[Enhancement] BiS gear and set-up discussion - Page 12

Vespasian · 09/01/09, 7:32 AM

I, like everyone else, do not open with LL when using my Electrifying Wind totem. Since this totem and the accompanying playstyle is the standard at the moment, any deviation would consistute a change from the status quo.

Ryethe · 09/01/09, 9:07 AM

Originally Posted by Vespasian

Updated

2) I took the work done by Mjoed and applied it to the basic parameters used for this thread, namely Tauren racials only. I was very surprised to see that the axes sim higher than the cudgels. It was my assumption, based on general anecdotal experience as well as consistent EP value calculations, that crit equates to more DPS when past the hit cap. Obviously this is not the case. I can only assume that hit does not scale in a linear fashion, and at this level provides slightly superior DPS vs. crit (axes provide roughly 12 DPS more than the cudgels). I am thinking this might have something to do with the increased ArP we have with this gear set, but did not test it in detail.

I did some playing around and found that Hit and Crit provide almost identical DPS until you hit about 550 Hit, when crit becomes better or when you fall below 420 Hit, where Hit becomes better. These are pure DPS benefit values, not EP values since any EP calculation I did consistently said that Crit was about .20 higher than Hit. I think this needs a lot more analysis as tbh, I cannot yet properly interpret the results.

TLDR: Hit is identical to, or better than, Crit at this amount of Hit in terms of DPS gained. This makes the Axes better than the Cudgels for all races.

What methods did you use for testing crit and hit? What step size did you use when increasing them? Or are you just talking about swapping weapons and reimporting RAWR XML files?

I find it interesting that the EP calculation doesn't bear out the same results as the regular simming. Since the sim is just running itself 14 times, there should be no justifiable reason for the EP calc to not jive with the the regular simmed results (never mind that EP is calculated from DPS).

Did you try setting the EP range for crit and hit higher to see the long term benefits of those stats?

Vespasian · 09/01/09, 10:47 AM

Originally Posted by Ryethe

What methods did you use for testing crit and hit? What step size did you use when increasing them? Or are you just talking about swapping weapons and reimporting RAWR XML files?

I find it interesting that the EP calculation doesn't bear out the same results as the regular simming. Since the sim is just running itself 14 times, there should be no justifiable reason for the EP calc to not jive with the the regular simmed results (never mind that EP is calculated from DPS).

Did you try setting the EP range for crit and hit higher to see the long term benefits of those stats?

I used increments of 10 Hit/10 crit either way to arrive at my results. I did not alter the EP ranges because, tbh, I have never been interested in EP values; only in actual DPS benefit.

I too found the results interesting, but the divergent EP/DPS result is not unheard of. It happens quite frequently in the sim, especially in the case of haste where higher EP values for haste can often result in DPS losses when applied in even the smallest of increments.

As I said, I cannot properly interpret these results and I think more work needs to be done in trying to understand them. Perhaps the sim is failing or perhaps there is something else that is elevating the value of hit above what we are otherwise use to.

Falcon213 · 09/01/09, 11:16 AM

Originally Posted by Vespasian

As I said, I cannot properly interpret these results and I think more work needs to be done in trying to understand them. Perhaps the sim is failing or perhaps there is something else that is elevating the value of hit above what we are otherwise use to.

I think much of this can be attributed to the loss of the T8 4pc bonus. Less LBs means less GCD-locking, so ironically building up those MW stacks faster becomes better.

Hit has always simmed well for me, so your results aren't too surprising (hit diminishes much faster than crit, which is probably the cause for your 550 rating cut-off). I also think hit is the key to understanding the true value of haste better, as they are similar in function for us.

Nobias · 09/01/09, 11:28 AM

Originally Posted by Vespasian

I too found the results interesting, but the divergent EP/DPS result is not unheard of. It happens quite frequently in the sim, especially in the case of haste where higher EP values for haste can often result in DPS losses when applied in even the smallest of increments.

