See, I'm all for skepticism -- but unless you can PROVE that the assumptions MAXDPS is making will result in bad gear choices, you're not really theorycrafting. Here's a good assumption to start with:
Well, again, there are lots of reasons to use slower weapons, of which the cooldown is only one. However, weapon choice is a bit of failure. If you enter a pair of 1.3s weapons, you get a "DPS value" that's about 4% lower than if you enter a pair of 2.7s weapons (on Yo's it's more like 17%). As a result, there are some bad suggestions in there such as Stellaris and the Stormreaver Warblades, and Guile of Khoraazi is way over-represented. In any case, the best weapons are 2.6.
As for hit rating: using 5-man dungeon stats (1000 AP, 15% crit, 15% hit), your "DPS per 10" ratings are Hit Rating: 1.60, Crit Rating: 2.73, Strength: 4.18 with 1.3s weapons, vs 1.77, 2.85 and 4.38 for 2.7s. So actually, it's upvaluing HR with SLOW weapons (just checeked, yo's does the same) -- but it's also making it blatantly obvious that strength and crit rating are more important. So, there's one for MaxDPS.
So it has a ways to go. Would it offer MaxDPS? Well, the gear this site claims is best is the SAME as the gear other methods claims is best, so maybe. As an armor upgrade selection site, it does really well. But its weapon choices are, as you hypothesized, off. However, I don't think it's impossible to find some way to generally express the 3s cooldown's effect on DPS in a flat equation -- say, each .1s of speed beyond 1.3s is worth 1.2% additional damage -- derived from simulations.
I just guessed that the faster weapons would over value hit. I was wrong on that (and am still pretty surprised by that... I'll have to think more about why that is), but since it was an example of a possible outcome, I don't think it matters. As for it recommending gear improperly, I already showed 2 examples (ring and something else, in an ealier post) where it ranks them wrong.
Well, wrong if you consider Yo!'s sim to be a more accurate test. I know I sure do.
It is possible that it assumes I'm hitting from the front, thus doubling the effect of expertise?
(Probably not because the attack table has Parry/Dodge at 5% rather than ~~15 if it included parry)
If this is the case it's made my day 'cause Supremus has only dropped his necklace twice in over 20 kills.
Not to mention a solid chunk of H P.
He doesn't say how he uses expertise. I have no doubt he doesn't use it correctly. Don't pay attention to crazy things like that. By the way, i use the Brooch of Deftness - it's nothing personal I have against expertise (I'm at 56 rating = 14 expertise atm). I would love to upgrade to the supremus neck, but our guild has a huge backlog on them and I'm just a casual.
There ya go toots, another good example of bad gear ranking from the site.
Well, what are your other stats looking like? That Brooch doesn't show up on my list until I enter an absurdly high value for unbuffed crit rating and an absurdly low value for hit rating, like 42% with 0% hit. In which case the extra crit on the choker would be pretty useless, the hit would cancel out, and you'd be effectively comparing 1.25% additional hits vs 72 AP.
You aren't entering raid buffed values in those fields are you, are you? It's looking for your un-raid buffed paper doll values for crit percentage and hit rating (yo's expects a hit percentage).
Disq: I feel that, due to a general trend towards good gear suggestions, this site could be really useful. In the areas where it fails, I'd rather provide constructive feedback than write it off as unreliable. In any event, I'll still be using Yo's to dream up and test hare brained new gear combos -- it's an uncanny resource.
Last edited by Toots Hepcat : 12/13/07 at 6:22 PM.
He doesn't say how he uses expertise. I have no doubt he doesn't use it correctly. Don't pay attention to crazy things like that. By the way, i use the Brooch of Deftness - it's nothing personal I have against expertise (I'm at 56 rating = 14 expertise atm). I would love to upgrade to the supremus neck, but our guild has a huge backlog on them and I'm just a casual.
There ya go toots, another good example of bad gear ranking from the site.
I don't see how its possible he's getting that neck there.. from my stats it's not even listed on the list for necklaces. Something seems very wrong and I don't think its the site.
Rest of the items are 99% identical to my thottbot lists, with afew differences such as belts, but that because thott doesn't have expertise yet.
