Scrapper Melee Primary Testing: "Standard" environment

[Yomo Kimyata]

5 hours ago, Galaxy Brain said:

The trouble with 1 person is that it is also just 1 data point unless that person is a master of all sets.

 

What I mean by parity though is like... I dont expect like massive 2 minute average swings unless somebody is really bad vs somebody really good. On average it should wiggle a little but not that much.

I would argue that if you are trying to identify the factor [weapon damage], and one of your variable inputs is also [player speed], you need to isolate and ideally eliminate the player speed factor.  If there is a 30 second swing between player A and dual blades and player B and radiation melee, for example, you need to statistically identify what component is responsible for what.  Maybe the difference between Player A and Player B maxes out at 60 seconds, assuming that both players are experienced CoHers. But your thesis statement [Dual blades is x seconds slower than radiation melee] gets destroyed in the noise.

 

I would expect that a good player would be consistent across most if not all melee sets, though. 

 

If I were paying for this project, I'd want to isolate [player speed].  Maybe I'm wrong and player A and player B would be within a few seconds with the same set, but I doubt it.

 

Just my opinion; you all are doing this on your own time, so you can do it however you want!  I totally appreciate your dedication.

[Galaxy Brain]

@Yomo KimyataI agree, getting 1 person who is somehow equally good at all sets to run through would be the best option, but I don't think we have that at our disposal 😞 

 

Thank you a ton for your feedback @tinytim1183, we are looking into ways to make the enemies more dangerous to make primary mitigation meaningful.

 

Just a note on Primary/Secondary pairings: 9/13 secondaries have some sort of endurance management tool making it somewhat uncommon *not* to have something to mitigate costs. WP sort of gives everything an advantage as it can stack up with any secondary mitigation from the primary fairly equally IMO, and it is common to have +End there too.

 

What I have noticed though is that the Accolades may be providing too much benefit on Pineapple... running the numbers:

 

• 3 slotted QR and Stamina = 3.45 end/sec - 0.69 end drain/sec (4 WP toggles + CJ) = 2.76 net end / sec
• With accolades, this is 3.8/s ---> 3.11/s!
• No accolades, Miracle proc -> 3.01/s
• No accolades, Miracle + Numina -> 3.18/s
• No Accolades, Panacea (~0.375/sec) -> 3.135/s

 

Closest equivalent is having the Panacea Proc slotted into Health with no other IO's with these Accolades active... hmm

 

 

 

 

 

 

 

 

Edited September 21, 2019 by Galaxy Brain

[Hopeling]

@tinytim1183, you were right, herding is still the dominant strategy. I had done it wrong before, but I just clocked 5:30 with Broadsword via aggressive herding, a dramatic improvement over the 6:30-7:00 range I'd been hitting before.

 

I'm giving the <Punching Bags> back their "normal" number of attacks instead of just a single attack. According to a quick test, you should still be in no danger while fighting a single spawn, but 2 spawns together is now right about the limit of what can be survived unless you have mitigation from your primary. So a little bit of herding will still work, but gathering multiple full spawns will be sketchy, especially with bosses.

2 hours ago, tinytim1183 said:

I also don't see how to pick any secondary that wouldn't give at least a few primaries and inherent advantage.

This is definitely true. I'd like to test with Invulnerability and Ninjitsu later, to see if losing bonus endurance or losing a taunt aura scrambles the rankings.

[Infinitum]

4 hours ago, Hopeling said:

@tinytim1183, you were right, herding is still the dominant strategy. I had done it wrong before, but I just clocked 5:30 with Broadsword via aggressive herding, a dramatic improvement over the 6:30-7:00 range I'd been hitting before.

 

I'm giving the <Punching Bags> back their "normal" number of attacks instead of just a single attack. According to a quick test, you should still be in no danger while fighting a single spawn, but 2 spawns together is now right about the limit of what can be survived unless you have mitigation from your primary. So a little bit of herding will still work, but gathering multiple full spawns will be sketchy, especially with bosses.

This is definitely true. I'd like to test with Invulnerability and Ninjitsu later, to see if losing bonus endurance or losing a taunt aura scrambles the rankings.

Are you talking herding around a corner or just gathering/clustering as you go?  thats a crazy big improvment.

[tinytim1183]

That feels much better!! I think it will at least hint at the secondary effects, if not show their value. Definitely feels more "real world". btw, i snicker every time i write "real world", because its still game testing.

 

Next question, do you only want "quality" runs recorded, or runs with deaths as well, and just notate the death. I can see doing this two ways, playing it relatively safe and death never enters the equation, or riding the edge and pushing the limit just to see what these sets can do, which leads to bad things happening, but loads of fun.

 

10 hours ago, Infinitum said:

Are you talking herding around a corner or just gathering/clustering as you go?  thats a crazy big improvment.

 

For me personally, herding in a run on this map was just always fighting the groups two at a time. Levering as much damage as I can from the AoEs.

[Galaxy Brain]

In my re-runs with BA and BS, BS comes out on top due to Parry allowing for way more mitigation than random KD. With BS< I can reliably herd about 2 groups per area and come out on top, where BA is a gamble to do so thanks to the mitigation being random.

 

I think ultimately, herding the whole map up / as much as you can will lose time due to aggro caps and target caps, but being able to safely bunch up 2 groups to mow down at a time appears to be the meta.

[Bill Z Bubba]

18 hours ago, tinytim1183 said:

Third, claws just optimizes i think the best of all sets on its own.

I love this sentence so much.

[Hopeling]

4 hours ago, tinytim1183 said:

Next question, do you only want "quality" runs recorded, or runs with deaths as well,

I'm OK with just excluding runs where you die or otherwise lose a bunch of time for some reason.

[TheAdjustor]

Some times for the data I haven't had much time to test these lately, or for that matter to actually play the game but here are some data points.

