Scrapper Melee Primary Testing: "Standard" environment

[Lord_Cyclone91]

6 minutes ago, Galaxy Brain said:

@Lord_Cyclone91, what is your strat going into each mob?

 

 

On Elec I didn't mention, sorry. 

 

Group up two spawns, then BU (if available) -> LR (when up) -> Jacobs Ladder (as they get up) -> Thunderstike -> Chain Induction -> Havoc -> Jacobs Ladder-> clean up with Havoc, Charged and Chain...repeat. 

 

When LR is not up, then I lead with Thunderstrike, Jacobs Ladder, Chain, Havoc, Jacobs Ladder... 

 

I tended to be a little more careful how I pulled them, because I was having some trouble with double bosses at times with the trash, especially when the conning was at +1.  Sometimes the KD doesn't hit as many as you need when you need it.  I would often drag one along with 1/3 or so health to the nearby spawn if it looked reasonable. 

Edited September 27, 2019 by Lord_Cyclone91
Added spawn movement stuff

[Galaxy Brain]

I've started to do that too with the bosses if it looks like I am just standing there wailing on them, if the set has AoE to chew through a spawn on minions/LTs at least. 

 

@Hopeling, was your strat similar?

[Haijinx]

Ive found Ele to be underwhelming except for the Awesome Telenuke. 

 

On SOs how many nukes would you even get in 6 minutes?  And would the spawns be big enough for it to really matter?  

[Haijinx]

Since a lot of IO increase survivability im curious how this will affect something like FM.

 

 

[Lord_Cyclone91]

4 minutes ago, Haijinx said:

Ive found Ele to be underwhelming except for the Awesome Telenuke. 

 

On SOs how many nukes would you even get in 6 minutes?  And would the spawns be big enough for it to really matter?  

Yeah, I would say I used it 4-5 times per run, as an opener on a new 2-spawn grouping.  Of course, it was very helpful to have 6-8 minions gone and 2-4 Lts taken down to 1/3 or 1/4 but overall, the uptime is...not great.  It needs more recharge than this basic setup will allow for.  I found myself wanting Cross Punch again.  

[Lord_Cyclone91]

15 minutes ago, Haijinx said:

Since a lot of IO increase survivability im curious how this will affect something like FM.

 

 

In my runs, Fire felt like I was dancing on the edge of knife with the basic setup.  It wasn't uncommon for me to bite off more than I could chew and didn't have a lot of tools to slow the damage down. I did need to do some dancing and hokey pokey around an obstacle here and there, not going to lie. I had to analyze a fair amount on the fly, because the spawn sizes and conns did matter, with how I could survive them.  I think I had 3 deaths in the 13 runs resulting in DNFs, with the 10 I reported being ones where I had no deaths.  That said, I had numerous times where I was in the danger zone (below 1/4) and some pulsing sliver moments as well.  Like I said, better players may have a different experience but that was mine.  Fire was easy to get precarious with, particualrly when I'm used to having some end games builds where I don't need to think overly much about things other than extreme debuffers and such. 

 

This is in complete contrast to Ice, for example, where my health bar almost wasn't moving because everything was falling down, slowed, or held.  The price for it, is the time needless to say, but what a difference in styles. 

Edited September 27, 2019 by Lord_Cyclone91
Added note on Ice

[Shoulung]

I read through this, and I didn’t see inspirations mentioned anywhere. I think they could really tank the readings because player usage varies widely. 

[Galaxy Brain]

7 minutes ago, Shoulung said:

I read through this, and I didn’t see inspirations mentioned anywhere. I think they could really tank the readings because player usage varies widely. 

Inspirations, temp powers, most pool powers and epic powers are banned at the moment to limit variables

[Sir Myshkin]

2 hours ago, Shoulung said:

I read through this, and I didn’t see inspirations mentioned anywhere. I think they could really tank the readings because player usage varies widely. 

The very reason why inspiration specifically can't be used in a test like this is because they completely invalidate this specific style of test. If anyone remembers CEBR self-farming technique, that was applicable to any given character. Basic premise was the stock up a full tray of Reds, burn a bunch, stock up again, walk into the AE mission and start rolling around at the damage cap. The map it used had ambushes and spawn triggers at the opening so that the player didn't have to go anywhere. Take Claws, hit level 6, cap the mission at level 6, and just collect the XP wave after wave. Took about 5-10 minutes to complete, restock, and then repeat over and over. It was possible to self-level (specifically a Claws/Elec Brute) in 6 hours. That particular mission was worth about three of the one being tested right now, and was run at x8 (no boss).