Although it would be nice if the site would cache or have a save option for your inputted values.
Hit rating does not support faster weapons without cooldown due to fact that WF is normalised by flat proc rate. Meaning it will be equal for all speeds without cooldown. The reason why slow weapons still pull ahead is SS.
Ok, I maintain a list here for my own reference as to which bosses I need/don't need, and haven't updated it since ZA to see what's new. Also, now that I'm finally in BT and have access to Epic gems, that changes the weights on some items for me, with some gemmed items overtaking non socketed ones.
So using Loot Rank and Maxdps, I made a double chart of top items for each slot (for me) with base weighted values for the gem slots, and then maxdps did it's own thing with socketing epic gems and weighting based on my stats (mix t5/6). Now I know maxdps has formulas that are still under construction, and I may tweak it again later, but these items feel pretty solid to me as to their rankings. I know that a lot of you prefer to run a sim based on gear changes, which may effect a few items, but I'm fairly comfortable with generating a list and going by that, and hopefully this will help others as well.
The one issue I did have, which seems to mainly effect gloves, is expertise rating. I think I may have over-weighed it, and will have to look into it more later.
I also made an effort to extend each list far enough to include at least one boe crafted or pvp/badge item. Which means that at least one item in these end game gear lists should be obtainable by anyone without relying on luck. I couldn't really do it with rings though, so the pvp rings are your best bet most likely. Ideally a person can essentially get top 10 gearing without needing to rely on a random drop by using these lists.
Loot Rank URL used * I don't remember where I got my expertise rating from, it may be wildly inaccurate.
Thottbot URL (Used it to verify at first, but lootrank came out even enough that Thott wasn't needed)
Head (sans gems, but still weighting slots using thottbot/lootrank):
*Note, the variance here is huge compared to other items. Mainly due to the presense of weapon expertise on some items, in combination with epic gem slots. I'm not sure where Gloves of the Searing grip should actually fall.
*Note, Band of the Eternal Champion does not have its proc taken into account for Loot Rank, so is much lower there.
Maxdps does seem to model it however, as it has a much higher ranking compared to the non proc version before it. Band of Devastation did not appear in the Loot Rank list.
I actually disagree with spamming LHW, it heals for so little over what you get hit for even with SR up. I would rather take my chances with my combined dodge/block/parry/miss in a raid setting and let the healers do that job instead.
If I'm pulling aggro on a boss fight, you're right, I wouldn't be spamming LHW - I'd be /praying. On a heroic trash mob, LHW is likely to actually go off in between their hits on me, and I'd avoid spell knockback. I find it more effective than regular HW for that reason, and it helps to keep me up while the healer switches healing targets. I PUG a lot; I don't like to assume that the healer is good/aware/responsive.
I know that a lot of you prefer to run a sim based on gear changes, which may effect a few items, but I'm fairly comfortable with generating a list and going by that, and hopefully this will help others as well.
Everything you just listed in that huge post is fully explained in the OP for people to do on their own.
I checked out that Loot Rank site just now. It will be a nice replacement for Thottbot and Lootzor (which has long since ceased being useful) if they can get the bugs fixed. (Specifically, haste rating and expertise rating are definitely not working properly at the moment.) I still don't get the seeming insistence these types of sites have with "normalizing" the scores. If they would just show me what the item would be scored based on my values, it would make it a lot easier to realize when scores for specific items are bugged in some way. Frankly, I hope Wowhead comes out with something soon that ends this problem once and for all.
Hi Rob,
Loot Rank has been up for around 1 week now and we have fixed the expertise rating and haste rating calculations. We do not normalize the scores since it is such normalizing tends to confuse rather than help (and also allows for easier checking of score accuracy).
Disq: I feel that, due to a general trend towards good gear suggestions, this site could be really useful. In the areas where it fails, I'd rather provide constructive feedback than write it off as unreliable. In any event, I'll still be using Yo's to dream up and test hare brained new gear combos -- it's an uncanny resource.
I absolutely agree and that's why I've written a detailed explanation of every problematic area I've found in his forums where he provides the formulas. I *want* that site to succeed, because I like the combination of maximizing dps and providing gear to help do that in the same place. But, in its current incarnation, it doesn't do that.