 

1st Seeing as I had an Archery/Tactical arrow blaster on Pineapple that I hadn't put any IOs into the build yet I thought just for the sake interest I would load her up with SOs and see how she did

 

Start 2019-09-21 20:30:33 You activated the Snap Shot power.
Dead: 2019-09-21 20:36:19

No Idea how deaths without a self rez should be scored

 

Pulling out the War Mace/ Will Power

First two runs

2019-09-21 20:41:24 You activated the Clobber power.
2019-09-21 20:48:54 You have defeated Heavy Punching Bag

 

2019-09-21 20:59:09 You activated the Clobber power.
2019-09-21 21:06:04 You have defeated Medium Punching Bag

 

Oddly enough the third run wasn't so lucky and he died twice

 

2019-09-21 21:08:59 You activated the Clobber power.
2019-09-21 21:10:12 [Local] TW argument: <color #010101>dead
2019-09-21 21:10:21 You activated the Resurgence power.
2019-09-21 21:14:18 You activated the Resurgence power.
22019-09-21 21:16:17 You have defeated Medium Punching Bag

 

Something which only cost him a minute or so on his time

 

Just a note the blaster was taking groups as they were placed and was not using any immob or hold except electrified net even so progress was rather abysmal.

 

 

 

[TheAdjustor]

On 9/20/2019 at 8:34 PM, Yomo Kimyata said:

I would argue that if you are trying to identify the factor [weapon damage], and one of your variable inputs is also [player speed], you need to isolate and ideally eliminate the player speed factor.  If there is a 30 second swing between player A and dual blades and player B and radiation melee, for example, you need to statistically identify what component is responsible for what.  Maybe the difference between Player A and Player B maxes out at 60 seconds, assuming that both players are experienced CoHers. But your thesis statement [Dual blades is x seconds slower than radiation melee] gets destroyed in the noise.

 

I would expect that a good player would be consistent across most if not all melee sets, though. 

 

If I were paying for this project, I'd want to isolate [player speed].  Maybe I'm wrong and player A and player B would be within a few seconds with the same set, but I doubt it.

 

Just my opinion; you all are doing this on your own time, so you can do it however you want!  I totally appreciate your dedication.

This is much my feeling that each player will have their own time profile depending on their playstyle.

[Galaxy Brain]

Thanks for the results Adjustor!

 

Perhaps there could be a way to isolate player results as a sort of trial? I've already seen that the results can vary depending on the set and playstyle. I can herd with broadsword and not battle axe due to parry, but it seems BA handles individual mobs cleaner so the same style is not applicable between the two. 

 

Maybe a select few sets with varied gameplans should be used as a preliminary to see if we can gather consistent results. If somebody varies we would discuss what they did to try and stick to effective playstyles.

[Sir Myshkin]

I've been away from my PC for a couple days, come back to find this thread and I spark a lot of questions in the first post and see a lot of those questions are being landed on through further experimentation, so I guess that saves me the time of asking a lot of the "But what about this factor..."

 

I think you both have fallen on the biggest dilemma of practicing a "standard" rule. Play style is going to change the dynamic of each interpretation, as well as build choices. Trying to define a static metric with Willpower is understandable, but in the SO game, each player always looked for the best way to bend abilities no differently than we do with IO's/Sets now. Hasten alone dramatically changes the speed a set operates at, even at the SO level. Other mechanics like +Rech in Super Reflexes and Electric Armor, how Fire Melee/Fire Armor plays out so much faster than just Fire/WP. You're talking about trying to rebuild a baseline that the original Dev team had/used years worth of data to benchmark from and still gave birth to Titan Weapons.

 

The short end of the stick, I think this prospect is an interesting endeavor to explore, but it isn't something that can be weighed on from just one study of how a singular defensive set operates with a slew of primary attack sets. This really needs weighed equally for every set and pairing combination to find where (or even if) something abnormally sticks out. At ten runs a piece, 19 primary, 13 secondary, 2,470 tests that we must assume will average somewhere between 5-8 minutes a piece. Right now the one singular test you're looking to complete is already 20-25 hours worth of testing alone.

 

[Galaxy Brain]

I dont think testing every secondary would be necessary as many give bonuses that are found in others, or the secondary sort of acts on it's own and would give similar value to each primary its paired with.

 

I do think SR should be on the table at some point for quickness to see how even a smidge more recharge impacts play, but the reason to stick to WP is to get the least invasive look at the primary set without special factors from the pairings. I think that after WP is done we should work in secondaries on certain primaries to see of performance jumps up for say... two sets with a ~5 sec difference or so?

 

 

Edit: 

 

To account for strategic differences, we could just have the tester write up a quick summary of how they played. Something like "I use parry often to allow me to herd groups up, then cycled attacks from highest to lowest DPA / AoEs till done" could suffice. That way we could at least compare notes and see like "oh, this set was able to be played this way much better, but this other set lacks the tools to do that / can be played this other way".

Edited September 23, 2019 by Galaxy Brain

[TheAdjustor]

2 hours ago, Sir Myshkin said:

I've been away from my PC for a couple days, come back to find this thread and I spark a lot of questions in the first post and see a lot of those questions are being landed on through further experimentation, so I guess that saves me the time of asking a lot of the "But what about this factor..."

 

I think you both have fallen on the biggest dilemma of practicing a "standard" rule. Play style is going to change the dynamic of each interpretation, as well as build choices. Trying to define a static metric with Willpower is understandable, but in the SO game, each player always looked for the best way to bend abilities no differently than we do with IO's/Sets now. Hasten alone dramatically changes the speed a set operates at, even at the SO level. Other mechanics like +Rech in Super Reflexes and Electric Armor, how Fire Melee/Fire Armor plays out so much faster than just Fire/WP. You're talking about trying to rebuild a baseline that the original Dev team had/used years worth of data to benchmark from and still gave birth to Titan Weapons.

