mean | 0.0002 | 0.0001 | 0.0002 | 0.0001 | 0.0001 |  0.0002 |
|          |        |        |        |        |        |         |
|     mean | 0.7415 | 0.8067 | 0.7011 | 0.4776 | 0.3025 |  1.0000 |
|      std | 0.0585 | 0.0512 | 0.0623 | 0.0483 | 0.0422 |  0.0712 |
| std_mean | 0.0185 | 0.0162 | 0.0197 | 0.0153 | 0.0134 |  0.0225 |

Mastery is a bit ahead of dodge and parry, while hit is not far behind dodge.  Expertise trails further, as expected, and haste trails by even more.  Mastery pulling ahead is partly a function of how the simulation handles SotR.  In earlier versions, I was lumping it in with the GCD-based casts.  Then it occurred to me that since it’s off-GCD, it would make more sense to attempt to cast it any time an event occurs (either a boss swing or the GCD).  So I moved it out of the GCD section and put it at the front of the loop.  Basically, any time anything should happen, we first check to see if we should cast SotR, and if we can, we do so.

This introduces a bit of an anomaly.  Essentially, we know exactly when every boss attack is coming, and we try to make sure we cast SotR immediately before it happens.  That gives us an unnaturally high effective uptime on SotR, because we’re using it preferentially.  If I put it at the back of the cast sequence, such that the code handles all of our GCD-based abilities and boss swings before checking if we cast SotR, I get the following results:

N=10, tau=10000, stat=1500

|          |  dodge |  parry |    hit |    exp |  haste | mastery |
|     mean | 0.0078 | 0.0081 | 0.0043 | 0.0027 | 0.0031 |  0.0071 |
|      std | 0.0005 | 0.0004 | 0.0002 | 0.0004 | 0.0005 |  0.0002 |
| std_mean | 0.0002 | 0.0001 | 0.0001 | 0.0001 | 0.0001 |  0.0001 |
|          |        |        |        |        |        |         |
|     mean | 1.0897 | 1.1391 | 0.6100 | 0.3758 | 0.4411 |  1.0000 |
|      std | 0.0729 | 0.0498 | 0.0333 | 0.0618 | 0.0654 |  0.0265 |
| std_mean | 0.0231 | 0.0158 | 0.0105 | 0.0195 | 0.0207 |  0.0084 |

As you can see, dodge and parry fare better here.  The first set gives us a better feel for what a careful tank who’s anticipating boss attacks will experience, while the second set is a better representation of sloppy play.  The stochastic average (which is what the spreadsheet calculates) should be somewhere closer to the latter.  And, if we compare all three on on table, we find that this is pretty much the case:

stat         SS     MC 1    MC 2
parry:     1.1418  0.8067  1.1391
dodge:     1.0931  0.7415  1.0897
strength:  1.1155
mastery:   1.0000  1.0000  1.0000
hit:       0.6866  0.7011  0.6100
exp:       0.4413  0.4776  0.3758
haste:     0.4810  0.3025  0.4411

Most of the stat weights line up pretty well, at least within a reasonable margin of error.  The major outlier is haste, which seems to come out consistently low in these Monte-Carlo simulations.  I’m not sure exactly why, either – so far I’ve looked in all of the places I could think of that might cause that sort of depression in value, and couldn’t find a bug in any of them.

My guess at this point is that it’s a discretization error.  Because of the way we’re using SotR extra haste is less effective at creating overlap between SotR and boss swings.  If we were casting SotR pseudo-randomly within time slots, haste should make more difference, because it would increase the number of times that SotR and the boss swings overlap.  It’s still doing that to some extent, of course, but it’s less effective when we limit SotR casts to be only on the GCD or coincident with a boss swing.  It’s an open question which model is more accurate.

In any event, we now know that the Monte-Carlo and the Spreadsheet are in pretty good agreement.  We’re not getting results that we can’t substantiate, like haste or expertise topping dodge and parry.  That gives us a level of confidence in both tools, and leads us to our next line of inquiry: what about spike damage?  The entire debate over whether to use avoidance gearing or hit/exp/mastery/haste gearing  hinges on whether you value TDR or “damage smoothing.”  In the next blog post on this topic, I’ll use the Monte-Carlo to try and answer these questions.

This entry was posted in Tanking, Theck's Pounding Headaches, Theorycrafting and tagged , , , , , , , , , , , . Bookmark the permalink.