I don't just complain, I've done my best to help the site author fix the problems. It's in his court now. I hope for all of our sakes that he does so.
It's the same that I've done with each sim or calculator where I've been able to see the code -
Pater's
Mine
Friedrick's
Yo!'s
Enshamulator (I forgot the author's name)
an Excel based on that's faded into obscurity.
I really do try to help, and put effort into it. I know my critiques seem harsh, but after going through my 7th simulator now, and finding some of the same issues, but in this case worse, than in the previous 6, I don't have the effort left in me to sugar coat my critiques. That doesn't mean that I don't appreciate the authors' efforts immensely. Having made one of these myself, and not even building the functional and pretty front end that enshamulator, Yo!'s, or maxdps's does, I can really appreciate how hard these people work on them.
Originally Posted by Polar
Hi Rob,
Loot Rank has been up for around 1 week now and we have fixed the expertise rating and haste rating calculations. We do not normalize the scores since it is such normalizing tends to confuse rather than help (and also allows for easier checking of score accuracy).
Polar
Lootrank author
I really like the site and have my personal stats bookmarked (yay for querystrings instead of post-data forms!)
Originally Posted by Atren
Hit rating does not support faster weapons without cooldown due to fact that WF is normalised by flat proc rate. Meaning it will be equal for all speeds without cooldown. The reason why slow weapons still pull ahead is SS.
Just looking over the first post. In short, it says:
[Merciless Gladiator's Cleaver]
97.5 DPS (MH: x9.03 = 880.4) (OH: x3.70 = 360.75)
+27 Stamina
Equip: Improves hit rating by 10. (x1.4 = 14)
Equip: Improves critical strike rating by 19. (x2 = 38)
Equip: Improves your resilience rating by 12.
Equip: Increases attack power by 30. (x1 = 30)
Total MH EP = 962.4
Total OH EP = 442.75
[Netherbane]
96.5 DPS (MH: x9.03 = 871.4) (OH: x3.7 = 357)
+25 Agility (x2 = 50)
+21 Stamina
Equip: Increases attack power by 40. (x1 = 40)
Total MH EP = 961.4
Total OH EP = 447.05
Comparing the two weapons we can see that the difference in the MH for each weapon is relatively minor with a 1 EP difference, but in the OH the Netherbane performs better with a 5 EP difference.
So if [Netherbane] outperforms [Merciless Gladiator's Cleaver] as a OH, why isn't it listed in the choice OH section? I assume this was a simple oversight, so I thought I would bring it up.
Loot Rank has been up for around 1 week now and we have fixed the expertise rating and haste rating calculations. We do not normalize the scores since it is such normalizing tends to confuse rather than help (and also allows for easier checking of score accuracy).
Polar
Lootrank author
Finally looked at the site. I think it looks great!
Two points to try to help out:
1 - can there be an option for limiting gear to a certain armor type? I try to stick with mail when possible (ducks from the pro-leather based comments) and removing leather options would be nice.
2 - I notice it crashes it when you replace numbers you don't need in the stats by using the space bar to clear them. I had to go back and delete the space in order to get the list. Just FYI for bug fixing (or anyone else who tried using it but crashed like I did)
So if [Netherbane] outperforms [Merciless Gladiator's Cleaver] as a OH, why isn't it listed in the choice OH section? I assume this was a simple oversight, so I thought I would bring it up.
It's highly dependant on your stat weights honestly. I did the numbers for my weights, and the Merc glad was higher by a few points. Not to mention, I think he just listed items that were attainable pre-t5 zones, as all of those are either Kara, Arena, ZA, or crafted.
I absolutely agree and that's why I've written a detailed explanation of every problematic area I've found in his forums where he provides the formulas. I *want* that site to succeed, because I like the combination of maximizing dps and providing gear to help do that in the same place. But, in its current incarnation, it doesn't do that.
I don't just complain, I've done my best to help the site author fix the problems. It's in his court now. I hope for all of our sakes that he does so.
