Thread: Trinks Question

Page 2 of 3 FirstFirst
1
2
3
LastLast
  1. #21
    Quote Originally Posted by Grunkh View Post
    I´m gonna hog the thread a little bit.

    I have gotten so used to playing with Carbonic Carbuncle.. But it´s time to replace it it seems.

    The trinkets i have available are:

    Relic of xuen
    Carbonic Carbuncle
    Lei shin (Normal)
    Lessons of darkmaster
    Darkmist Vortex (normal)
    Skullrender Medallion (missing 100 vps will get it soon)

    Hey dude!
    Seeing you're a Titan's Grip Warrior without Engineering so your Best trinkets are the Lei Shin heroic and the 5.1 trinket. Seeing you don't have LS heroic, just take the normal one.

    TG with eng: Lei shin + 5.1 trinket
    TG without eng: Lei shin + 5.1 trinket
    SMF with eng: Lei shin + Relic of Xuen
    SMF without eng: Lei shin + 5.1 trinket

  2. #22
    Quote Originally Posted by Beauterham View Post
    Hey dude!
    Seeing you're a Titan's Grip Warrior without Engineering so your Best trinkets are the Lei Shin heroic and the 5.1 trinket. Seeing you don't have LS heroic, just take the normal one.

    TG with eng: Lei shin + 5.1 trinket
    TG without eng: Lei shin + 5.1 trinket
    SMF with eng: Lei shin + Relic of Xuen
    SMF without eng: Lei shin + 5.1 trinket
    This is not correct. Relic of Xuen beats the 5.1 trinket in every situation. The only time it may lose to the 5.1 trinket is around 505ish ilvl for TG.

  3. #23
    Got anything to back that up? Simcraft disagrees.

  4. #24
    Simcraft is very unreliable, most people know this. Too many people think whatever Simcraft says is law these days. But the reasoning is that crit procs just don't compare to strength ones. The uptime on Xuen and Lei Shen is really high due to such low ICD's, and having str is more reliable since it will always give you a dps increase. All top warriors I've seen are all using Lei Shen's and Xuen.

  5. #25
    They're also engineers, which obviously means they wouldn't use the rep trinket.

    So your basis is the armory of a few top engineer warriors, no actual math or simulation of your own?

    "Having str is more reliable since it will always give you a dps increase" - does that mean we should gem strength because it will always give the same increase, while even by gemming crit we will sometimes have bad runs of RNG where we crit several % lower than crit rate suggests we should? Obviously not, and being able to stack a use trinket with bloodbath every time is going to go a long way to off-set the lower value of a crit use during recklessness.

    By comparison proc trinkets frequently fall out of sync with bloodbath every couple of minutes.
    Last edited by Rhia; 2012-12-31 at 07:15 PM.

  6. #26
    Quote Originally Posted by Rhia View Post
    They're also engineers, which obviously means they wouldn't use the rep trinket.

    So your basis is the armory of a few top engineer warriors, no actual math or simulation of your own?

    "Having str is more reliable since it will always give you a dps increase" - does that mean we should gem strength because it will always give the same increase, while even by gemming crit we will sometimes have bad runs of RNG where we crit several % lower than crit rate suggests we should? Obviously not, and being able to stack a use trinket with bloodbath every time is going to go a long way to off-set the lower value of a crit use during recklessness.

    By comparison proc trinkets frequently fall out of sync with bloodbath every couple of minutes.
    No, we shouldn't gem str because str gems only give 160 str where crit gives 320. Landsoul himself has said many times that Xuen/Lei Shen is best. And basic knowledge is better then doing a random simulation that is outdated and known to be unreliable.

  7. #27
    So basically, all you've got is anecdotal evidence.

    What if Xuen gave 2500 strength rather than 3027? How about 2000? Is there any point in your mind where a crit use would overtake it? Does your "basic knowledge" give you a concrete method of evaluating trinkets?

    Not sure how simcraft is outdated or unreliable, given its constantly updated and matches logs from real fights fairly closely, down to ability breakdown, buff uptime, etc.

  8. #28
    Herald of the Titans Darkfriend's Avatar
    Join Date
    Mar 2011
    Location
    Wisconsin
    Posts
    2,835
    Quote Originally Posted by Rhia View Post
    So basically, all you've got is anecdotal evidence.

    What if Xuen gave 2500 strength rather than 3027? How about 2000? Is there any point in your mind where a crit use would overtake it? Does your "basic knowledge" give you a concrete method of evaluating trinkets?