Throughout this entire patch, I've tried to find the point where I should start stacking haste instead of AP, but still haven't. My Haste EP has been higher that AP for a very long time, suggesting I should stack haste, but it's always resulted in an actual DPS loss. So these inital results arent really suprising for me.

I'm wondering if the extra hit (more successful white attacks) will help off-set the t8.5 4-piece bonus loss, seems like it should.

Ryethe · 09/01/09, 12:15 PM

Originally Posted by Nobias

Throughout this entire patch, I've tried to find the point where I should start stacking haste instead of AP, but still haven't. My Haste EP has been higher that AP for a very long time, suggesting I should stack haste, but it's always resulted in an actual DPS loss. So these inital results arent really suprising for me.

I'm wondering if the extra hit (more successful white attacks) will help off-set the t8.5 4-piece bonus loss, seems like it should.

Haste is a sustained stat, meaning the more you are on the boss the more useful it is. It has absolutely no value if you're constantly moving around or regularly having to get off and on the boss (jousting). A simplified way of looking at it is that haste is useful over an interval of time whereas Crit/Hit/AP/ArP are instantaneously useful. This is not 100% accurate, but it is useful as a rule of thumb.

Vespasian: I'm sure you've noticed that you can see the actual DPS to the right of the EP value. There should be no difference between setting an EP range of 20 crit rating and adding 20 crit rating to your character sheet and manually running the sim. Silly question: Were you adding crit/hit % to both spell and melee crit/hit? I'm going to try to reproduce your results today using Mjoed's RAWR XML.

clliche · 09/01/09, 1:51 PM

Originally Posted by Vespasian

Updated

TLDR: Hit is identical to, or better than, Crit at this amount of Hit in terms of DPS gained. This makes the Axes better than the Cudgels for all races.

Hit and crit are both log relationships in terms of dps benefit. More hit means more flametongue, windfury, maelstrom and static shock procs while crit is basically just more damage and increased flurry uptime (which leads to more flametongue, windfury, maelstrom weapon, and static shock procs). I think the question that has to be answered here is at what point does increased flurry uptime give more benefit than hit? I would agree that armor pen is making hit more valuable but I dont think hit and crit can be considered the same dps upgrade. I would guess that after you reach a point where flurry is almost always up, getting more white hit becomes better.

Cochice · 09/01/09, 1:54 PM

Originally Posted by Vespasian

I, like everyone else, do not open with LL when using my Electrifying Wind totem. Since this totem and the accompanying playstyle is the standard at the moment, any deviation would consistute a change from the status quo.

Which is why I specified "until recently." Because everyone should have been using the gladiator's totem. Before Electrifying Wind was available, the priority wasn't LL first here, so why would it change now that we will have another LL totem.

Ryethe · 09/01/09, 2:22 PM

Originally Posted by Cochice

Which is why I specified "until recently." Because everyone should have been using the gladiator's totem. Before Electrifying Wind was available, the priority wasn't LL first here, so why would it change now that we will have another LL totem.

Because this totem grants more static AP? As we have learned countless time in the past, different items yield different results for priority, glyphing and even gemming. Throw your assumptions out the window and sim it.

Rouncer · 09/01/09, 2:45 PM

Originally Posted by Ryethe

Because this totem grants more static AP? As we have learned countless time in the past, different items yield different results for priority, glyphing and even gemming. Throw your assumptions out the window and sim it.

Problem is that there is no option to utilize a specific attack first but then to put that attack at a low priority. It's just one of those things, similar to Flame Shocking while running in to melee range, that people can do to maximize their performance but that can't be properly accounted for with the Sim.

Ryethe · 09/01/09, 8:08 PM

Originally Posted by Vespasian

I used increments of 10 Hit/10 crit either way to arrive at my results. I did not alter the EP ranges because, tbh, I have never been interested in EP values; only in actual DPS benefit.