 

Damage-Capping and walking into this mission... I'd doubt it'd take more than 2-3 minutes (depending on how quickly you herd).

[Galaxy Brain]

It appears the mission is down on Pineapple...

[Hopeling]

9 hours ago, Galaxy Brain said:

@Hopeling, was your strat similar?

Yep, basically the same. If Lightning Rod came up and I was at high health but only had a few mobs left, I would run to the next spawn and try to hit everything at once. I didn't take Lightning Clap though.

9 hours ago, Haijinx said:

On SOs how many nukes would you even get in 6 minutes?

If you slot it 3dam/3rech, it has a 45-second recharge to match Build Up, so you can get 8. In practice probably 6-7, since you're probably not using it exactly on cooldown.

35 minutes ago, Galaxy Brain said:

It appears the mission is down on Pineapple...

I didn't change anything, but when I logged in just now, under My Creations it had an error message that I hadn't set the contact. I put him back and republished, so it should be available now. Not sure why that happened, maybe the patch broke something somehow...?

[Lord_Cyclone91]

14 hours ago, Hopeling said:

Yep, basically the same. If Lightning Rod came up and I was at high health but only had a few mobs left, I would run to the next spawn and try to hit everything at once. I didn't take Lightning Clap though.

If you slot it 3dam/3rech, it has a 45-second recharge to match Build Up, so you can get 8. In practice probably 6-7, since you're probably not using it exactly on cooldown.

I didn't change anything, but when I logged in just now, under My Creations it had an error message that I hadn't set the contact. I put him back and republished, so it should be available now. Not sure why that happened, maybe the patch broke something somehow...?

Yeah, I think that slotting would have helped me clear the spawns a little better, but I think I was slowed down at least as much by the Boss spawns, numbers wise and locations.  My best runs were clearly where I was light a boss or two and then also helped if they weren't grouped up by another one.

 

That and the 'me' factor. As always, there is some player difference here as well, as I know I'm not the best player out there, and pretty sure I don't know Elec well enough to squeeze every ounce of performance out of it. 

 

I could re-run some tests with that slotting, but perhaps we can just note that my LR slotting was a little sub-optimal and likely slowed me down a little bit.  Have others run their tests with the improved slotting

[Galaxy Brain]

I ran Dark Melee (7:26 avg) and dual blades (6:06 avg)

 

Dark Melee really, really suffers from a lack of AoE. Even played optimally, I could tab target repeatedly to spread -ToHit to a crowd and gather max Soul Drain -> Siphon Energy -> Shadow Maul a few times and still have to spend a lot of time bopping foes over and over. The ST damage did not feel fast either all things considered, though vs 1 guy the safety is hilarious.  I feel Dark Melee is very akward at a base level and really needs a secondary with AoE to take advantage of Soul Drain, or a *ton* of rech to stack up soul drain to hand out ST naps to each foe you target. Even still that seems slow... As I mentioned earlier, AoE matters for survival even at the easiest difficulty.  Locking down 1 enemy doesnt stop the other two from hurting you unlike Ace which can bounce multiple foes or Sword which can buff your defense vs all attackers.

 

Speaking of swords, Dual Blades feels like claws lite. You have 3 AoEs, decent ST, guaranteed knockdowns... but the cost is that it does everything slower than claws. 1000 cuts is awesome as a 10 target cone, but the animation lasts forever and does not knock down till the end unlike Shockwave which is instant. Same with the strong ST attack which isnt ranged, takes a while to animate and does not knock till the end. Missing with DB hurts a lot ad you ruin combo effects and miss out on CC, but also being stuck in an animation for quite a while.

 

 

[Frosticus]

On 9/26/2019 at 11:54 AM, Galaxy Brain said:

Thank you @Lord_Cyclone91! I've updated the numbers in the sheet as well (I need to update the OP come to think of it...)

