5.2 Agility Trinkets

[Zeherah]

Unless they fix it not proccing on Explosive Shot, it most certainly will not be BiS for Survival.

For what it's worth, with the way RPPM works, the frequency of abilities used which can proc the trinket should not have a significant effect on the trinket uptime. The proc chance scales with the last chance the trinket had to proc, so it helps compensate for less frequent proc chances. That said, I'm fairly sure the explosive shot not proccing thing is a bug and should be fixed. I do find it very strange that RPPM trinkets proc off of Kill Command since that is treated as a pet ability.

[Conjor]

For what it's worth, with the way RPPM works, the frequency of abilities used which can proc the trinket should not have a significant effect on the trinket uptime. The proc chance scales with the last chance the trinket had to proc, so it helps compensate for less frequent proc chances. That said, I'm fairly sure the explosive shot not proccing thing is a bug and should be fixed. I do find it very strange that RPPM trinkets proc off of Kill Command since that is treated as a pet ability.

Explosive shot also has an issue with not proccing Misdirection. Been like that since 5.0 IIRC.

[Herecius]

For what it's worth, with the way RPPM works, the frequency of abilities used which can proc the trinket should not have a significant effect on the trinket uptime. The proc chance scales with the last chance the trinket had to proc, so it helps compensate for less frequent proc chances. That said, I'm fairly sure the explosive shot not proccing thing is a bug and should be fixed. I do find it very strange that RPPM trinkets proc off of Kill Command since that is treated as a pet ability.

I feel like it's pretty significant for the sort of trinket that Talisman is, for Survival at any rate, because I've noticed that if you happen to get a LnL proc while it's procced, you are SIGNIFICANTLY less likely to get a second (or third etc.) stack, because you're not hitting stuff that can proc it, and it's only your autoshot that has a chance of proccing it during all of those GCDs. You only get 10 seconds to hopefully refresh that stack, if you get a LnL proc in the first 3/4 seconds of the 10 second proc, you'll be spending a full 4 GCDs of the proc's uptime on shots that are incapable of proccing it. It really sucks.

If I instead spam arcane shot as much as I can, it's noticeably easier to stack. Of course, spamming AS when you're sitting on a LnL isn't exactly smart.

[Conjor]

I feel like it's pretty significant for the sort of trinket that Talisman is, for Survival at any rate, because I've noticed that if you happen to get a LnL proc while it's procced, you are SIGNIFICANTLY less likely to get a second (or third etc.) stack, because you're not hitting stuff that can proc it, and it's only your autoshot that has a chance of proccing it during all of those GCDs. You only get 10 seconds to hopefully refresh that stack, if you get a LnL proc in the first 3/4 seconds of the 10 second proc, you'll be spending a full 4 GCDs of the proc's uptime on shots that are incapable of proccing it. It really sucks.

If I instead spam arcane shot as much as I can, it's noticeably easier to stack. Of course, spamming AS when you're sitting on a LnL isn't exactly smart.

That is not how RPPM works. If you fire a shot that has no chance of proccing it, that is not a "wasted" chance. That chance gets "banked" (as a blue put it) and used later. Anything you see to the contrary is just luck (good or bad).

[Herecius]

That is not how RPPM works. If you fire a shot that has no chance of proccing it, that is not a "wasted" chance. That chance gets "banked" (as a blue put it) and used later. Anything you see to the contrary is just luck (good or bad).

I know how RPPM works, but the issue is that 4 GCDs are spent doing nothing that can proc it. That means that the only thing that can proc it during that time is the auto-shots (which will only go off every 2+ seconds), and with an only 10-second long proc that you want to have proc a second time within its own duration, that's a problem.

Either way though, it really should be proccing from Explosive Shot. It doesn't make sense that it can't, was mostly my point. Wanted to make the bug known.

[Tehstool]

I know how RPPM works, but the issue is that 4 GCDs are spent doing nothing that can proc it. That means that the only thing that can proc it during that time is the auto-shots (which will only go off every 2+ seconds), and with an only 10-second long proc that you want to have proc a second time within its own duration, that's a problem.