I too found the results interesting, but the divergent EP/DPS result is not unheard of. It happens quite frequently in the sim, especially in the case of haste where higher EP values for haste can often result in DPS losses when applied in even the smallest of increments.

As I said, I cannot properly interpret these results and I think more work needs to be done in trying to understand them. Perhaps the sim is failing or perhaps there is something else that is elevating the value of hit above what we are otherwise use to.

I did some looking into this and here are my results:

First order of business:
Is there a quirk in EP simming that does not bear out in normal simming?

Testing Method:
1. I simmed and verified I was getting the same results as Mjoed with his config (accurate to within 3-4DPS).
2. In order to test the addition of hit and crit I activated the BS socket on the gloves in RAWR and socketed a 20hit/crit gem (this is pure stat addition and avoids possible rounding errors, human errors and talent interaction errors that can appear when modifying the EnhSim config directly).
3. I then did 3x 20k hour sims for hit and 3x 20k hour sims for crit and recorded the data.
4. I then removed the BS socket in RAWR and simmed to ensure I arrived back at the results in 1.
5. I then did 3x 20k hour sims with EP steps of 20 for hit and crit. All other EP calculations were turned off (to save time). I recorded this data.

Results:
The DPS values bore out using the standard simulation matched the DPS values seen with EP simulation to within 1-2DPS. There appears to be no difference between standard simulation and EP simulation. The crit having higher EP was probably just the regular variance you see while simming (more on that in the next section).

Second order of business:
How do crit and hit sim at top end gear levels?

Testing Method:
1. Based on the results from the first trial, I felt comfortable using the EP simulation test to parallelize the hit and crit calculations. I'm less concerned with the EP values here and more concerned with the DPS values so ignore the EP values for now.
2. I set the crit and hit step size to 20 and ran 3x 20k hour EP simulations and here are the results:

EP                        value     DPS       total DPS      difference

baseline                                      9445.06
ap                        1.00      0.74      9594.05        148.99
crit rating               1.77      1.32      9471.37        26.31
hit rating                1.79      1.34      9471.79        26.73
expertise rating          0.00      0.00      0.00           -9445.06
haste rating              0.00      0.00      0.00           -9445.06
armor penetration rating  0.00      0.00      0.00           -9445.06
spellpower                0.00      0.00      0.00           -9445.06
mh dps                    0.00      0.00      0.00           -9445.06
oh dps                    0.00      0.00      0.00           -9445.06
strength                  1.10
agility                   1.69
intelligence*             1.39
spirit                    0.00      0.00      0.00           -9445.06
mana                      0.00      0.00      0.00           -9445.06
mp5                       0.00      0.00      0.00           -9445.06

EP                        value     DPS       total DPS      difference

baseline                                      9445.19
ap                        1.00      0.74      9593.28        148.10
crit rating               1.83      1.36      9472.31        27.13
hit rating                1.81      1.34      9472.04        26.85
expertise rating          0.00      0.00      0.00           -9445.19
haste rating              0.00      0.00      0.00           -9445.19
armor penetration rating  0.00      0.00      0.00           -9445.19
spellpower                0.00      0.00      0.00           -9445.19
mh dps                    0.00      0.00      0.00           -9445.19
oh dps                    0.00      0.00      0.00           -9445.19
strength                  1.10
agility                   1.71
intelligence*             1.40
spirit                    0.00      0.00      0.00           -9445.19
mana                      0.00      0.00      0.00           -9445.19
mp5                       0.00      0.00      0.00           -9445.19