 

The short end of the stick, I think this prospect is an interesting endeavor to explore, but it isn't something that can be weighed on from just one study of how a singular defensive set operates with a slew of primary attack sets. This really needs weighed equally for every set and pairing combination to find where (or even if) something abnormally sticks out. At ten runs a piece, 19 primary, 13 secondary, 2,470 tests that we must assume will average somewhere between 5-8 minutes a piece. Right now the one singular test you're looking to complete is already 20-25 hours worth of testing alone.

 

I have to disagree. Just from my own testing there's certain elements that are already very apparent.

1.) AOE >> ST

2.) If you are pushing the envelope having a self res >> everything else

3.) Blasters with just SOs are still kind of crap no matter what you do. (Admittedly I could take hover and float around these maps but I hardly have to test that to know just how badly that would gimp the toon).

4.) Hasten will likely make a gigantic difference

 

If I feel well tomorrow I'll try to take a savage melee/WP scrapper through this test and possible a WP/Energy Melee tank and a WP/Savage Melee Tank. It might be nice to have a baseline on just where EM is after this.

 

The above does bring us back around to the big questions of just how well are things performing ? and how much of FOTM is just madness of the crowds.  I suspect we may find much of what was done in the "holy name of balance" was in fact just the paragon devs rattling our chains.

 

 

 

[Sir Myshkin]

On 9/19/2019 at 11:11 PM, Galaxy Brain said:

MISSION ID 3821: MISSION SIMULATOR

Might want to fix this to 3281. Went looking for the mission and came up empty until I saw the number in the image and realized this was a typo.

 

4 hours ago, Galaxy Brain said:

I dont think testing every secondary would be necessary as many give bonuses that are found in others, or the secondary sort of acts on it's own and would give similar value to each primary its paired with.

May get away with not testing every single one, but there are considerable aspects to each that alter the way an SO build can function in a mission, and how many mobs it can handle at any given time. If I did this same test with Dark Armor I'd have absolutely no concern about herding 16 mobs at once because Oppressive Gloom would eliminate 90% of the damage this test is capable of dishing out, while also shortening my "1 HP" scenarios with Death Shroud. The fact that four secondary sets have damage components in themselves changes the dynamic of this test. Granted you've very specifically chosen just one (to start) here, but it's an important aspect on how things would pair. To boot, Elec, Dark, and Fire all approach combat in different ways. Sap, Mez, and Kill-it-with-fire, respectively. Frankly Elec and Fire would fair better with access to endurance correction tools, but all are still viable in an SO world, especially at base level.

 

3 hours ago, TheAdjustor said:

I have to disagree. Just from my own testing there's certain elements that are already very apparent.

1.) AOE >> ST

2.) If you are pushing the envelope having a self res >> everything else

3.) Blasters with just SOs are still kind of crap no matter what you do. (Admittedly I could take hover and float around these maps but I hardly have to test that to know just how badly that would gimp the toon).

4.) Hasten will likely make a gigantic difference

1.) AoE is a given, and dramatically favored in this kind of style test "clear all." That's one of the biggest things about the Pylon Challenge that's been brought up, how ST take-down doesn't fully explore all avenues of what a build can do. In the realm of a "Clear All" style mission, AoE will always be superior.

2.) I question if Self Rez should be removed from the table as an option here. Only a few sets actually have access to them, and in terms of SO's they were explicitly added for the fact that the set needed to self-revive because it was almost a given it would die. Is the Rez a handicap in this situation where other sets are forced to evaluate a scenario more carefully when the Rez user can just barrel in constantly without concern? It changes the play style by a significant degree. Should the Rez be disqualified from testing, and if you die, that run be considered a DNF. Should we not have a category that exemplifies when a set fails?

3.) The OP of this thread was specifically targeting Scrappers, and as such, any other AT can't really be quantified fairly here because each has their own faults and require their own discussion (IMO).

4.) Yes, it will/does, and I'll have an example of such...

 

 

An Actual Test:

Cause why not, I took out a Claws/WP over on Pineapple and executed the given test. This is the build I put together because I did want to test out using Hasten, and see if I felt force-herding with Provoke might dramatically change things versus the fact that WP naturally taunts in RttC anyway. So I did more than just the base-line tests.

Spoiler

Hero Plan by Hero Hero Designer 2.23
https://github.com/ImaginaryDevelopment/imaginary-hero-designer

Click this DataLink to open the build!

Level 50 Magic Scrapper
Primary Power Set: Claws
Secondary Power Set: Willpower
Power Pool: Fighting
Power Pool: Speed
Power Pool: Leaping
Power Pool: Presence
Ancillary Pool: Body Mastery

