1. #1
    The Patient
    Join Date
    Sep 2010
    Location
    Dallas/College Station, Tx
    Posts
    327

    Trinket Question

    Armory:
    http://us.battle.net/wow/en/characte...heese/advanced

    After simming my toon many, many times, with both trinkets I have available besides Relic of Yu'lon (which are my Zen Alchemist Stone and Vision of the Predator) on various types of fights, my DPS is basically in the neighborhood of 72k sustained dps (or so says SimC), and Mastery is like... 1.1 whereas haste is 0.8, stat-weight wise.

    Here is my question, though: Askmrrobot is telling me that Zen Alchemist Stone is literally the absolute best thing I can get outside of a raid of any sort. Better than Static-Caster Medallion, better than Blossom of Pure Snow, better than even LFR Essence of Terror. I have a hard time believing this. Yes, having mastery on a trinket, and the proc alone, are very good, but I'm not sure if I'm screwing up somewhere along the line to get results like this. Is static intellect trinkets, even with good procs, like, not as good as mastery and a 15 second proc?

    Any and all helpful criticism would be welcome, even not about my trinket situation. Thanks in advance.

  2. #2
    While the Zen Alchemist Stone is amazing pre raid, I'm not sure why AskMrRobot is telling you it is as good as it's said. On my own shaman I wouldn't trade anything for my Relic of Yu'lon and LFR Essence of Terror. And static int trinks with good procs are ideal for most caster classes. That being said, I'd place Zen Alchemist Stone over Vision of the Predator, because ~800 mastery and with ~4k int proc is by far better than ~800 int and 3.3k crit.

  3. #3
    You can also just test it out on a dummy and see what kind of results you get. 10 minutes with each trinkets should tell you much more than just having a website tell you. Get a feel of it instead of relying on indicators.

  4. #4
    Hey Joshdatroll, I'm from AskMrRobot. First, you might be interested in this page: it lists all of the trinket procs average value. We get that value by simulating trinkets on every class and spec.

    Next, I don't see Zen Alch stone being listed at BiS. Here's what I see when I select the following gear filters:
    "PRE-Raid" (No LFR): I see Relic of Yu'lon and Static-Caster's Medallion.
    "LFR": I see Relic of Yu'lon and Light of the Cosmos (LFR)
    If you do in fact see Zen being suggested, can you save your profile and post the link? I'll make sure there isn't a bug. Also note, if you change the stat weights, it's possible Zen ends up ranking higher.

    Based on the proc info and default stat weights, here is how the trinkets rank. Feel free to re-do the math with your own weights.
    Int weight: 3.74
    Haste weight: 1.65
    Mastery weight: 1.46
    Crit weight: 1.39

    Relic of Yu'lon:
    956 Int * 3.74 = 3575
    851 average Int from proc * 3.74 = 3183
    TOTAL SCORE: 6758

    Light of the Cosmos (LFR)
    956 Haste * 1.65 = 1577
    1202 average Int from proc * 3.74 = 4496
    TOTAL SCORE: 6073

    Essence of Terror (LFR)
    1021 Int * 3.74 = 3819
    1211 average Haste from proc * 1.65 = 1998
    TOTAL SCORE: 5817

    Static Caster's Medallion
    1152 Haste * 1.65 = 1901
    1022 average Int from Proc * 3.74 = 3822
    TOTAL SCORE = 5723

    Zen Alchemist Stone:
    809 Mastery * 1.46 = 1181
    1159 average Int from proc * 3.74 = 4335
    TOTAL SCORE: 5516

    Vision of the Predator (H)
    847 Int * 3.74 = 3168
    989 average crit from proc * 1.39 = 1374
    TOTAL SCORE: 4543
    Optimize your gear on AskMrRobot.com
    From Mr. Robot: Find the best gear in your bags, help & faq, and Preview our combat log tool.
    Mr. Robot is also on Twitter

  5. #5
    Zoopercat, first off thanks for mr.robot is an amazing site. I'll probably disagree with 2 of your trinket calculations though.

    First off, you should use the 466 ilvl Zen Alchemist Stone in the comparisons, because all serious raiders have the upgraded version (+8ilvl). It's practically a free upgrade: just 1500JP, so it doesn't hinder any gearup progression for most raiders at this point, unlike VP upgrades. Imo when you're at an ilvl of say 475+, mr.robot should use upgraded blues as default in the gearscore lists, because the upgrade is practially free.

    Zen Alchemist Stone (466)
    871 Mastery * 1.46 = 1271
    1249 average Int from proc * 3.74 = 4671
    TOTAL SCORE: about 5942 instead of 5516. Significant difference.