EP                        value     DPS       total DPS      difference

baseline                                      9445.03
ap                        1.00      0.74      9593.53        148.51
crit rating               1.92      1.43      9473.54        28.52
hit rating                1.79      1.33      9471.66        26.63
expertise rating          0.00      0.00      0.00           -9445.03
haste rating              0.00      0.00      0.00           -9445.03
armor penetration rating  0.00      0.00      0.00           -9445.03
spellpower                0.00      0.00      0.00           -9445.03
mh dps                    0.00      0.00      0.00           -9445.03
oh dps                    0.00      0.00      0.00           -9445.03
strength                  1.10
agility                   1.75
intelligence*             1.41
spirit                    0.00      0.00      0.00           -9445.03
mana                      0.00      0.00      0.00           -9445.03
mp5                       0.00      0.00      0.00           -9445.03

3. To test more drastic changes I simmed 50hit and 50crit steps (2.5-3 gems worth). I didn't want to go much higher because of the hit breakpoint that was mentioned earlier. These were the results:

EP                        value     DPS       total DPS      difference

baseline                                      9445.26
ap                        1.00      0.74      9593.86        148.60
crit rating               1.88      1.40      9515.28        70.02
hit rating                1.81      1.35      9512.61        67.35
expertise rating          0.00      0.00      0.00           -9445.26
haste rating              0.00      0.00      0.00           -9445.26
armor penetration rating  0.00      0.00      0.00           -9445.26
spellpower                0.00      0.00      0.00           -9445.26
mh dps                    0.00      0.00      0.00           -9445.26
oh dps                    0.00      0.00      0.00           -9445.26
strength                  1.10
agility                   1.73
intelligence*             1.41
spirit                    0.00      0.00      0.00           -9445.26
mana                      0.00      0.00      0.00           -9445.26
mp5                       0.00      0.00      0.00           -9445.26

EP                        value     DPS       total DPS      difference

baseline                                      9444.99
ap                        1.00      0.75      9594.02        149.04
crit rating               1.90      1.42      9515.81        70.83
hit rating                1.85      1.38      9513.76        68.77
expertise rating          0.00      0.00      0.00           -9444.99
haste rating              0.00      0.00      0.00           -9444.99
armor penetration rating  0.00      0.00      0.00           -9444.99
spellpower                0.00      0.00      0.00           -9444.99
mh dps                    0.00      0.00      0.00           -9444.99
oh dps                    0.00      0.00      0.00           -9444.99
strength                  1.10
agility                   1.74
intelligence*             1.41
spirit                    0.00      0.00      0.00           -9444.99
mana                      0.00      0.00      0.00           -9444.99
mp5                       0.00      0.00      0.00           -9444.99

EnhSim version 1.8.3.2

EP                        value     DPS       total DPS      difference

baseline                                      9445.46
ap                        1.00      0.74      9593.87        148.41
crit rating               1.86      1.38      9514.61        69.15
hit rating                1.83      1.36      9513.46        68.00
expertise rating          0.00      0.00      0.00           -9445.46
haste rating              0.00      0.00      0.00           -9445.46
armor penetration rating  0.00      0.00      0.00           -9445.46
spellpower                0.00      0.00      0.00           -9445.46
mh dps                    0.00      0.00      0.00           -9445.46
oh dps                    0.00      0.00      0.00           -9445.46
strength                  1.10
agility                   1.73
intelligence*             1.40
spirit                    0.00      0.00      0.00           -9445.46
mana                      0.00      0.00      0.00           -9445.46
mp5                       0.00      0.00      0.00           -9445.46

Overall Results:
Hit and crit scale roughly equally. Note this line:

crit rating               1.92      1.43      9473.54        28.52
hit rating                1.79      1.33      9471.66        26.63

Notice that crit spiked up 1DPS and hit dropped down 1 DPS? This 2DPS difference resulted in a 0.13 EP difference between hit and crit. These kind of fluctuations are normal for a simulation and would likely account for the 0.2 EP difference you're seeing. This should further reinforce a) The need for lengthly simulations and b) The need to sim actual gear sets rather than relying on EP (ie. what we're doing in this thread).

As for why this is happening there are a number of potential factors:
1) GCDs to be used for additional MW procs have opened up with the loss of 4pT8.
2) More hits mean more SS procs (especially with the 2pT9)
3) High amounts of ArP on coliseum gear makes white hits more valuable.