Hero Profile:
Level 1: Strike -- Acc(A), Acc(31), Dmg(31), Dmg(33), Dmg(33), EndRdx(33)
Level 1: High Pain Tolerance -- Heal(A), Heal(29), Heal(31), ResDam(46), ResDam(48), ResDam(48)
Level 2: Mind Over Body -- EndRdx(A), ResDam(25), ResDam(25), ResDam(29)
Level 4: Fast Healing -- Heal(A), Heal(5), Heal(5)
Level 6: Slash -- Acc(A), Dmg(7), Dmg(7), Dmg(19), EndRdx(19), RechRdx(23)
Level 8: Spin -- Acc(A), Dmg(9), Dmg(9), Dmg(11), RechRdx(11), RechRdx(17)
Level 10: Indomitable Will -- EndRdx(A), DefBuff(42)
Level 12: Follow Up -- Acc(A), Acc(13), EndRdx(13), RechRdx(15), RechRdx(15), RechRdx(17)
Level 14: Kick -- Empty(A)
Level 16: Rise to the Challenge -- EndRdx(A), Heal(39), Heal(39), Heal(40), Taunt(40), ToHitDeb(45)
Level 18: Focus -- Acc(A), Dmg(36), Dmg(37), Dmg(37), EndRdx(37), RechRdx(39)
Level 20: Quick Recovery -- EndMod(A), EndMod(21), EndMod(21)
Level 22: Tough -- EndRdx(A), ResDam(43), ResDam(45), ResDam(50)
Level 24: Weave -- EndRdx(A), DefBuff(43), DefBuff(43), DefBuff(48)
Level 26: Hasten -- RechRdx(A), RechRdx(27), RechRdx(27)
Level 28: Heightened Senses -- EndRdx(A), DefBuff(40), DefBuff(42), DefBuff(42)
Level 30: Super Speed -- EndRdx(A), EndRdx(46)
Level 32: Shockwave -- Acc(A), Dmg(34), Dmg(34), Dmg(34), EndRdx(36), RechRdx(36)
Level 35: Resurgence -- RechRdx(A)
Level 38: Strength of Will -- ResDam(A)
Level 41: Conserve Power -- RechRdx(A)
Level 44: Physical Perfection -- EndMod(A), Heal(46)
Level 47: Combat Jumping -- DefBuff(A)
Level 49: Provoke -- Acc(A), Acc(50), RechRdx(50)
Level 1: Brawl -- Empty(A)
Level 1: Critical Hit
Level 1: Prestige Power Dash -- Empty(A)
Level 1: Prestige Power Slide -- Empty(A)
Level 1: Prestige Power Quick -- Empty(A)
Level 1: Prestige Power Rush -- Empty(A)
Level 1: Prestige Power Surge -- Empty(A)
Level 1: Sprint -- Empty(A)
Level 2: Rest -- Empty(A)
Level 4: Ninja Run
Level 2: Swift -- Run(A)
Level 2: Health -- Heal(A), Heal(3), Heal(23)
Level 2: Hurdle -- Jump(A)
Level 2: Stamina -- EndMod(A), EndMod(3), EndMod(45)
------------

 

| Copy & Paste this data into Mids' Hero Designer to view the build |
		|-------------------------------------------------------------------|
		|MxDz;1433;559;1118;HEX;|
		|78DA6D946D6FD25014C7FBC42A0CD80ADB188C0D05E7E6061D44DF9BE8D43723594|
		|2E27C260DBB62B5691BC0CCBDF4CBF8D26FE6C337C043CFFFE2DADAD0FCB8FFF3D0|
		|73CFE96DFFEB69FEC7F36F8F14B5F8C473A6D3E1603471C2504C327D67EC8E94C56|
		|5D2BD21F5615F7842D8E47C35AD2DC553F141F853615FB89E17065762A214CE83C0|
		|B39FB9E38F33D71FE7A2D52014E2321FFD3D134E48BA152DFA8EEF865F3C67E606B|
		|EF5347447F6E3E0F27AD877A63331B9DEA6C7EFD3ED9A0AAE792682A12B7B923AF3|
		|769226F3C5822A6273ACD5C13DB053629E247841B12AC7EA2AF26D97E2AC832FC9A|
		|EB3AFA623AF91E027F2C9601F19D4BFA2C75992C4F3B62CE66793C7B18835117B4B|
		|8F3327899875F0150564B936358BBCF70DA647EBD56893F3CC2AF2E625E15BB4E29|
		|47935B217B826A530A74B794DCA1AB2AD21FA5E2ECE03C90AB35D65FAE46F618716|
		|2A686147779394B951C91B5A97D1FDF20A6B1B09EAE4B389696E22FE08D36B833DD|
		|0207B05BE15E96BC4790236C95EE5BAB52AEAA925F8967C76906F07F90E8CF82424|
		|5B64DFE57CEA2E7C3B60486CA0470DF4E88EFE7FB610D3420DEF68DDC4B49AD0DE9|
		|B7CC216DA3EF63EA0F8436887F0FB4E633EC61B748C7E7630C72DCA61B34DB1B187|
		|1A695DF4A48B3A7B92C8593796A79A7E8BB7A7612CCFEAFC2865EDA6945E4A79905|
		|21EA694B394726EE0B1A4A891925DA7D3CCCAFC676EF91550D432D7FEFB9FA6A9E8|
		|8326FB01FEBA195764EDCFCD38F45103DBE05FBF1DA2B6|
		|-------------------------------------------------------------------|

 

My times were pretty consistent; within a +/-5 sec variance of 6:05. A couple of scenarios took me further out, and it brought out an interesting irritation about this test that added unnecessary time (runners).

 

6:20, 6:00, 6:03, 5:55, 6:10, 6:05, 6:01, 6:00, 6:09, 6:08

 

The first run I over extended without consideration of being SO-only. I know the OP was stated that we could invalidate an initial run or two, but I don't think that I should as it reflects the fragility of SO's appropriately. No mulligans. In that run I ended up kicking the bucket and sat there for a couple of seconds debating if I should count that as a failed run, or use Resurgence. To me, rezzing on the spot felt like I'd cheated the test as it showed an inability of the primary to stop from poor play performance. In the consideration of getting a complete time, however, I did resurrect and complete the run to end up with that exaggerated time. However, in my book, that's a "DNF" (Did Not Finish).

 

The other two slight offsets are the fact that I kept running into issues with mobs running from me, despite RttC. I had to back track to find lost targets (not far, generally, but enough that running back around a corner added effort and time). Having two somewhat ranged attacks was a big saving grace in most cases as I could usually tag the runners before they got far, but I couldn't always stop multiples.