    Secondly: How does mr.Robot model /use trinkets? I guess Mr.Robot probably takes a (15sec each 3 mins =) 8,33% Ascendance uptime into account as overall average, but that shouldn't be the case with /use trinkets. You can line the /use trinket up exactly with Ascendance, making the Ascendance uptime for that /use much higher than with a random proc trinket. For instance, for Static Caster's Medaillion's /use (15sec each minute), every third use has full Ascendance line-up (Ascendance = 15sec each 3mins), meaning there's not 8,33% but 30-33% Ascendance uptime for that /use. That makes the /use significantly more powerful, because DPS is practically doubled while Ascendance is up, hence increased /use score. The /use's score should not be calculated with the standard 8,33% average Ascendance uptime, but usually a much higher one, depending on the cooldown of the /use. As you can see it can go from 8,33% all the way to 30%+.

    Static Caster's Medallion
    1152 Haste * 1.65 = 1901
    1022 average Int from Proc * 3.74 = 3822
    TOTAL SCORE = 5723

    Lets take into account Ascendance has about 30% uptime (33,33% in theory) instead of 8,33% on the /use of 3838 intel (I thought 15sec out of 60sec +3838 intel is 959 average intel by the way, not 1022), because you can have full Ascendance uptime every third use. Then, instead of 959 average intel (15/60*3838) it's practically 959/1,0833*1,3 = 1150 effective average intel due to the much higher Ascendance uptime. Not sure I used the right calculation, but it's at least much closer than a 8,33% Ascendance uptime. This would make >>

    Static Caster's Medallion
    1152 Haste * 1.65 = 1901
    1150 average Int from Proc * 3.74 = 4301
    TOTAL SCORE = 6202

    If I'd use 1022 average intel (how did mr.robot come up with that?) and then take the higher Ascendance uptime into account it would be 1022/1,0833*1,3 = 1266 effective average intel from the /use, making a TOTAL SCORE = 6636 Now, better than Essence of Terror 496... Crazy isn't it. So what's it gonna be?
    Last edited by Skippert; 2012-12-12 at 02:55 PM.

  6. #6
    Quote Originally Posted by Skippert View Post
    Zoopercat, first off thanks for mr.robot is an amazing site. I'll probably disagree with 2 of your trinket calculations though.

    First off, you should use the 466 ilvl Zen Alchemist Stone in the comparisons, because all serious raiders have the upgraded version (+8ilvl). It's practically a free upgrade: just 1500JP, so it doesn't hinder any gearup progression for most raiders at this point, unlike VP upgrades.

    Zen Alchemist Stone (466)
    871 Mastery * 1.46 = 1271
    1249 average Int from proc * 3.74 = 4671
    TOTAL SCORE: 5942 instead of 5516. Significant difference.

    QUESTION: How does mr.Robot model /use trinkets? I guess Mr.Robot probably takes a (15sec each 3 mins =) 8,33% Ascendance uptime into account as overall average, but that shouldn't be the case with /use trinkets. You can line the /use trinket up exactly with Ascendance, making the Ascendance uptime for that /use much higher than with a random proc trinket. For instance, for Static Caster's Medaillion's /use (15sec each minute), every third use has full Ascendance line-up (Ascendance = 15sec each 3mins), meaning there's not 8,33% but 30-33% Ascendance uptime for that /use. That makes the /use significantly more powerful, because DPS is practically doubled while Ascendance is up, hence also the score of that proc. The /use's score should be calculated with the standard 8,33% average Ascendance uptime, but a much higher one, depending on the cooldown of the /use. As you can see it can go from 8,33% all the way to 30%+.

    Static Caster's Medallion
    1152 Haste * 1.65 = 1901
    1022 average Int from Proc * 3.74 = 3822
    TOTAL SCORE = 5723

    Lets take into account Ascendance has about 30% uptime (33,33% in theory) instead of 8,33% on the /use of 3838 intel (I thought 15sec out of 60sec +3838 intel is 959 average intel by the way, not 1022), because you can have full Ascendance uptime every third use. Then, instead of 959 average intel (15/60*3838) it's practically 959/1,0833*1,3 = 1150 effective average intel due to the much higher Ascendance uptime. Would make >>

    Static Caster's Medallion
    1152 Haste * 1.65 = 1901
    1150 average Int from Proc * 3.74 = 4301
    TOTAL SCORE = 6202

    Ofcourse it's a rough estimate, but it's at least much closer than when modelled with a 8,33% Ascendance uptime on the proc. Huge difference. Now, Static Caster's better than Light of the Cosmos LFR, because it's a /use trinket with much higher Ascendance uptime on the /use than on Light of the Cosmos' more random proc. So, it's quite important to calculate /use score in another way than 'random' procs.
    The issue with this is that not all elemental shammys line up /use trinkets with Ascendance. The trinket values have to be taken with no outside factors affecting them else others (those in this example who are not using Ascendance in conjunction with their 'on-use' trinket) will get some pretty crazy results when looking at trinket lists. I love AMR but have always been wary of how they rank trinkets and usually do my own research when choosing which two I want and then manually adjust the AMR trinket list for my own needs.