Either way though, it really should be proccing from Explosive Shot. It doesn't make sense that it can't, was mostly my point. Wanted to make the bug known.

Those 4 wasted GCDs aren't wasted. Their chances are added up and put onto the next attack that has a chance to proc it. So instead of (arbitrarily picking a number) having a 1% chance on the auto shot you'd have 2% since the attack before that didn't have a chance to proc it.

[Conjor]

I know how RPPM works, but the issue is that 4 GCDs are spent doing nothing that can proc it. That means that the only thing that can proc it during that time is the auto-shots (which will only go off every 2+ seconds), and with an only 10-second long proc that you want to have proc a second time within its own duration, that's a problem.

Either way though, it really should be proccing from Explosive Shot. It doesn't make sense that it can't, was mostly my point. Wanted to make the bug known.

No it isn't a problem. The chance to proc scales with the time between chances to proc the ability. On average it doesn't matter if you attack 10 times per second or once every 10 seconds. The only time that it would make a difference is if the LnL happened at the end of the trinkets proc duration. That could be mitigated by buff monitoring.

[Deleted]

Thought I would share some logs of a few fights in ToT since thurs using the voodoo gnome trinket, has good uptime on a couple of the fights

Ji kun, Durumu and Primordius: http://www.worldoflogs.com/reports/r...?s=8826&e=9226
Dark Animus, Iron Qon and Twin Consorts: http://www.worldoflogs.com/reports/r...=10540&e=11227

The buff is Juju Madness, hope this helps anyone that's interested, My armory is in sig

[Dracodraco]

Simcraft seems to put it at ~25% uptime. That gives the normal mode an average of 1467 secondary, and 1833 agility.

As a comparison, Bottle has a ~42% uptime, for 1652 agility and 1313 secondary stats. So yea, it's a pretty safe bet that Juju (even normal) now comes out ahead of any of the two old-tier trinkets.

The rep-trinket, whose uptime seems to be ~20%, is 1760 agility and 1467 secondary.

However, as the VP trinket is on a set ICD, and Juju is on PPM, it means that Juju can pull even further ahead of the VP trinket in lucky scenarios. You can also game your pots, CD's, etc around the VP trinket - you can't with Juju.


Blades of renataki's uptime seems to be around 20-25% - assuming a 10 minute fight with a ~25% uptime, that's 140 seconds (7 procs).
Agility per is:
1333 2
2666 4
3999 6
5332 8
6665 10
7998 12
9331 14
10664 16
11997 18
13330 20

for a total of 7332 agility average over 20 seconds.
7332*0.25 (uptime)=1832

Meaning renataki's is on average, surprisingly close to the Juju. The strength of the trinket is, of course, that it provides an absolutely HUGE amount of agility towards the end, making anything you line up with it (cds, pot, bloodlust etc) that much stronger. Personally, I wouldn't say this is worth inoting for a hunter, though. Mainly, because as a Surv hunter, you have one cooldown to line it up with (rapid fire), and if you're an orc, Blood fury. That's... Not alot.
Another reason would be the fact that it is a completly random proc, meaning you have no chance to know WHEN it'll happen - you can't delay rapid fire to boost the proc. Which means that it's also mostly useless for BM, as you can't delay BW for the proc.

Assuming the same behaviour for heroic mode trinkets, we're looking at 8278 agility average on a 25% uptime on renataki's for 2070, and we're looking at 2070 agility average from bad juju - the exact same.

So, those numbers in consideration, trinkets in order from best to worst:

Bad Juju HC (mastery > expertise on juju+renataki's, therefore juju is the best).
Renataki's HC
Bad Juju Norm
Renataki's Norm
Shado-pan Assault trinket
Bottle of infinite stars
Essence of Terror
Darkmoon Card
(Note: having an upgrade on either of the 3 last trinkets will put it infront of the others, assuming they do not have an upgrade. Meaning upgraded DMC trumphs non upgraded essence, for example, but fully downgraded and fully upgraded they remain at the same positions).
Throw thunderforged in between if you want, I can't be bothered.