    Not sure how simcraft is outdated or unreliable, given its constantly updated and matches logs from real fights fairly closely, down to ability breakdown, buff uptime, etc.
    Ok, static and proc'ed stats have slightly different values. Static crit is worth more (higher SEP) than the same amount of average crit, proced, thus 1k static crit is better than 4k procced crit with 25% uptime. Proced strength has a fairly static value because of static scaling with strength so you can figure out how good a proc strength trinket is. Simply using the base simcraft SEP of crit, and doing the math on a crit trinket with static numbers doesn't work. (EG skullrender is 1152+(3838x.25x1.2)=2303 SEP, but that is not an accurate number. Look at Relic which is 956+(3027x.33)=1954. According to the base numbers, Medallion would be better. But that is flawed math, which is why trinkets are a PITA.

    The reason crit is so much higher for fury than it is for arms is because of 1.) RB 2.) Double chance of BT crit 3.) Enrage scaling with mastery. Including those factors, as well as the rotation of fury makes figuring out the math a bit more complicated.
    Last edited by Darkfriend; 2013-01-01 at 12:18 AM.

  9. #29
    All of what you said is very obvious. You're neglecting the fact that the rep trinket's use is always going to be used in conjunction with bloodbath and a colossus smash, which both will increase crit's value relative to its steady state average value. On the other hand 1 in 3 will be used in conjunction with recklessness which will diminish the value of crit. But on the other hand, 1 in 3 will also be used with skull banner, which will increase the value.

    My main point is that Kurimlin's assertion that "it's basic knowledge" and other similar hand-waving rubbish is a pretty awful basis to make decisions upon, and unless he can point out exactly how simcraft is flawed, that's the best tool we have to evaluate trinkets.

  10. #30
    On the other hand 1 in 3 will be used in conjunction with recklessness which will diminish the value of crit. But on the other hand, 1 in 3 will also be used with skull banner, which will increase the value.
    To make things even more complicated, banner is pretty much always used on the same 1/3 as reck so it's not a clean increase in value for crit on that burst

  11. #31
    Herald of the Titans Darkfriend's Avatar
    Join Date
    Mar 2011
    Location
    Wisconsin
    Posts
    2,835
    Quote Originally Posted by Rhia View Post
    All of what you said is very obvious. You're neglecting the fact that the rep trinket's use is always going to be used in conjunction with bloodbath and a colossus smash, which both will increase crit's value relative to its steady state average value. On the other hand 1 in 3 will be used in conjunction with recklessness which will diminish the value of crit. But on the other hand, 1 in 3 will also be used with skull banner, which will increase the value.

    My main point is that Kurimlin's assertion that "it's basic knowledge" and other similar hand-waving rubbish is a pretty awful basis to make decisions upon, and unless he can point out exactly how simcraft is flawed, that's the best tool we have to evaluate trinkets.
    Wait so you call someone out on something being obvious, and then when answers are given you say "well that's obvious." Am I missing something?

    And ya Delath, unless its a fight where you can fit in 3 reck and only 2 banners without not having them all for execute phase, you should indeed be waiting for that banner, meaning crit is worth even less every 3rd use.
    Last edited by Darkfriend; 2013-01-01 at 09:50 AM.

  12. #32
    So, to those saying relic > skullrender (for non-engineers), how do you back that up? You say simcraft is not reliable, so what tool or math do you use to rank trinkets?

    Saying strength procs are better than crit procs doesn't cut it, because obviously things like amount of stats procced, uptime and synergy with other cooldowns matter. So you need some method of weighting stats. If it isn't simcraft, what is it?

  13. #33
    Saying strength procs are better than crit procs doesn't cut it, because obviously things like amount of stats procced, uptime and synergy with other cooldowns matter. So you need some method of weighting stats. If it isn't simcraft, what is it?
    Personally my biggest issue with skullrender (apart from me happening to be engineer, never mind my out of date sig.pic.) is the fact that it syncs badly with 4set reck. For SMF this is especially highlighted in comparison with relic since SMF typically values str higher in relation to crit than TF does at least for the time being. I don't see a problem with a simcraft approach as long as the action priority is fixed to work as realistically as possible, which is not something I'm able to create myself but I'm confident that some people know how to do this.

  14. #34
    Quote Originally Posted by Darkfriend View Post
    Wait so you call someone out on something being obvious, and then when answers are given you say "well that's obvious." Am I missing something?

    And ya Delath, unless its a fight where you can fit in 3 reck and only 2 banners without not having them all for execute phase, you should indeed be waiting for that banner, meaning crit is worth even less every 3rd use.
    Obvious: static and proc SEPs are not the same. You claim that crit proc SEPs are intrinsically worse without any actual basis though. I should hope it doesn't need explaining that a hypothetical 4k strength proc/use that lasts 15 seconds and procs every 60 seconds like clockwork is worth considerably more than 1000 passive strength - a situation where a proc/use outweighs static SEP. So given that is quite possible for a proc to outweigh the averaged value, why is it that crit is the exception? Because it's RNG? That seems to be what the previous poster was arguing. Newsflash, RNG governs multiple aspects of our rotation, we still value crit extremely highly.