 

Current standings:

 

Playtesters: 5

Total missions tracked: 94

Total Average completion time:  6:05:07

Current set rankings:

 

 

What's interesting is that if you look at TW every power deals approx 10% or more base damage than it *should* given their recharge

ie.  8 sec crushing blow = 112.9

vs  8 sec hack = 102.6

 

16 sec rend = 200.9

vs 16 sec clobber = 182.7

 

I imagine if you took away that bonus damage (not sure why it is on every power) then TW would likely be a median set on SO's  (and still a superstar on IO's).

[Galaxy Brain]

To be fair, TW has not had the full run of tests yet nor has every set been tried. 

[Hopeling]

2 hours ago, Frosticus said:

I imagine if you took away that bonus damage (not sure why it is on every power) then TW would likely be a median set on SO's 

It's worth pointing out that we only have eight data points for TW, all of them mine. Moreover, the TW average so far is split between two different approaches: five runs without Defensive Sweep (averaging 5:41, similar to Claws), and three with (averaging 5:04, far and away better than any other set). Parry-type buffs are really really powerful in this test, because they allow you to herd a ton of enemies safely, and TW has enough AoE to really take advantage of herding. That's also why Broadsword is doing so well; without Parry, I got times around 7:00, with one unusually good run at 6:27.

 

In real play, Parry buffs are not as useful as they look here - any set can get a similar effect by using inspirations, and not all incoming damage is flagged Melee. I'm tempted to say we should keep two entries for each of TW/Broadsword/Katana, one with their parry power and one without. (I don't think the same concern applies to eg Dark Melee with Siphon Life, since that's a strong attack that is worth using even when you don't need the heal.)

 

Without Defensive Sweep (which lots of builds don't take), it looks like TW is "only" a top-tier set on SOs, rather than in a tier of its own.

2 hours ago, Frosticus said:

(and still a superstar on IO's).

So far, TW does seem to be the only primary that is limited by endurance in this test, and very frequently has gaps when all its attacks are recharging. So it does look like it stands to benefit more from IOs than other sets - most sets take advantage of high recharge by replacing weak attacks with strong attacks, but TW will replace dead time with strong attacks, which seems like it should give a larger improvement in performance. We'll see though.

Edited September 29, 2019 by Hopeling

[Cawshun]

Just ran 10 runs with Staff with some pretty solid, but not too surprising results. Staff has a great kit for this sort of test with two cones and a pbaoe, a -res debuff from the pbaoe, a +def buff from a cone, and a +res buff from Sky Splitter to help with herding. Stayed in Form of Body for the damage bonus as well as the accompanying buff/debuff on the spenders. Serpent's Reach was useful for runners and deals decent single target for the bosses. I would usually dance around a little to optimize cones to hit as many mobs as possible, which I think is where a lot of variance in the time comes from. Pretty much all abilities were used often in order to keep stacks of form up as much as possible. The set is fairly consistent for the most part as long as I was getting good uses of the cones.

Staff Times
5:49
5:38
6:06 - experimented with maximizing resistance from Sky Splitter, but maximizing 3 stack Eye of the Storm was definitely better outside of bosses and EB.
5:32
5:51

5:34

5:38

5:59

5:44

5:20 - This run had ideal spawns and very few runners.

Average Time: 5:43

[nihilii]

So much great work done in here. Chosen parameters make a lot of sense as a baseline.

At the same time, seeing as this thread is already used in performance discussions elsewhere, I kinda want to say the testing parameters might make it even more removed from actual gameplay than Pylon testing, at least for some of us. Even on SOs.

Level 50 mission on +0/x3 with no insps and no temps with /wp specifically, vs level 1 character filling his tray with lucks, using insps as they drop, starting on +0/x6, buying amplifiers, recovery serum as needed, and using any secondary. Day and night.
 

Whereas if you need to deal ST damage during an extended period of time, your DPS baseline from a Pylon will mostly hold.

 

I know most of the participants here are well aware of that, and aren't claiming this is the be-all end-all of balance testing. Not trying to downplay the effort either. Just voicing some thoughts. 🙂

Edited September 29, 2019 by nihilii

[Galaxy Brain]

I would argue the parry type powers are great on the melee sets because they all give melee defense while in melee range. Where they get questionable is when other bonuses come in that make the parry superfluous...

 

The mitigation sets give matters for performance.  In this test, it allows you to safely herd up a bunch of enemies. In real play, this translates to being able to safely jump into encounters and handle yourself in order to deal your damage.