 

My process of execution was to run into a mob, hesitate for one tick, then swing into a second while taking a corner or box to collect at, then use Focus > Spin > (Jump) > Shockwave (In Air). This gave me a good PBAoE on the mob, then in the air I could direct the cone downwards and knock the spawns flat, with minor spread. Hit Focus a second time and now have two stacks, ST attack the toughest (yellow or orange con) in the group until Spin was back up 1-2/s later, and reapply. Then just tab-cycle one attack quickly on the remaining mobs. In the event I encountered an orange con Boss, I'd stick my targeting though them, AoE out the spawn and then drag them into the next group and repeat until the Boss was dead to avoid wasting time cycling attacks with nothing around me. RttC needs a spawn to punch-up Regen, so I wanted to stay surrounded as much as possible. For the initial entry-way into that specific map, I'd RttC Taunt the first spawn, then run down the hall to the center of the ramp/split level room and jump up, get the attention of the mobs above, and drag them all down to the bottom to let off AoE's. I regularly used Shockwave as a collection tool for a secondary spawn. I tested collecting two mobs at a time as well as starting in one and pulling a second into it, dwindling down, slowly moving towards the next and keep adding. That technique really only worked in the final area where there's 4-5 spawns relatively close to each other in a circle, separated by a wall and one corner/offset room.

 

Will also note a small thing about endurance, as the Accolades got brought up. For this build in particular I did toss in Physical Perfection to wastefully fill power selections, and even tossed a slot into it, because it has no appreciable effect on the build. Without the accolades, the build recovers so much EPS that I don't even have to slot end redux into Spin. At 9.15 for a cost, and an EPS of 3.45 on its own (no PP/Accolades), and an animation time of 2.64/s (arcana), I recover 9.108 end in that window. Endurance is a non-issue in this scenario. As for the HP side, they add ~7 HP/s net regen at 1 target in RttC, 12 HP/s at 10 targets in range. That's a little more significant, but not sure that really warrants much mental effort for concern since this isn't 100% about survival.

 

I also tried out using Provoke as a collection tool over Shockwave to try and pool groups together. It really didn't change anything compared to just jumping into the vicinity of the spawn and letting RttC aggro and pull them back a hair into the previous spawn.

 

Also additionally tested Hasten on two runs, got a reduced time of 5:05 to 5:15, which was the difference of being able to keep Follow Up double stacked versus a small gap. Without Hasten FU carried a 6.16 cooldown, w/ 4.53. Spin lost 1.6/s of rech time, and SW 4.1/s.

[Heraclea]

If the goal is to simulate actual gameplay, this will not work.  Early and late in the game you run into mobs that do debuffs like -tohit and -recharge.  And, of course, they are fighting back more effectively, making your scrapper drop what he is doing and look to his inspirations tray, or fire off a trouble power like RttC. 

[Hopeling]

@TheAdjustor and @Sir Myshkin, thanks for the data! TA, are those timestamps on the first and last power used in each run (ie, clear times of 7:30 and 6:55 with WM/WP)?

18 hours ago, Sir Myshkin said:

Other mechanics like +Rech in Super Reflexes and Electric Armor, how Fire Melee/Fire Armor plays out so much faster than just Fire/WP. [...] This really needs weighed equally for every set and pairing combination to find where (or even if) something abnormally sticks out.

Since we're looking to compare primaries here, I'm less interested in the secondary as a package and more in testing how the components of each secondary contribute. For example, it would be silly to test Quickness and Lightning Reflexes and IO recharge bonuses as three separate things; those can all just be folded into testing at various levels of global recharge. Similarly, while Power Sink + Energize is not identical to Quick Recovery or Consume or Energy Absorption, an endurance-heavy primary is going to benefit about equally from all of them, so comparing (a set with endurance) vs (a set without endurance) should roughly cover it. Repeat again for taunt auras and damage auras, and we've covered the majority of non-defensive differences between sets.

 

Some secondaries have truly unique effects, but things like Fiery Embrace or Against All Odds should be mostly agnostic to the primary. The big one to worry about, it seems to me, is Burn, since that becomes part of the attack chain.

15 hours ago, TheAdjustor said:

I have to disagree. Just from my own testing there's certain elements that are already very apparent.

1.) AOE >> ST

I consider this one a feature, not a bug. ST should matter, but AoE should matter more, just like it does in most missions.

 

It looks like all the runs so far have also used sets that are at least as good at ST as they are at AoE; multiple bosses and an EB might cause more slowdown with a set like Electric Melee. I'm going to try that later today.

10 hours ago, Heraclea said:

If the goal is to simulate actual gameplay, this will not work.  Early and late in the game you run into mobs that do debuffs like -tohit and -recharge.  And, of course, they are fighting back more effectively, making your scrapper drop what he is doing and look to his inspirations tray, or fire off a trouble power like RttC. 

I agree this won't simulate actual gameplay. The goal is just to get a metric that is less removed from actual gameplay than the existing ones. It won't be be perfect, but will (hopefully) still be more informative than no data at all.

 

Investigating the effects of tohit, recharge, and varying amounts of incoming damage is already on the agenda, and me, Galaxy Brain, and TheAdjustor have talked about some ideas for testing how much each set is affected by disruptions to its attack chain.

Edited September 23, 2019 by Hopeling

[TheAdjustor]

Quote

@TheAdjustor and @Sir Myshkin, thanks for the data! TA, are those timestamps on the first and last power used in each run (ie, clear times of 7:30 and 6:55 with WM/WP)?

That's correct. First activation to last activation.

[Galaxy Brain]

8 hours ago, Sir Myshkin said:

Might want to fix this to 3281. Went looking for the mission and came up empty until I saw the number in the image and realized this was a typo.

Fixed, thanks for the catch!