### 19 Responses to MoP Mitigation Stat Weights – Monte Carlo – b16048

1. Wrathblood says:

Fantastic post, thank you for doing all that work and also catching how ShoR’s sequencing was influencing the weights. Nails down a bunch of building-block things at 90 I’d been wondering about. I’m very curious to see how you go about addressing the value of damage smoothing, especially if you’re able to come up with a value for the “certainty” of ShoR usage a Control-based gearing strategy can give you, over and above uptime.

• Theck says:

That’s the plan. I would have had it done already, but these pesky Warrior sims are taking a lot longer than I expected. I’m hoping to carve out some time Friday or this weekend to work on the damage smoothing bit (and once the Warrior sims are working to the warrior community’s satisfaction, I can do the same for them).

2. Nooska says:

Looking forward to the next part – a request that may or may not be feasible – could you look at how well or how badly crit does for us in regards to dps value when looking at the HEM(h) gearing – mainly I’m curious whether we need 2 sets at all if going that way, and having a ret offspec.

• Theck says:

It’s feasible, but not with these sims. The usual MATLAB DPS code will address that much better.

3. Amorisse says:

In MC2, dodge is weighted higher than parry. Is that just due to the stats that you gave the model paladin? It was my understanding that dodge suffers more severely from DR.

• Amorisse says:

Edit: looking at your MC2 results and then the side-by-side comparison, its seems that you mixed up the dodge and parry labels on one or the other.

• Theck says:

You’re right, I copied them down in the wrong order. Fixing now.

4. Hightin says:

Im confused at how hit and expertise are pulling different stat weights below the caps. Unless Im understanding totally wrong you get the same % chance on the stat hit to miss and expertise to dodge reductions per point I thought. So the 2550(?) rating in either stat would net 7.5% chance to hit or dodge ignore. Again I could be totally wrong but given that shouldn’t they be giving the same state weight as they basically provide the same benifit?

Also that I just thought of but doens’t expertise actually add to spell hit now too? I dont know how much of our abilities are on the spell table but if any I would think it would pull ahead and not consistently behind.

• Hightin says:

Ignore the last bit after thinking it through it would still come out as even cause hit also provides spell hit. Took me a second for the gears to work.

• Hightin says:

So, after a few hours of thinking about the subject I decided to just make a large post over on maintankadin. I feel like thats a better place for me to get my point across.

• Theck says:

Hit and expertise aren’t identical anymore because of Judgment. When Judgment was a “spell” type ability, it was equally affected by hit and expertise, and hit and expertise were mathematically identical for survival.

However, Judgment was reverted back to a “melee” type ability that can’t be dodged or parried. As a result, expertise does *nothing* for Judgment anymore. And since Judgment is a significant portion of our holy power generation, that makes expertise less effective than hit.

5. ihaz says:

Why 1.5sec swing timer?

• Theck says:

That’s the default boss swing timer going forward in MoP. SotR was chosen to be 3 seconds for precisely that reason – so you don’t have to try and time it to get an extra swing’s worth of mitigation.

6. Benebarba says:

So these weightings and the analytic work from earlier seem to indicate that all values, haste actually beats out expertise for TDR? Is this weighting an artifact of the ‘steady-state’ assumptions (hundreds of thousands of actions instead hundreds)?

In short: Given the usual caveats regarding TDR, is expertise indeed a dump stat behind haste now, and so even 7.5% may not be needed? I’m curious, because intuitively it seems like expertise would greatly help holy power generation through CS and HoTR and thus fuel SoTR’s and WoGs. But I’m not really familiar enough with haste to know how much makes a big difference in how often you can use abilities over a typical encounter length.

• Theck says:

This comes down to a control vs. TDR argument. Expertise fares poorly for TDR because it only affects CS, and thus has no effect on 36% of our HPG (Judgment). That’s why haste beats it out – getting all of our HP generators off faster nets a larger HPG increase than getting one of those generators to hit more often.

That said, you should know by now that I don’t put a lot of stock in pure TDR assessment. Despite the fact that dodge and parry seem to “win,” I favor the hit/exp/haste/mastery “control” setup. I should have a blog post ready for next week that helps quantify that thought by looking at damage smoothing.

• Benebarba says:

That’s kind of what I figured, somehow I think I’d actually mis-rememebered the previous posts on stat weights and that’s why I was kind of surprised by that result.

It will be interesting to see the comparison of TDR vs damage smoothing and stat weights.

7. Pingback: Keep Calm and Parry On | Sacred Duty