You'll notice that I haven't done anything regarding the rune of reorigination. Blizz has (supposedly) hotfixed it to actually be usefull to hunters (no more 0.08 PPM, yay!), but as our stats are so extremely close to each other in value, I find it hard to imagine that it'll be better than any of these trinkets (especially as increasing, say, crit rating while lowering mastery will mean that you MIGHT get more crits, but they will hit for less). I also haven't been able to find any logs of the trinket in use.

But with a little napkin math - assuming that it works off of the same 25% uptime as the other two trinkets (which I'm not sure it does, I really haven't looked much into it), then we have a base 1467 agility (almost 400 lower than the other trinkets).

Using my own gear as an example (currently ranked in the top 30 west gearwise, so this is much more than most of the raiding population will get out of it):

7920 crit.
4335 mastery.
3353 haste.

4335+3353=7688*2=15376.

This is 7688 secondary stats on a 25% uptime, giving you 1922 secondary stats average. Basically? Not worth it, atleast in my eyes. Everything we have updates dynamically, and our stats are way too closely related to gamble away and expect a gain. Not gonna include it in the above list, because I have no idea if I'm correct.

Any comments?

[Tehstool]

Simcraft seems to put it at ~25% uptime. That gives the normal mode an average of 1467 secondary, and 1833 agility.

As a comparison, Bottle has a ~42% uptime, for 1652 agility and 1313 secondary stats. So yea, it's a pretty safe bet that Juju (even normal) now comes out ahead of any of the two old-tier trinkets.

The rep-trinket, whose uptime seems to be ~20%, is 1760 agility and 1467 secondary.

However, as the VP trinket is on a set ICD, and Juju is on PPM, it means that Juju can pull even further ahead of the VP trinket in lucky scenarios. You can also game your pots, CD's, etc around the VP trinket - you can't with Juju.


Blades of renataki's uptime seems to be around 20-25% - assuming a 10 minute fight with a ~25% uptime, that's 140 seconds (7 procs).
Agility per is:
1333 2
2666 4
3999 6
5332 8
6665 10
7998 12
9331 14
10664 16
11997 18
13330 20

for a total of 7332 agility average over 20 seconds.
7332*0.25 (uptime)=1832

Meaning renataki's is on average, surprisingly close to the Juju. The strength of the trinket is, of course, that it provides an absolutely HUGE amount of agility towards the end, making anything you line up with it (cds, pot, bloodlust etc) that much stronger. Personally, I wouldn't say this is worth inoting for a hunter, though. Mainly, because as a Surv hunter, you have one cooldown to line it up with (rapid fire), and if you're an orc, Blood fury. That's... Not alot.
Another reason would be the fact that it is a completly random proc, meaning you have no chance to know WHEN it'll happen - you can't delay rapid fire to boost the proc. Which means that it's also mostly useless for BM, as you can't delay BW for the proc.

Assuming the same behaviour for heroic mode trinkets, we're looking at 8278 agility average on a 25% uptime on renataki's for 2070, and we're looking at 2070 agility average from bad juju - the exact same.

So, those numbers in consideration, trinkets in order from best to worst:

Bad Juju HC (mastery > expertise on juju+renataki's, therefore juju is the best).
Renataki's HC
Bad Juju Norm
Renataki's Norm
Shado-pan Assault trinket
Bottle of infinite stars
Essence of Terror
Darkmoon Card
(Note: having an upgrade on either of the 3 last trinkets will put it infront of the others, assuming they do not have an upgrade. Meaning upgraded DMC trumphs non upgraded essence, for example, but fully downgraded and fully upgraded they remain at the same positions).
Throw thunderforged in between if you want, I can't be bothered.


You'll notice that I haven't done anything regarding the rune of reorigination. Blizz has (supposedly) hotfixed it to actually be usefull to hunters (no more 0.08 PPM, yay!), but as our stats are so extremely close to each other in value, I find it hard to imagine that it'll be better than any of these trinkets (especially as increasing, say, crit rating while lowering mastery will mean that you MIGHT get more crits, but they will hit for less). I also haven't been able to find any logs of the trinket in use.