I checked the forums over on maxdps.com. Disquette provided the exact kind of bug reports that any developer would want, and though he even apologizes there for being 'harsh' he certainly isn't - he's just factual. And to the site's credit, it appears Nuuga is actively trying to act on all the suggestions that Disquette made. So much ado about nothing here I think.
Much appreciation to both Nuuga and Disquette for hashing it all out. As great as Yo's sim/Loot Rank are, as a casual player maxdps.com is great for a stuff-in-my-stats-and-get-suggestions out without needing to do a lot of leg work (ie., yes I'm lazy).
So if [Netherbane] outperforms [Merciless Gladiator's Cleaver] as a OH, why isn't it listed in the choice OH section? I assume this was a simple oversight, so I thought I would bring it up.
Its not listed because that section on Offhands was intended to tell people where they could quickly pick one up and compare them, the idea was not to list each and every single OH. It was one of the original portions of the post that was trying to inform people to use slow weapons.
I'm the author of MaxDPS.com and just wanted to let you know that the page is in beta. I'm working very hard trying to make it as accurate as possible, and I realize that peer review is the only way to do so. I'm willing to open up any part of my code if it will help the site become more accurate. If you're willing to help (As Disquette has done, thanks!) - rather than muddle up this thread/forum, feel free to drop me a line or post on the forum on my site.
To respond to one thing in particular that was posted:
MaxDps 'values' put the [Brooch of Deftness] at 46.5
And the [Choker of Endless Nightmares] at 42.51
He doesn't say how he uses expertise. I have no doubt he doesn't use it correctly. Don't pay attention to crazy things like that. By the way, i use the Brooch of Deftness - it's nothing personal I have against expertise (I'm at 56 rating = 14 expertise atm). I would love to upgrade to the supremus neck, but our guild has a huge backlog on them and I'm just a casual.
There ya go toots, another good example of bad gear ranking from the site.
I had an error in the code that was using expertise rating as expertise percentage. This has been fixed (again thanks to Disquette).
As you can see, I'm constantly working on the site trying to make things better. But as it is in Beta, please use at your own risk. I'm trying to help, not harm, the community.
First of all escuse my english thats not very good. I'm french, so I'll try to make my best.
Thanks for this tread that I find very interesting.
I couldn't read the whole tread because there are to many pages and it'll take me to much time.
I have a Shaman as you can suppose. With my guild, we have clean SSC and we are in front of kael'thas that (I think) will be down soon.
So I've begin to look at item that are dropped in BT and Hijal.
My shaman is currently (unbuff) :
1592 AP
30,01% crit
112 hit
The most interesting items in BT/Hijal has a lot of crit/agi but not a lot of AP/force or hit in comparaison with T6.
My question is: Does it exist a cap crit or is crit/agi always so interresting when you have a lot of it? Have we to consider crit as < than 2EP after a specific value?
I'm surprised that force/AP as not a higher value after the 2.3 that make increase the horion damage with the AP. Could you explain?
Sorry if the aswers are somewhere in the thread and thank a lot for aswering.
I am a first time poster, and knowing how things are done around here ill get right to the point.
For quite some time i have used the information on this thread as a form of refference, at best a very solid base to maximize what I do as an enhancement shaman.
I have found that while there is a lot of mathematics that have been shown here, there are things that dont translate very well to actual practice.
My main question is regarding Hit Rating as a whole. I'm not looking for a specific number, because there honestly isn't. To illustrate my point I need to give a bit of an explanation first.
AEP is currently being used as a form of measurement for the potential DPS of particular items, and lately I have begun to question the exact accuracy of these numbers - though admittedly its under a different light.
Im sure like myself, to all experienced ENH shamans, gear upgrades wont often times be about the individual piece vrs another, but instead the synergy between the new gear upgrade and the rest of the set you have created for yourself. Having a PvP set and a PvE set has allowed me to shift some stats and do some testing for myself on how much of an effect hit currently has. I am not here to show crazy math, Im actually here to get you geniuses look into this since you have more tools than i do. Here is the Dilema:
The difference between my PvP set vrs my PvE set stats wise is +50atk, +.4 crit - 35 hit for PvP, where as my PvE is obviously -50atk, -.4, and +35 hit. In short my DPS with my pvp set is lower. Now, to some it may seem obvious but here is the thing:
Per AEP calculations each piece that i replaced (a total of 2) was replaced by a higher AEP value piece, however the overall result was lower DPS.
I realize that hit may become more crucial vrs certain mobs..and there are a few of them, but even stepping down to Gruul's/Mag to do some DPS testing the differential in DPS totals per recount/sw was +- 50 DPS, sometimes a little higher, all in favor of the PvE gear.
For those of you who will have this question, my total + hit with PvE is at 173, where my PvP is at 138.
I have read in numerous places (mind you, as obvious opinions) that a "good" + hit is around 120, but frankly, from what I am seeing in actual results, it seems to me the sweet spot is between 175-200, which i realize it seems like a lot.
Part of my concern is that around the SSC part of progression which is where im currently at; the variety of upgrades offer a crit and a atk power boots but at the entire cost of +hit on the item replacing. Most items at this level have anywhere between 15-18 +hit on them, so its roughly 1% hit (sometimes more sometimes less) that is being lost to items that on paper, due to higher ATk and crit values, have a higher AEP number.
I realize that many of you will ask for examples, and I intend to add 2 items for questioning as soon as i find out how to do links where the item shows. If not ill just write down the stats.
If anyone has noticed, or perhaps is curious to analyze this, please do so. Im sure any/all results will be posted here.
ok here is a good example.
Cataclysm Gauntlets (143.92 AEP) vs Tricksters Stickyfingers (152.20 AEP).
My issue is sitting at 173 +hit (and I have all talented bonuses for +9 total) loosing 1 percent actually lowers my DPS by loosing 25 + hit, where the gain on paper for the tricksters doesnt make up for it (i hope that makes sense).
If i figured out the AEP wrong, please let me know, otherwise - for theory craft, has anyone considered having certain +hit thresholds not necessarily based on gear but perhaps a reflection of the content you are in? or does it seem completely unrelated?
Buffed, passed 2400+ ATK, and over 35 crit (before procs), the 50 gain in atk really doesnt seem to stack up to loosing 1% of hit in what i am seeing.
For actual DPS numbers I can hover between 1050-1130s in DPS In PvE gear, while in -hit gear (PvP) it was between 970-1070's DPS. These results were consistant over 3 weeks of Gruul/Mag.
In MAXdps attack table you cant see relative mob level difference.
dodge is 5% instead of 5.6%
crit is 25% instead of Yo's 22%.(what is right value?)
MAXdps dps list.
Current Stats
White DPS: 239.93
Windfury DPS: 229.23
Stormstrike DPS: 75.10
Earth Shock DPS: 50.91
Flame Shock DPS: 60.11
Total DPS: 655.27
Yo's
Current Stats
White DPS: 237
Windfury DPS: 178
Stormstrike DPS: 68
Shocks DPS: 104
Total DPS: 588
White damage look almoust perfect(higher dodge% and lower crit% put these numbers par)
Windfury damage is way too high. Almoust +30% too high.
Stormstrike damage look almoust OK.
Shocks damage is good. Cant say how MAXdps value SS debuff.
I fail to see how you can be doing less DPS with 50 ap .4 crit vs 35 hit, and I really think Gruul/Mag are horrible fights to cite for testing. There are a ton of elements to DPS like raid/group composition, boss abilities, and buffs. With such a small sample size (3 may seem like a lot but it is pretty insignificant) and I find it very difficult to believe that 1% hit makes a 5% damage difference.
As you progress through content your hit drops, I was at ~200 hit rating during Gruul/Mag(THANKS SEBUDAI), 150 through SSC/TK, and to my current 110(which will be back up to 150 RNG willing). Many, many people have came here and had revelations about how insignificant hit is, myself included, it is best to embrace countless hours of work and simulations instead of fighting based on a very minimal sample size on very random fights.
Son, to me a robot's just a garbage can with sparks comin' out it.
I keep forgetting to ask one addition to Yo simulator -- could you add 4-set bonus of Arena set (9 sec SS) to it? Also does anyone know if Ashtongue trinket has any cooldown to prevent back to back SS with 9 seconds?
12/13/07, 6:04 PM
Similar Threads