8 hours ago, Sir Myshkin said:

 

May get away with not testing every single one, but there are considerable aspects to each that alter the way an SO build can function in a mission, and how many mobs it can handle at any given time. If I did this same test with Dark Armor I'd have absolutely no concern about herding 16 mobs at once because Oppressive Gloom would eliminate 90% of the damage this test is capable of dishing out, while also shortening my "1 HP" scenarios with Death Shroud. The fact that four secondary sets have damage components in themselves changes the dynamic of this test. Granted you've very specifically chosen just one (to start) here, but it's an important aspect on how things would pair. To boot, Elec, Dark, and Fire all approach combat in different ways. Sap, Mez, and Kill-it-with-fire, respectively. Frankly Elec and Fire would fair better with access to endurance correction tools, but all are still viable in an SO world, especially at base level.

I very much agree that different secondaries would yield different results. I would love to dive into what the next runs should be using, though that becomes much more dynamic given the goal of singling out primaries. I think we can circle back once we get data on all primaries with WP as the "least invasive" secondary. (Also, probably shorten runs to 5 instead of 10 with multiple pairings in consideration...) 

 

8 hours ago, Sir Myshkin said:

 

1.) AoE is a given, and dramatically favored in this kind of style test "clear all." That's one of the biggest things about the Pylon Challenge that's been brought up, how ST take-down doesn't fully explore all avenues of what a build can do. In the realm of a "Clear All" style mission, AoE will always be superior.

This is a fair assessment. +0/x3 was chosen to give groups that would leverage AoE caps for Melee on occasion. Though to be fair, in most content AoE is already heavily favored due to just sheer amount of enemies on even the lowest setting on a team or even when x1 and you face a couple of minions at once. If there is a better way, maybe +(Number) / x 2 that would be better weighted we could try that?

 

8 hours ago, Sir Myshkin said:

2.) I question if Self Rez should be removed from the table as an option here. Only a few sets actually have access to them, and in terms of SO's they were explicitly added for the fact that the set needed to self-revive because it was almost a given it would die. Is the Rez a handicap in this situation where other sets are forced to evaluate a scenario more carefully when the Rez user can just barrel in constantly without concern? It changes the play style by a significant degree. Should the Rez be disqualified from testing, and if you die, that run be considered a DNF. Should we not have a category that exemplifies when a set fails?

This was something I did not really consider / communicate well on my end. I had been going in ignoring SoW / Revival as powers in that if I die the run ends / DNF. I think the test should rely more on the primary's mitigation if possible and if you die you need to try again.

 

8 hours ago, Sir Myshkin said:

3.) The OP of this thread was specifically targeting Scrappers, and as such, any other AT can't really be quantified fairly here because each has their own faults and require their own discussion (IMO).

Other AT's will have other hurdles. For example, most Tanks and Trollers could probably handle x8 no sweat with just the time being a factor.

 

8 hours ago, Sir Myshkin said:

An Actual Test:

Cause why not, I took out a Claws/WP over on Pineapple and executed the given test. This is the build I put together because I did want to test out using Hasten, and see if I felt force-herding with Provoke might dramatically change things versus the fact that WP naturally taunts in RttC anyway. So I did more than just the base-line tests.

  Reveal hidden contents

Hero Plan by Hero Hero Designer 2.23
https://github.com/ImaginaryDevelopment/imaginary-hero-designer

Click this DataLink to open the build!

Level 50 Magic Scrapper
Primary Power Set: Claws
Secondary Power Set: Willpower
Power Pool: Fighting
Power Pool: Speed
Power Pool: Leaping
Power Pool: Presence
Ancillary Pool: Body Mastery

Hero Profile:
Level 1: Strike -- Acc(A), Acc(31), Dmg(31), Dmg(33), Dmg(33), EndRdx(33)
Level 1: High Pain Tolerance -- Heal(A), Heal(29), Heal(31), ResDam(46), ResDam(48), ResDam(48)
Level 2: Mind Over Body -- EndRdx(A), ResDam(25), ResDam(25), ResDam(29)
Level 4: Fast Healing -- Heal(A), Heal(5), Heal(5)
Level 6: Slash -- Acc(A), Dmg(7), Dmg(7), Dmg(19), EndRdx(19), RechRdx(23)
Level 8: Spin -- Acc(A), Dmg(9), Dmg(9), Dmg(11), RechRdx(11), RechRdx(17)
Level 10: Indomitable Will -- EndRdx(A), DefBuff(42)
Level 12: Follow Up -- Acc(A), Acc(13), EndRdx(13), RechRdx(15), RechRdx(15), RechRdx(17)
Level 14: Kick -- Empty(A)
Level 16: Rise to the Challenge -- EndRdx(A), Heal(39), Heal(39), Heal(40), Taunt(40), ToHitDeb(45)
Level 18: Focus -- Acc(A), Dmg(36), Dmg(37), Dmg(37), EndRdx(37), RechRdx(39)
Level 20: Quick Recovery -- EndMod(A), EndMod(21), EndMod(21)
Level 22: Tough -- EndRdx(A), ResDam(43), ResDam(45), ResDam(50)
Level 24: Weave -- EndRdx(A), DefBuff(43), DefBuff(43), DefBuff(48)
Level 26: Hasten -- RechRdx(A), RechRdx(27), RechRdx(27)
Level 28: Heightened Senses -- EndRdx(A), DefBuff(40), DefBuff(42), DefBuff(42)
Level 30: Super Speed -- EndRdx(A), EndRdx(46)
Level 32: Shockwave -- Acc(A), Dmg(34), Dmg(34), Dmg(34), EndRdx(36), RechRdx(36)
Level 35: Resurgence -- RechRdx(A)
Level 38: Strength of Will -- ResDam(A)
Level 41: Conserve Power -- RechRdx(A)
Level 44: Physical Perfection -- EndMod(A), Heal(46)
Level 47: Combat Jumping -- DefBuff(A)
Level 49: Provoke -- Acc(A), Acc(50), RechRdx(50)
Level 1: Brawl -- Empty(A)
Level 1: Critical Hit
Level 1: Prestige Power Dash -- Empty(A)
Level 1: Prestige Power Slide -- Empty(A)
Level 1: Prestige Power Quick -- Empty(A)
Level 1: Prestige Power Rush -- Empty(A)
Level 1: Prestige Power Surge -- Empty(A)
Level 1: Sprint -- Empty(A)
Level 2: Rest -- Empty(A)
Level 4: Ninja Run
Level 2: Swift -- Run(A)
Level 2: Health -- Heal(A), Heal(3), Heal(23)
Level 2: Hurdle -- Jump(A)
Level 2: Stamina -- EndMod(A), EndMod(3), EndMod(45)
------------