But with a little napkin math - assuming that it works off of the same 25% uptime as the other two trinkets (which I'm not sure it does, I really haven't looked much into it), then we have a base 1467 agility (almost 400 lower than the other trinkets).

Using my own gear as an example (currently ranked in the top 30 west gearwise, so this is much more than most of the raiding population will get out of it):

7920 crit.
4335 mastery.
3353 haste.

4335+3353=7688*2=15376.

This is 7688 secondary stats on a 25% uptime, giving you 1922 secondary stats average. Basically? Not worth it, atleast in my eyes. Everything we have updates dynamically, and our stats are way too closely related to gamble away and expect a gain. Not gonna include it in the above list, because I have no idea if I'm correct.

Any comments?

Yeah, never looked at it like that, doesn't look worth it since you are losing stats too.

[Deleted]

So, those numbers in consideration, trinkets in order from best to worst:

Bad Juju HC (mastery > expertise on juju+renataki's, therefore juju is the best).
Renataki's HC
Bad Juju Norm
Renataki's Norm
Shado-pan Assault trinket
Bottle of infinite stars
Essence of Terror
Darkmoon Card
(Note: having an upgrade on either of the 3 last trinkets will put it infront of the others, assuming they do not have an upgrade. Meaning upgraded DMC trumphs non upgraded essence, for example, but fully downgraded and fully upgraded they remain at the same positions).
Throw thunderforged in between if you want, I can't be bothered.

Where does the Talisman of Bloodlust fit in here? Better or worse than Rentaki's?

[Dracodraco]

Where does the Talisman of Bloodlust fit in here? Better or worse than Rentaki's?

I'd need a log with the Talisman post-buff to be sure. Dela tested it before the buff went live, and we found it gave ~1300 haste on average, I think.
Generally, though, Talisman SHOULD be the lowest of all the trinkets - it has haste (our worst stat) and it has "only" 1467 agility for normal (almost 400 less than the proc trinkets). Not even sure if the heroic mode one would beat out the other two normal modes, but I'd have to say no, without actually knowing the math - even then, with our relationship to haste, I'd not take the trinket unless absolutely everyone else has it.

[Deleted]

I'd need a log with the Talisman post-buff to be sure. Dela tested it before the buff went live, and we found it gave ~1300 haste on average, I think.
Generally, though, Talisman SHOULD be the lowest of all the trinkets - it has haste (our worst stat) and it has "only" 1467 agility for normal (almost 400 less than the proc trinkets). Not even sure if the heroic mode one would beat out the other two normal modes, but I'd have to say no, without actually knowing the math - even then, with our relationship to haste, I'd not take the trinket unless absolutely everyone else has it.

With with how RPPM scales with hastes, and that our Meta gem, 2 set, and 2 trinkets will be using the RPPM formula wouldn't that make haste higher than mastery once all those are obtained?

[Dracodraco]

With with how RPPM scales with hastes, and that our Meta gem, 2 set, and 2 trinkets will be using the RPPM formula wouldn't that make haste higher than mastery once all those are obtained?

I doubt it. Remember that while the procs scale with haste, most of the damage you push out scale with mastery - meaning an agility proc with 5% more mastery will do more damage than if those were 5% haste, but it'll happen less often. Basicly, going into one stat will affect the other negatively.

[Rashomon Gate]

fully agree with this. i think same about haste, since 2t15 benefits from haste, and in t15 content even in BM additional atack of dire beast better better than you will reforge haste into mastery

[Conando]

Bad Juju HC (mastery > expertise on juju+renataki's, therefore juju is the best).
Renataki's HC
Bad Juju Norm
Renataki's Norm
Shado-pan Assault trinket
Bottle of infinite stars
Essence of Terror
Darkmoon Card
(Note: having an upgrade on either of the 3 last trinkets will put it infront of the others, assuming they do not have an upgrade. Meaning upgraded DMC trumphs non upgraded essence, for example, but fully downgraded and fully upgraded they remain at the same positions).
Throw thunderforged in between if you want, I can't be bothered.

I'd tend to agree. I honestly think I'm just going to go with whichever one gives out the most stats on average. Rentataki MIGHT end up having some bad RNG (as losing out on the last few seconds is a huge DPS loss), but I'm not sure that will be any kind of dealbreaker.

The only thing I'd somewhat disagree with is that I'm fairly neutral on which secondary stats they give, provided that it's not the last stat on your priority list (let's say haste). The reason being is that the difference will often be lessened by reforging - you either have a mastery trinket that causes you to reforge other pieces into expertise or an expertise trinket that allows to to reforge other pieces to mastery.

That's not to say that there's no difference, it's not a perfect circle like that, but it makes me way less concerned about the secondary stats on a trinket.

[wimp]

So many agi trinket threads to keep track of... reposting here some stuff I posted on EJ earlier.

I think we're getting there with the modelling of bad luck protection and proc overlaps. This is what I'm working with at the moment:

Let the average proc rate, avg_rate = (RPPM*(1+Haste))/60

For trinkets with ICD: proc overlap is obviously not a concern here. The uptime is just [proc duration / average time between procs]. Average time between procs is [1/avg_rate] + ICD

For trinkets without ICD: here we can't just measure proc uptime by [proc duration / average time between procs], because procs can overlap. Instead, we can model the RPPM mechanic as a Poisson distribution, which gives us an uptime = 1-EXP((-avg_rate*Proc_duration))

Bad luck protection:
There are two approaches here. One is to multiply the average proc time by 13.07%. For the mathematical derivation of this assertion, please see this link to work by Jay (@Hamlet EJ): pbs.twimg.com/media/BFRHvJ4CQAANzmS.jpg:large

There is a second approach, which uses mathematica to get a formula which takes into account bad luck protections (and so eliminates the need to multiply by 13.07%, which feels like a bit of a fudge). The approach here is to integrate the equations over the two time periods (the 'regular time' and the 'time after bad luck protection kicks in'). Conjor is working on that at the moment (links to his thread and work above on this thread), but for now multiplying by 13.07% seems more accurate to me.

Using the math above, and assuming 23% haste (rogue here), I'm getting the following average uptimes:

Bad Juju: 22.5%
Renataki's: 22.7%
Talisman: 53.5%
RoRo: here the uptime increases with ilvl. 16.9% for 541, 16.3% for 535, 15.6% for 528 and 15.1% for for the 522 one.

I'm not providing a ranked trinket list, as that would depend on your EP values. E.g. the rogue ranking would be different to the hunter ranking.

Any and all feedback or corrections on the math are very welcome.

[Zeherah]

Wimp, I'm a little confused by the bad luck protection part. Is that meant to cover the bad luck protection from the initial portion of the fight when the proc is greatly increased by having been out of combat, or is it meant to compensate for the effects of bad luck at any point later or does it cover both cases?

[wimp]

Wimp, I'm a little confused by the bad luck protection part. Is that meant to cover the bad luck protection from the initial portion of the fight when the proc is greatly increased by having been out of combat, or is it meant to compensate for the effects of bad luck at any point later or does it cover both cases?

Ah, I'm afraid only the latter: it compensates for the bad luck protection during the fight. A lot of this modelling has been done with 'infinite fights' in mind, in which case the higher chance to proc on pull doesn't matter. But I suppose it will matter for actual fights, which don't tend to go on forever... I don't think you will be able to include this in a formula though, since the percentage of the chance to proc on pull will depend on how long you waited between your last proc and the pull. But what we can work out is how long you should wait in order to have good chances/certainty of a proc on the pull (and I believe Conjor has detailed that maths, so I won't repost).

[Tehstool]

Wimp, I'm a little confused by the bad luck protection part. Is that meant to cover the bad luck protection from the initial portion of the fight when the proc is greatly increased by having been out of combat, or is it meant to compensate for the effects of bad luck at any point later or does it cover both cases?

Covers both cases according to the blue post.