    Reck devaluing the crit use is certainly a valid point, but you can't just ignore the fact that being guaranteed to stack with skullbanner, bloodbath and a full colossus smash all work in the opposite direction to promote the value of crit. Without math or simulation it's just speculation, hand-waving and anecdote, which is what I object to.

    Also obvious: why fury values crit. Quite aware of that. You didn't actually give any sort of answer to my question of how the previous poster would find the cut-off point where one trinket would overtake another beyond his claims of "basic knowledge".

  15. #35
    Herald of the Titans Darkfriend's Avatar
    Join Date
    Mar 2011
    Location
    Wisconsin
    Posts
    2,835
    Quote Originally Posted by Delath View Post
    Personally my biggest issue with skullrender (apart from me happening to be engineer, never mind my out of date sig.pic.) is the fact that it syncs badly with 4set reck. For SMF this is especially highlighted in comparison with relic since SMF typically values str higher in relation to crit than TF does at least for the time being. I don't see a problem with a simcraft approach as long as the action priority is fixed to work as realistically as possible, which is not something I'm able to create myself but I'm confident that some people know how to do this.
    Also, berserker rage slightly devalues on use crit. The issue with BR is one of balancing how often it is used with how often it is available to be used, but is not. Sit on BR for too long each time, and you end up with a situation where you could have used BR more time total in a fight, but didn't. (The same problem sitting on any CD has, really.)

    The other factor is that the ICD of both of the two BiS trinkets, relic and LS are the same, although relic has an uptime of 15 second and LS is 20 (~30% overall uptime and ~40-43% uptime) meaning they are both up at the same time, giving even more value to the proc'ed strength.

    Now, to explain why Skullrender isn't as good as the SEP would seem to make it. The reason Crit is so high for fury is because of all the reasons I listed before. Take away the value of any particular crit, and the value of the accompanying crit value drops massively. After a certain gear point and with 4 set, every 3rd use of the trinket will line up with Reck/Banner-not optimal because you'll already be critting w/ BT, be enraged, and procing RB, meaning the crit buff is useful for non-BT abilities, but still worth far less than normal. Or you could delay the on use, which means its also worth less, since it's not lining up with other cooldowns. On the other side, the proc'ed strength loses much less for occurring at an non-optimal time, such as right after you finished using CDs, and gains even more for timing it with CDs.
    To make it a bit more clear, for sake of argument say every BT crit. You'd always be enraged and you'd always have RB to use after. If this were true, the value of crit would drop to ~.5 SEP (Because DR always crits, but Skullbanner increases crit damage, so it's not exactly .5) This is similar to the case of every 3rd reck.

    Honestly, it's not as simple as simply looking at stat weights and doing math. You have to compare uptime, ICD, if its a proc how often it will (on average) line up with CS/CDs, etc.
    Last edited by Darkfriend; 2013-01-02 at 12:05 AM.

  16. #36
    Quote Originally Posted by Darkfriend View Post
    Honestly, it's not as simple as simply looking at stat weights and doing math. You have to compare uptime, ICD, if its a proc how often it will (on average) line up with CS/CDs, etc.
    Thing is I've not once mentioned looking at stat weights as a basis for trinket selection. Simcraft shows skullrender + lei shen heroic ahead for non-engineers, and simcraft doesn't work out its results by looking at stat weights, it simulates a run, sees what the dps is, and repeats that however many iterations. This obviously takes into account uptime, ICD, how often it will proc, how it lines up with other CDs etc. The other guy is claiming that simcraft is faulty without any sort of explanation why, and not offering any sort of alternative method of comparing trinkets beyond anecdote.

    While criticising someone (presumably me, though I never talked about using SEP as a basis of trinket comparison) for taking a limited view of the situation, you're discounting the factors that increase the value of the crit use, only focusing on the factors that decrease the value. As you're pointing out, trying to figure out which trinket is better analytically is very difficult, but Simcraft doesn't try to analytically model any of this so don't know why we're talking about analytical models at all really.

    Strength procs out of bloodbath and CS do lose a hell of a lot of value by the way, all stats do. During bloodbath and CS you're doing roughly 82% more damage than when neither of those are up. Enrage + CS + BB is roughly 149% more damage than nothing up. Yes of course crit trinket stacking with recklessness devalues it, although not entirely - you still go from ~75% to 80% chance on specials, 25% to 30% on white hits, but you can't just look at one factor and decide based on that. It's even more important to be enraged when bloodbath and cs are up, and an extra 10% bt crit helps a lot in that regard on the 2 out of 3 occasions reck isn't up. And when the difference between a crit and non-crit execute is around 500k damage, the difference between 75% and 80% execute crit is fairly significant.