 

 

| Copy & Paste this data into Mids' Hero Designer to view the build |
		|-------------------------------------------------------------------|
		|MxDz;1433;559;1118;HEX;|
		|78DA6D946D6FD25014C7FBC42A0CD80ADB188C0D05E7E6061D44DF9BE8D43723594|
		|2E27C260DBB62B5691BC0CCBDF4CBF8D26FE6C337C043CFFFE2DADAD0FCB8FFF3D0|
		|73CFE96DFFEB69FEC7F36F8F14B5F8C473A6D3E1603471C2504C327D67EC8E94C56|
		|5D2BD21F5615F7842D8E47C35AD2DC553F141F853615FB89E17065762A214CE83C0|
		|B39FB9E38F33D71FE7A2D52014E2321FFD3D134E48BA152DFA8EEF865F3C67E606B|
		|EF5347447F6E3E0F27AD877A63331B9DEA6C7EFD3ED9A0AAE792682A12B7B923AF3|
		|769226F3C5822A6273ACD5C13DB053629E247841B12AC7EA2AF26D97E2AC832FC9A|
		|EB3AFA623AF91E027F2C9601F19D4BFA2C75992C4F3B62CE66793C7B18835117B4B|
		|8F3327899875F0150564B936358BBCF70DA647EBD56893F3CC2AF2E625E15BB4E29|
		|47935B217B826A530A74B794DCA1AB2AD21FA5E2ECE03C90AB35D65FAE46F618716|
		|2A686147779394B951C91B5A97D1FDF20A6B1B09EAE4B389696E22FE08D36B833DD|
		|0207B05BE15E96BC4790236C95EE5BAB52AEAA925F8967C76906F07F90E8CF82424|
		|5B64DFE57CEA2E7C3B60486CA0470DF4E88EFE7FB610D3420DEF68DDC4B49AD0DE9|
		|B7CC216DA3EF63EA0F8436887F0FB4E633EC61B748C7E7630C72DCA61B34DB1B187|
		|1A695DF4A48B3A7B92C8593796A79A7E8BB7A7612CCFEAFC2865EDA6945E4A79905|
		|21EA694B394726EE0B1A4A891925DA7D3CCCAFC676EF91550D432D7FEFB9FA6A9E8|
		|8326FB01FEBA195764EDCFCD38F45103DBE05FBF1DA2B6|
		|-------------------------------------------------------------------|

 

This raises another excellent point looking at your slotting. We should probably agree upon the /WP build as a static variable when it comes to RTTC slots, and so on. I did not include Taunt or -Tohit for example on my runs, nor +Res in the T1 passive.

 

Of the Ancillary pools, Body mastery is again the least invasive though it does eat into the "these sets are balanced by end consumption" hypothesis.

 

8 hours ago, Sir Myshkin said:

Excellent summary of the tests.

This is awesome feedback! 

 

8 hours ago, Sir Myshkin said:

 

Will also note a small thing about endurance, as the Accolades got brought up. For this build in particular I did toss in Physical Perfection to wastefully fill power selections, and even tossed a slot into it, because it has no appreciable effect on the build. Without the accolades, the build recovers so much EPS that I don't even have to slot end redux into Spin. At 9.15 for a cost, and an EPS of 3.45 on its own (no PP/Accolades), and an animation time of 2.64/s (arcana), I recover 9.108 end in that window. Endurance is a non-issue in this scenario. As for the HP side, they add ~7 HP/s net regen at 1 target in RttC, 12 HP/s at 10 targets in range. That's a little more significant, but not sure that really warrants much mental effort for concern since this isn't 100% about survival.

 

I also tried out using Provoke as a collection tool over Shockwave to try and pool groups together. It really didn't change anything compared to just jumping into the vicinity of the spawn and letting RttC aggro and pull them back a hair into the previous spawn.

 

Also additionally tested Hasten on two runs, got a reduced time of 5:05 to 5:15, which was the difference of being able to keep Follow Up double stacked versus a small gap. Without Hasten FU carried a 6.16 cooldown, w/ 4.53. Spin lost 1.6/s of rech time, and SW 4.1/s.

These are all interesting points.

 

The added recovery from accolades & Body mastery can make certain sets seem much better if their supposed endurance issues are alleviated. Granted, this is very easy to do given you could do it in the SO environment (same with Accolades), but still 🤔

 

The taunting aspect is something that @Hopelingand I are trying to hash out. It seems that safer sets can afford to taunt to save time while some others cannot (like BS vs BA), and this may be a variable we need to account for going forward. 

 

Hasten will make a huge impact... but I am very curious as to when it was used. In the beginning? Towards the end when the EB came into play? Just curious on if there are swings either way since I do not think hasten could last the entirety of the encounter no matter what we do (until we get to crazy decked out builds)

 

[TheAdjustor]

Quote

1.) AoE is a given, and dramatically favored in this kind of style test "clear all." That's one of the biggest things about the Pylon Challenge that's been brought up, how ST take-down doesn't fully explore all avenues of what a build can do. In the realm of a "Clear All" style mission, AoE will always be superior.

2.) I question if Self Rez should be removed from the table as an option here. Only a few sets actually have access to them, and in terms of SO's they were explicitly added for the fact that the set needed to self-revive because it was almost a given it would die. Is the Rez a handicap in this situation where other sets are forced to evaluate a scenario more carefully when the Rez user can just barrel in constantly without concern? It changes the play style by a significant degree. Should the Rez be disqualified from testing, and if you die, that run be considered a DNF. Should we not have a category that exemplifies when a set fails?

3.) The OP of this thread was specifically targeting Scrappers, and as such, any other AT can't really be quantified fairly here because each has their own faults and require their own discussion (IMO).

4.) Yes, it will/does, and I'll have an example of such...

@Sir Myshkin

 

1) Fully agreed , with the caveat that this sort of mission is much more representative of solo/ team  game play  than say the pylon challenge.  This is doubly so when you mention your troubles with runners something that war mace really didn't have because it was stunning everything.

 

2) I agree self rez shouldn't be removed but it does have to be acknowledged. Having the self rez allowed me to press the envelope and herd larger than I normally would with this sort of build.  Allowing the self rez to be used and recording the times is a position that needs to be examined, the same way excluding them is , and so would be formulating a penalty for it.

 

3) I actually see this as a very good test for any AT and any combination. It's fundamentally the basic scenario for the game. Go into a mission and defeat some number of enemies.

 

4) Thank you for the hasten data. As you said no great surprise.

[Sir Myshkin]

6 minutes ago, Galaxy Brain said:

This raises another excellent point looking at your slotting. We should probably agree upon the /WP build as a static variable when it comes to RTTC slots, and so on. I did not include Taunt or -Tohit for example on my runs, nor +Res in the T1 passive.

 

Of the Ancillary pools, Body mastery is again the least invasive though it does eat into the "these sets are balanced by end consumption" hypothesis.

Even with the Taunt aspect, I still ran into issues of mobs running away from me and not staying engaged. While it was useful to initially collect, it didn't inherently change the fact that things didn't stay glued to me, so even sets that don't have a taunt (all the other ones), this was still somewhat fair play. In the realm of sets with Damage Auras, that would act just as effectively a threat-modifier for short distance collection. But I did try and maximize what was available to the ability, and adding a little -ToHit gave that much more wiggle room on the elemental aspect of defense. Although what was placed for attacks was still only Smashing anyway wasn't it? So really only reliant on one resistance shield. Maybe there out to be a little variety in the attacks (range, melee, aoe, s/l, elemental) to diversify the need to keep all toggles on beyond "good faith."

 

And to be fair, any character can select the Body Ancillary, so do you consider that a fair-use policy to balance consumption if needed? It still has to play within the realm of the SO system and should be considered in the same way that Hasten must be (somehow).

 

12 minutes ago, Galaxy Brain said:

Hasten will make a huge impact... but I am very curious as to when it was used. In the beginning? Towards the end when the EB came into play? Just curious on if there are swings either way since I do not think hasten could last the entirety of the encounter no matter what we do (until we get to crazy decked out builds)

I used it immediately. Claws is a generally fast attack set and my chain was "fine" out of the box with just FU > Strike > Slash > Focus, Hasten was only a caveat to fuel Spin and Shockwave speed. I ran in with it tirggered after the first use of Spin which gave me two full minutes of carnage. When it came to Bosses+, there wasn't really anything different for me to exercise in a ST realm other than dragging them into the next spawn and applying more AoE. Getting Spin faster made it so I didn't have to spend an extra 2/s occasionally tabbing-attacking individual kills. That's what added up over time.

[TheAdjustor]

13 minutes ago, Sir Myshkin said:

I used it immediately. Claws is a generally fast attack set and my chain was "fine" out of the box with just FU > Strike > Slash > Focus, Hasten was only a caveat to fuel Spin and Shockwave speed. I ran in with it tirggered after the first use of Spin which gave me two full minutes of carnage. When it came to Bosses+, there wasn't really anything different for me to exercise in a ST realm other than dragging them into the next spawn and applying more AoE. Getting Spin faster made it so I didn't have to spend an extra 2/s occasionally tabbing-attacking individual kills. That's what added up over time.

 

Claws is an interesting case here as well. IIRC the cone sets aren't actually cones so much as a circular sector. Also there is the rather large difference in performace if KB=>KD IOs are allowed.

[Galaxy Brain]

Given this is solo, I don't think taunt will be that big a factor given there isnt much to steal aggro off of you anyways. That said I do think we should settle on a build for /WP that we share per tester to try and iron out variables.

 

Body Mastery and Hasten can be taken by anyone as well, but I personally think they are a bit game changing for certain sets. We should definitely include them in at some point tho.

[TheAdjustor]

40 minutes ago, Galaxy Brain said:

Given this is solo, I don't think taunt will be that big a factor given there isnt much to steal aggro off of you anyways. That said I do think we should settle on a build for /WP that we share per tester to try and iron out variables.

 

Body Mastery and Hasten can be taken by anyone as well, but I personally think they are a bit game changing for certain sets. We should definitely include them in at some point tho.

Hasten is kind of up there with the self res. Even the blaster can have access to a self res via epics.

It strikes me that a run over 2 minutes should be done as a base sample and the same setup with hasten. Which would give an insight into just how much recharge is a factor for the the set.

 

Claws is a good example here. Hasten will let you stack followup higher but it's a very short window, while with sets that have long recharge attacks and some form of build up ability, it will let you get those combos off much more frequently.

 

The self rez is more problematic as eliminating runs where you die will skew the data, and may well understate how much mitigation the sets provide.

[Haijinx]

So if i follow correctly, the punching bag damage was upped to discourage too much herding, but now people dying during runs has become a real concern?