 

After this initial run, we should look into a way of fighting more advanced punching bags (maybe they have guns lol) to see how much the mitigation can matter? 

 

Edit: Giving enemies Flares  would work, as all the parry type moves defend vs Melee and S or L damage, not Ranged + Fire.

Edited September 29, 2019 by Galaxy Brain

[Hopeling]

6 hours ago, nihilii said:

this thread is already used in performance discussions elsewhere

What?? Where? I'll go yell at them, haha.

[Galaxy Brain]

I had seen this quoted in the tanker beta thread, had to make a point that this is  not finalized yet 😛

[Haijinx]

Once tested with IO builds the parry powers will matter a lot less.

 

[Galaxy Brain]

I mean yeah, but for the non IO world they have a ton of power, esp when paired with like a straight resist set 🤔

 

Though that is sort of a big side issue related to IOs in general. The ability to get a ton of defense negates these powers useful side effects to where they become variably useful attacks, though I guess they allow you to build for other stuff besides defense?

Edited September 30, 2019 by Galaxy Brain

[Galaxy Brain]

Took some time to go over resistances across 61 enemy factions. This is based on an older list that was shared on discord, but it should still be relevant for the game today as it covers a ton of common enemies. It does lack Praetoria and other newer enemy types though, so take it with a grain of salt.

 

The sheet is broken down as such:

• Totals - Shows the culmination of all the data
• Group Averages - Averaged resists per enemy faction
• All Minions - All minion rank enemies per faction, averaged
• Minions 1-20, 20-40, 40+ - All minions within the respective ranges
• All LT - All lieutenant rank enemies per faction, averaged
• LT 1-20, 20-40, 40+ - All lieutenants within the respective ranges
• All Boss - All boss rank enemies per faction, averaged
• Boss 1-20, 20-40, 40+ - All bosses within the respective ranges

I use Averages due to each enemy faction not being comprised of a single enemy type. Sure, that Council mission may have a Robot that laughs at lethal or psy damage, but for every bot there are probably 5 normal dudes who don't resist it at all. So in that example, an enemy group with 25% lethal resist on one type of unit, and 0% resist on an other would average out to 12.5% resist if you were to just run through a bunch at random.

 

To factor the spread of different units across factions, I sorted the following data: 

 

This covers the raw, average resistance value as well as the odds of running into an enemy with said resistances, either positive or negative (for the enemy), and then factors those together for an "Adjusted Average". 

 

 

Putting everything together, sorted by the Highest to Lowest "ADJ Resist" we have the above ranking. Toxic is the most resisted overall throughout the game and has no enemies weak to it (in my sample of 61 groups), while Negative is the least resisted with the highest amount of enemies that are specifically weak to it. Smashing damage is the most likely to be resisted, and Fire is least likely. Further breakdowns per enemy rank, and level range are in the sheet below.

 

https://docs.google.com/spreadsheets/d/1arWBGWuwGCWqSrkLgnWK26xNTrILUO_QNKn8034A3L0/edit#gid=1737774505 

 

 

How does this relate to our testing though? Well, as we are trying to gauge "overall, standard performance", we can compare sets to these values. 

 

For example, Fire Melee deals Fire, Smashing and Lethal damage types in various proportions. Going by the AVG resist, those would be resisted by 4.88%, 12.29% and 13.97% each... or by 0.56%, 5.24%, and 5.96% each used ADJ Resists.

 

So, if Fire Melee has an average time to clear of 6:44, and Broadsword 6:16, if we were to just apply the ADJ resist to each for Fire/Lethal we could adjust times to 6:46 for Fire and 6:39 for Broadsword. The gap in theory closes from 27 seconds down to 7 seconds due to Lethal being much more resisted.

 

Thoughts?

 

@Hopeling

Edited October 2, 2019 by Galaxy Brain

[Galaxy Brain]

Ran elec melee 5 times, did not  significantly change the average, Added a "Run Count" column so we can pool mission runs for data points.

 

Elec Melee is super binary on SO's. You just wreck minions and LTs easily, but then you are suddenly blocked by boss battles. The ST damage is very lacking, and the mitigation is not reliable. It almost feels like reverse Dark Melee, but like that set it also seems to get much better when you pair it up or boost via IO than by itself.