  7. #7
    The issue with this is that not all elemental shammys line up /use trinkets with Ascendance
    Don't you think the people that regularly visit mr.robot are generally the people that care about min/maxing, hence also care to line up trinket /use with major DPS CD's? Would not be that weird to at least have the option (like there's many options at mr.robot) to "calculate on-use trinket score as if used in line with (major) DPS cooldowns" or something.

    Ofcourse it can be hard to code. If it's impossible to code or if there's no resources to do so, it would be a good idea to add something like a disclaimer for /use trinket score calculations, as they can be so far off, that they can change up the scorelist significantly.
    Last edited by Skippert; 2012-12-12 at 03:30 PM.

  8. #8
    Quote Originally Posted by Skippert View Post
    Don't you think the people that regularly visit mr.robot are generally the people that care about min/maxing, hence also care to line up trinket /use with major DPS CD's? Would not be that weird to at least have the option (like there's many options at mr.robot) to "calculate on-use trinket score according when used in line with (major) DPS cooldowns" or something.

    Ofcourse it can be hard to code, but with the above example you see it can make a HUGE difference and significantly shift around /use trinkets in the BiS list. The difference is so big, that at least something like a disclaimer for /use trinkets should be added, because the current calculations are far off for people that line-up /use procs with major DPS CD's.
    people who go to mr robot dont care about true minmaxing otherwise they would just use simcraft, which is infinitely more accurate than mr robot.

  9. #9
    Quote Originally Posted by Sarevokcz View Post
    people who go to mr robot dont care about true minmaxing otherwise they would just use simcraft, which is infinitely more accurate than mr robot.
    for sure it is. but AMR is some quick lookup or rough guide what item benefits you more. I use both. ARM for first lookup when something drops (or before something drops to get an idea what to hunt for) and later SIMC to fine tune as reforging tools tend to reforge all or nothing to one stat which is not optimal for elemental.

  10. #10
    I would seriously doubt any system that sees Cosmos ranked higher than Terror.

    Using the LFR versions.
    Essence of Terror
    Int: 1021
    Haste: 6121 for 20 seconds, 105 second cooldown, 15% chance. 1166 haste on average

    Light of the Cosmos
    Haste: 956
    Int: 2866 for 20 seconds, 45 second cooldown, 15% chance. 1273 int on average

    The primary issue with the cooldowns is that since Cosmos is a dot proc, while Terror is a hit proc, Cosmos is less likely to proc. We can therefore adjust the cooldowns to 110 & 55 seconds.

    Terror: 1021 Int, 1113 Haste
    Essence: 1041 int, 956 Haste
    Difference (Terror - Essence): -20 Int, 157 haste

    The difference gives 7.85 haste for every 1 Int, and there's no way that Int is ever going to be worth that many multiples of haste. Therefore Terror is better and we don't even have to look at the stat weights (which, incidentally, are incorrect).
    Sentry Totem: The new community news and advice site!

  11. #11
    Quote Originally Posted by binkenstein View Post
    I would seriously doubt any system that sees Cosmos ranked higher than Terror.

    Using the LFR versions.
    Essence of Terror
    Int: 1021
    Haste: 6121 for 20 seconds, 105 second cooldown, 15% chance. 1166 haste on average

    Light of the Cosmos
    Haste: 956
    Int: 2866 for 20 seconds, 45 second cooldown, 15% chance. 1273 int on average

    The primary issue with the cooldowns is that since Cosmos is a dot proc, while Terror is a hit proc, Cosmos is less likely to proc. We can therefore adjust the cooldowns to 110 & 55 seconds.

    Terror: 1021 Int, 1113 Haste
    Essence: 1041 int, 956 Haste
    Difference (Terror - Essence): -20 Int, 157 haste

    The difference gives 7.85 haste for every 1 Int, and there's no way that Int is ever going to be worth that many multiples of haste. Therefore Terror is better and we don't even have to look at the stat weights (which, incidentally, are incorrect).
    Ay-ohhhhhh.

    This.

Posting Permissions

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