    Basically yes obviously trinket loses some value from being stacked with reck, but it also gains value from being stacked with other cds, trying to figure out which is the greater effect is extraordinarily difficult to solve analytically, so we use simcraft which will model both of those effects. Blindly criticising simcraft as being inaccurate without any actual examples of why it's inaccurate, or providing any alternatives, as the previous poster did, doesn't help anything.

    Simcraft shows engineering + relic + lei shen roughly 1-1.2k dps ahead of rep trinket + lei shen + a 320 str profession, as you would expect, if anyone's curious.

    It also shows rep trinket + lei shen roughly 500-700 dps ahead of relic + lei shen for non-engineers, so at least if simcraft is accurate, the benefit of stacking the use with bloodbath + a full cs every time outweighs the diminishment in its value from reck every 3rd bloodbath.

    Edit: if someone doesn't understand what I mean when I talk abot analytical modelling vs simulation, read this. Of course there are bound to be mistakes in simcraft, possibly in its engine (e.g. it might model a certain ability incorrectly - e.g. it was/is calculating dragon roar vs multiple targets incorrectly compared to how that ability works in game), and possibly in the action priority list (for example it doesn't try to stack execute burn with trinket procs, which a real player might). But the creators of simcraft aren't intentionally trying to produce something that's inaccurate, and will listen to, and implement, feedback in an effort to try to make their tool as accurate as possible. So just stating "simcraft's useless" without any actual useful reasoning as to why, is pretty unhelpful.
    Last edited by Rhia; 2013-01-02 at 03:40 PM.

  17. #37
    It also shows rep trinket + lei shen roughly 500-700 dps ahead of relic + lei shen for non-engineers, so at least if simcraft is accurate, the benefit of stacking the use with bloodbath + a full cs every time outweighs the diminishment in its value from reck every 3rd bloodbath.
    At roughly what ilvl did you do the sims and did you use SMF/TG? Just curious, no criticism intended

  18. #38
    My first sims were for my own character, 494 ilevel, SMF. Decided to sim better geared SMF/TG, and the gap between engineering + relic and enchanting + rep trinket appears to close significantly as weapons improve (better weapons increase the value of secondary stats much faster than the value of strength), almost surprisingly so.

    For Landsoul (ilevel 500 TG, bloodbath/dragon's roar, oh changed to dancing steel):

    Engineering, relic, heroic lei shen: 120,399
    Enchanting, relic, heroic lei shen: 119,712
    Enchanting, rep trinket, heroic lei shen: 120,710

    For Sinnermighty (ilevel 504 SMF, bloodbath/dragon's roar):

    Engineering, relic, heroic lei shen: 123,612
    Enchanting, relic, heroic lei shen: 122,783
    Enchanting, rep trinket, heroic lei shen: 123,611

    Just to check that the results make sense: with relic + lei shen + engineering, the average value of all hits was higher than with rep trinket + lei shen + enchanting but enrage uptime was 2% lower in without rep trinket, with 2 less raging blows cast without rep trinket, and the crit % on special abilities was 1.5 - 2% lower, 3% lower on bloodthirst. So the expected effeclts from losing strength procs and gaining a crit use are reflected by the sim, the magnitude of the effect is perhaps a bit surprising though (ie going down in average hit size but having more abilities crit, as well as more raging blows and enrage uptime), so that's at least somewhat reassuring that the sim isn't doing something fundamentally wrong.

    This action priority list doesn't attempt to stack execute burn with trinket procs though, which means it is underestimating the power of engineering + double proc trinkets a bit.

  19. #39
    Field Marshal
    Join Date
    Oct 2012
    Location
    Belgium
    Posts
    63
    After reading this and almost had my head exploding from all the info, im still stuck on my trinkets (upgrade wise) and not sure where else to go for advice.
    (Im not a numbers guy, will never be, so hope i could get some responses from guru's like you guys)
    http://eu.battle.net/wow/en/characte...oshke/advanced

    I got relic and 5.1 rep trinket, i wont be doing Will HC for a while as our progress will be HoF so i wont get LS HC trinket.
    With what i just read, for non-engineers, 5.1 shows better, so this means i would upgrade that first? Or am i really just failing at understanding this.

    edit: i heard also to save up VP, but im close to cap, so its time to upgrade something.

    Thanks in advance.

  20. #40
    @Mooshke
    Upgrading your weapons is always a safe DPS increase, main-hand more so than off-hand ofcourse.
    And the cloak you got there is probably the best one you will acquire and so it will be a safe investment for now if you're unsure about which trinket to upgrade at this point(both should be a solid dps increase though)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •