[2.1 Mechanics] Spirit Generators: Quick Ref.

Monk
Prev 1 23 24 25 26 Next
Donchu have timber to harvest or an igloo to build?
What is the optimal IAS to achieve from gear? Please check out the gear from my Bulvairules toon in my profile. I have IAS on both weapons, my Witching Hour, and my gloves. Is that overkill for hitting any optimal breakpoints?

I'm just wondering if it would be prudent to possibly switch the Radiance rune from Dashing Strike to Blinding Speed for increased survivability if I already have enough IAS from gear. An alternative to that would be to drop Alacrity for something else or even the Mantra of Retribution. Please note that the build in my profile is for running group in high Grifts.
02/05/2017 09:03 AMPosted by Vox
Quad gen (Q,WFF,BW,SG


Hmmm is there any specific reason you went with BW and SG instead of foresight and Mangle

or added a fourth element C F L H P ... was less damage than adding... I mean tho hard to get a decent lightning or holy and still use WFF
02/22/2017 06:27 PMPosted by Darth
02/05/2017 09:03 AMPosted by Vox
Quad gen (Q,WFF,BW,SG


Hmmm is there any specific reason you went with BW and SG instead of foresight and Mangle

or added a fourth element C F L H P ... was less damage than adding... I mean tho hard to get a decent lightning or holy and still use WFF


Foresight is bad bc it costs a lot of attacks to maintain. BW far better as a support gen. And BW helps WD dps.

SG for dmg during fire.

I'm not sure I understand your last question. A fourth gen provides a minimum of 10% DIBS through combo strike. If it is a bit more dps during its COE flavor, then it probably makes sense over something like FITL.
02/05/2017 09:05 AMPosted by Vox

1. Cold WFF - Tri gen (WFF,Q,BW,FITL)

[/quote]

Can you explain why you're using two physical runes? Wouldn't it be better to get 3 CoE rotations vs 10% more damage for 3 seconds? I see some people running WFF, Q, SG, BoH to get 3 uses of the CoE rotation.
02/24/2017 03:28 AMPosted by BaesedGod
02/05/2017 09:05 AMPosted by Vox

1. Cold WFF - Tri gen (WFF,Q,BW,FITL)



Can you explain why you're using two physical runes? Wouldn't it be better to get 3 CoE rotations vs 10% more damage for 3 seconds? I see some people running WFF, Q, SG, BoH to get 3 uses of the CoE rotation.[/quote]

CW is more WD DPS. On ST burns, it is used (tapped 1 time) only right after a dash so that it never replaces the high dmg 3rd strike of your dps gen.

It isn't necessarily better or worse than a third element dps gen. Maybe someone has mathed/modeled it out that a DPS gen is better on avg?
02/24/2017 03:28 AMPosted by BaesedGod
Can you explain why you're using two physical runes? Wouldn't it be better to get 3 CoE rotations vs 10% more damage for 3 seconds? I see some people running WFF, Q, SG, BoH to get 3 uses of the CoE rotation.

Maybe he's simulating the case where you weave in BW hits into your other elements as well?

That said, I don't think ghom testing is a good way to test things like these, because breaking wave is much better in solo than in groups - in groups you already have lots of additive skill damage, so breaking wave adding a bit more is almost irrelevant. Solo, you get a much bigger boost from it. A better way to test the benefits from BW is with something like d3planner, although annoyingly a lot of things are still bugged there.

I think in practice, the cold/physical/fire trigen (BFF,Q,SG) is still going to be the best, although I question the inclusion of BoH instead of BF-FitL. I don't seem to have spirit problems even without quickening, testing this spec on my own. But I have no experience playing it in groups (I'm a solo player, sadly) so I can't test for certain. Lack of a target dummy really hurts.
I'm wondering if Davlok could add an analysis and/or comment to this thread on the speed of stacking Bane of the Stricken with the various generators. Is it just based on the total frames per attack, so a generator with fewer total frames over the three segments of the attack would stack Stricken faster?
02/25/2017 10:54 AMPosted by Set
I'm wondering if Davlok could add an analysis and/or comment to this thread on the speed of stacking Bane of the Stricken with the various generators. Is it just based on the total frames per attack, so a generator with fewer total frames over the three segments of the attack would stack Stricken faster?


Sorry, I have been slacking on testing mechanics lately, especially on the generator monk side since I don't play gen monk so my only recommendations are based on some out-of-context mechanics testing on PTR and some excel plotting.

The one artifact that was brought up recently was Thunderclap. For those of us old-school monks that played back in the day, know that TC was essentially the ONLY viable generator from launch till RoS. But with the major generator changes (giving TC's teleport to all FoT runes, etc) everyone forgot (including me) that the 3rd strike has always been a 'double-hit' since launch in that, the "clap" of the 3rd hit also hits the main target. (the first 2 single target strike's "claps" did not hit the double-hit) So the weapon damage amount on a single target is actually 200% + 200% + 400%/1 + 120% = 920% weapon damage per combo. This is slightly higher than Windforce Flurry's 880% so I had to add TC to the Single Target DPS chart, linked below ^_^

2.4.3 Single Target DPS Chart
http://i.imgur.com/t2yCyPd.png

Looking at the 4p leaderboards now, seems like Q/WF/SG is the top generator loadout at the moment, so I have doubts TC/FoF would displace Q/WF which looks to have a more useful spirit/aoe combination... but who knows, maybe if gen monks are still used in S10 we can find out!

As far as Stricken stacking, it looks like on paper if we are to accept the BotS ICD = CEILING ( .9 * Frame Per Last Attack ) formula, all generators except Fists of Fury are only adding 2 stacks every 3 strikes at nearly all breakpoints. The reason is the 3rd strike for every generator takes longer to "cast" than the first two strikes so you will essentially only be applying BotS stacks on the 2nd and 3rd strike every cycle since the ICD from the 3rd strike will still be counting down when the 1st strike lands. So theoretically, Fists of Fury should be the generator that stacks BotS the fastest because the constant 12 frame DoT can trigger between the 2nd and 3rd strikes, refreshing the faster ICD from Strike 1 or 2 instead of using the 3rd strike ICD. On my napkin, it yields between 15-25% more BotS stacks over time. The one anomaly I did notice, is that CW seems to be applying BotS stacks on all 3 hits, almost as if it is not applying the longer ICD on the third strike; or doing something else funky like that ^_^;
^ Do you have a constant damage setup (2-3 damage weapon and +1 min damage legacy ring)?

Because if you do, another interesting thing you could measure is the effective BotS stacks over time in practice, by comparing the damage numbers. i.e. using your stock generator setup, hit a mob for a minute and see how many stacks the numbers go up by after that minute. You could also analyze the individual generator hits to see if they're all providing a higher number than the previous or not.

I don't know if it's still possible to create such a setup, but I know I used to have one back in season 1. I might still have an adequate ring in my non-season stash.
So you're sure TC's on top now? That's interesting. Why are you pairing FoF with it though?

Either way, this is good. I like having a choice in the gen to use. Good SoJ and elemental TP's are difficult to get, so the more elemental choices that are viable the better it is for players. Looks like both cold and lightning are now quite viable at least.

I mean what's interesting is that pretty much any Fists of Thunder rune is viable as the generator right now. They all do more than mangle, and people have been getting by with mangle... so... yeah. I'm looking forward to seeing what people do. I know some high position leaderboards on season have even been using physical gen with FoT's quickening rune. Like the guy literally has physical on every slot. Wonderful variances.
And double post but I wanted to ask about solo play. Right now I think most people are using Mangle for soloing? I know why that is, but I've always hated Mangle and it's crappy FD proc rates (and trash for LoH). When I learned about holy gen, I liked it much better because it felt much more stable. I know FoF is totally nerfed now, but is there a gen that kind of strikes the middle line between Mangle and FoF (pre-nerf) for solo play? I don't care if it does less damage, I just hate Mangle.
02/26/2017 09:59 AMPosted by nand
Do you have a constant damage setup (2-3 damage weapon and +1 min damage legacy ring)?

I do, I usually only do that to verify multipliers. For APS with frame counting, I can record at 60 fps, but have a 144hz monitor and not sure if something weird happened going from 32 to 64 bit or whatever they did to support > 60 fps monitors.

02/26/2017 10:37 AMPosted by StoleOwnCar
So you're sure TC's on top now? That's interesting. Why are you pairing FoF with it though?

TC is only the top generator for SINGLE target. As soon as you are in combat with more than 1 target, you're better off using any other generator. So I'd still be in the Q/WF/SG boat overall.

As for FoF, it was more of a "how to stack BotS as fast as possible" option.
02/26/2017 10:37 AMPosted by StoleOwnCar
Either way, this is good. I like having a choice in the gen to use. Good SoJ and elemental TP's are difficult to get, so the more elemental choices that are viable the better it is for players. Looks like both cold and lightning are now quite viable at least.

I agree. R6 group gen is the most interesting spec in the game at the moment because there are so many open questions, variables and configurations you can try.

The biggest open question I have right now is the spirit. Since I play R6 gen only on paper (no group to test with), I don't know what spirit management in group GRs is like. d3planner tells me it's best to just hold down WFF regardless of your spirit level. Even if I specifically tell it to switch to quickening whenever my spirit drops below 90% (and I'm not in some other generator's element), it's less DPS than just holding down WFF during CoE downtime. But I imagine this happens mainly because d3planner doesn't simulate the BotS ICD correctly. I may have to write my own dedicated simulator after all. (I held off on the plan when discovering d3planner's sim feature)

There's also the consideration of the proc coefficients and FD uptime. Higher proc coefficient = higher FD uptime = better DPS, in principle.

For APS with frame counting, I can record at 60 fps, but have a 144hz monitor and not sure if something weird happened going from 32 to 64 bit or whatever they did to support > 60 fps monitors.

I would imagine that the 60 frame time base comes from the server's calculations; so it should be independent of the client framerate. That said, I don't know how this interaction works. Is the animation cooldown enforced client-side?

Could the FPS of your game possibly affect your attack speed / damage output? What if you modified your client to remove animations, or did something like CE speedhacking to it? Fun to think about, but not worth getting banned just to try it. :p

That said, I think that the right way to measure stuff like this is not by counting frames but by amortizing it over long durations. Compare how many stricken stacks you have after X 3-hit combos with the expected amount. Work from there.

02/26/2017 10:42 AMPosted by StoleOwnCar
And double post but I wanted to ask about solo play. Right now I think most people are using Mangle for soloing?

Honestly when I play R6 gen solo I prefer to just keep using static charge / FoF. I know it's totally nerfed, but it's just so much more fun. :p

Ps. With combination strike, you need to work a single hit of each of your other generators in once per element swap, right?
I may have to write my own dedicated simulator after all. (I held off on the plan when discovering d3planner's sim feature)


I don't suggest it. I looked at that thing's source code and it's a bit of a monster in a lot of ways. It takes a lot of variables into account, and for many builds it actually does it properly. The problem is that it relies on subject matter experts for some of these esoteric game principles. IE like Davlok. So there's probably a lot of stuff in there that just needs tweaks. It's just that it's hard to figure out which spots to edit. I imagine you'd be better off simply taking some time to parse what he did for various skills and then tweaking it and suggesting he merge in your changes.
02/26/2017 05:23 PMPosted by StoleOwnCar
I looked at that thing's source code and it's a bit of a monster in a lot of ways.

That's exactly the problem. It's a big ball of interlocked javascript that won't even let me find the right places to edit. I'm not sure if the “framework” it's built upon even supports these extremely quirky bane of the stricken mechanics. It could be that the simulator was built on the assumption that the game made more sense.

At any rate, when I say “simulator” I usually mean “100 lines of code”. It's just “one step more complex than a mathematical formula”.

Either way, the difficult part is figuring out the exact mechanics first - and that requires scientific experiments. What we do with the results is a much less interesting question.

I may get around to setting up a recording setup and making some frame measurements of my own some day. That said, if anybody else wants to record themselves performing experiments, I'll gladly look at the recorded footage and count the frames - I'm just too lazy to figure out how to record. :p

The most important thing would be getting a constant damage setup, so I can easily use the damage numbers as a reference point. It would also be good to remove as much crit chance as possible, as well as other dynamic effects and procs like flying dragon.
As much as we trash it here for inaccuracies though, I personally think d3planner and its simulator is still impressive. The web site portion especially. The way you can drag drop, drop down, set constraints, etc... it's an impressive and overall pretty user friendly GUI. So I'd prefer to help it succeed if at all possible. I'm personally more of an advocate for helping the author of it work out the kinks, because otherwise it's pretty impressive. But yeah perhaps an overhaul of some tidbits of the engine might be in order. I remember him linking me to the line where he fixed how the Swami went from multiplicative to additive and I was squinting at it for a bit and then... "oh, I see". But then again I've never been that great at reading code, just churning it out. That's why I prefer to just roll my own if some framework doesn't do what I need it to.

But even then, I think we should give d3planner more of a chance. >_>
D3planner is a great tool, the guy that wrote it (Riv) actually dropped by this thread many pages back to provide some of the frames data that I used to develop the 2.4.2 charts.

Edit: nevermind! I figured out the stupid thing I was doing epic failing my APS testing. So just the weird result I had of CW being able to stack BotS the best, which seems to line up with Vox's results. :)
02/26/2017 06:45 PMPosted by Davlok
D3planner is a great tool, the guy that wrote it (Riv) actually dropped by this thread many pages back to provide some of the frames data that I used to develop the 2.4.2 charts.

Yep, not trying to knock it at all. It's a great tool for what it does. But I still think we need both: General-purpose planners/simulators that try to cover as broad as spectrum as possible, and specialized single-purpose simulators for answering specific questions that are too complicated for the general simulator.

One of the problems I have, for example, is that with d3planner I can't perform any sort of automatic optimization. With a dedicated simulator for R6 mechanics we could program it to automatically optimize the setup for the maximum possible RG clear speeds under group circumstances and for a specific rift guardian.

The search space of possible generator combinations is in the ballpark of Sum[ n∈[1,4], binom(4,n) * 5^n * n ] = 4320 setups. This would be trivial for a computer to brute force with a simulation, but very difficult to do by hand with d3planner.

(For the interested: The binom(4,n) term is for the choice of generators, 5^n is for the choice of runes, and n is for which of those generators you want to dedicated your gear rolls to - going by the assumption that both the element and the gen skill% damage will be attuned to this generator)

Join the Conversation

Return to Forum