Regardless of why, the bottom line is that hit is as valuable as crit at our current top gear levels.

TLDR:
1) There is no bug with EP simming. It is indiscernible from regular simming.
2) Crit and hit are virtually identical at our current top gear levels. Both are still consistently worse than AP.

Perty · 09/02/09, 2:46 AM

Had a chance to check out the anub 25 TLC style trinket: Reign of the Unliving - Items - Sigrie

I can confirm the trinket DOES proc from FT crits, i did a run on the new vault boss with it and a quick duel test. I'm not entirely sure how to properly pick out the ft crits proccing the buff from the log that i've uploaded, but i'm sure someone smarter than me could do so. Here's the log, I'm Sharwa.

World of Logs - Real Time Raid Analysis

ophidiandusk · 09/03/09, 3:55 AM

As far as set bonuses go...

How long do you think it will be to replace the tier 8 4 piece with the new gear? As far as I can see all the other BiS listed sync extremely nicely with MW, and subsequently the tier 8 4set.

I can see the tier 9 4set overtaking tier 8 by a marked amount at item levels of 245+; provided that the increased flameshock damage is inclusive of the ticks as well. Does anyone know if this is the case?

Frontier · 09/03/09, 4:21 AM

Originally Posted by ophidiandusk

As far as set bonuses go...

How long do you think it will be to replace the tier 8 4 piece with the new gear? As far as I can see all the other BiS listed sync extremely nicely with MW, and subsequently the tier 8 4set.

I can see the tier 9 4set overtaking tier 8 by a marked amount at item levels of 245+; provided that the increased flameshock damage is inclusive of the ticks as well. Does anyone know if this is the case?

Your mileage may vary, but with the gear I had going into Trial (close to the ulduar best in slot list), I simmed higher at two piece 232 and two piece Worldbreaker, and higher still at four piece 232, than I did with four piece Worldbreaker. From what I was finding in the simulator, no single 245 piece would beat out a set bonus loss, but any two pieces of even the 232 would beat out the Worldbreaker four piece bonus.

This may depend a lot on your gear, so your best bet would be to get Rawr, plug your gear in, and then test various setups in the Enhsim.

Zensai · 09/03/09, 10:05 AM

The good, the bad the ugly... The way I am prioriting my loot, is picking up the BiS off pieces, and banking badges. After I have 3 or so BiS off pieces, I will pickup 2 or 3pc T9, and replace my gear in chunks. The good news about breaking 4pc-T8 is the fact our priority becomes alot easier to manage. Currently with BiS Ulduar gear, your fighting MW always being at 4/5. If you figure they fix our Totem to rely on us using LL to keep up the 400ap, and we break 4pc-T8 it will be alot more managable. Pre-ToC keeping up the PVP totem was a choice of LB or LL for the AP, then LB.

As with any regearing, you are usually alot better off stock piling a few pieces then doing a full gem/enchant switch. With how tight heroic 25m ToC is, you can't really afford to mess with loosing your set bonuses until you know for a fact you are increasing your DPS. 2 Offpieces + 2pc T9 should accomplish this. The ugly thing about the new items is the major lack of expertise, and forcing you to regem alot and go over caps because of the epic gems. I'm sitting on ~170 Triumph badges, and haven't been taking trophy's yet because breaking the set bonuses and current off pieces could lower my effectiveness and stuff is tuned very tight at this point. If your mainly doing 10m then it doesn't matter much knocking out heroic 10m ToC is pretty trivial. 25m is way to tight at this point to gamble.

Great analysis of the hit/crit. I've never shyed away from having alot of hit. Running with 420 to 460 hit, and I've been happy with it. With the new set bonuses and all of the ArP gear white hits will become a little more important, so getting closer to DW hit cap only makes sense if you can without sacrificing other key stats.