GU50, Tunso Testing, and Player Testing

We learned two key things this last go-around.
1) Devs don't test midrange, they test point-blank melee and full range
2) Players don't test point-blank melee and full range, they test midrange

So to really get on page, we need to all test all three. I've done some of that last night and tonight using the Tunso spec and constraints (and then deviating separate tests for mid-range and WM) and had some interesting insights.



I found that for Nature the results of a Tunso Test will be 1.3x-1.4x-1.5x less than what we can expect via a fully modded, fully spec'd, fully proficiencied, fully generator mod, and fully maximized toon as we tried using in GU50 PC Test Testing Feedback (and some reports on live). With high SP or simply with other powersets, maybe you can make a case for close to 1.5x. We all know there's some variability there. But I'm starting to learn that's not necessarily the most important thing to focus on at the moment.

I think we're going to find that pointblank melee and pure range is actually pretty balanced across the powersets. I'm sure there's some exceptions worth noting. But anyway, those two serve as the high water and low water marks for what our powersts can achieve. When we pull back to midrange, we start sacrificing some of that damage potential we can achieve at pointblank melee. The powersets we think of as OP are the ones that can pull back to midrange and sacrifice the least amount of damage. In the case of Fire/Ice they can fundamentally preserve all their pointblank damage at midrange. But when you compare pointblank to pointblank they are close to equal.

It's an interesting way to conceptualize where we stand. It would have been more apparent if (A) the devs weren't so headstrong to acknowledge and test midrange, and (B) players weren't so headstrong to acknowledge and test pointblank melee and full range. Even if pointblank melee and full range isn't ideal in a number of combat situations, it certainly shouldn't be overlooked in our analysis anymore. It really opens our eyes about where things are balanced and what aspects of them are making a powerset OP (at midrange).

If anything, this is maybe where we need to restart from. We'll probably learn other interesting insights. Many of them will probably be about how different powersets scale to live results from a Tunso Test. But if we can test our powersets more comprehensively in the future and share our results, we probably can have a more productive discussion about where we stand and what adjustments in which places are apt.

Here's a sample of one of my Tunso Tests in detail (my screenshot above was a summary)....

Human Melee (Swarm)
Might:8255
Precision: 2796

Human: Savage Growth, Vine Lash, Briar, Swarm, Harvest, Carnage
Weapon: Dual Wield

Rotation
> Setup: Savage Growth, Vine Lash, Briar, Swarm
> Main: 6xHold/Harvest, 6xHold/Swarm, 6xHold/Harvest/Carnage

Sequence
> Setup->Main->Main->Main->Main->Main->Main->Main->Main

Distance: 0m (entire time)
Damage Range: 12,185-24,076
Damage Mean: 19,474
Damage Median: 19,107

Parser Summaries from Combat Log
[Summary] Damage [10.0s] 12185/s - 121845 total - 281 hits (5471 max) - 62 (22.1%) crits - 8 targets
[Summary] Damage [10.0s] 22342/s - 223418 total - 416 hits (13253 max) - 112 (26.9%) crits - 8 targets
[Summary] Damage [10.0s] 18534/s - 185340 total - 382 hits (18091 max) - 115 (30.1%) crits - 8 targets
[Summary] Damage [10.0s] 21949/s - 219492 total - 407 hits (13730 max) - 121 (29.7%) crits - 8 targets
[Summary] Damage [10.0s] 17649/s - 176486 total - 423 hits (9651 max) - 117 (27.7%) crits - 8 targets
[Summary] Damage [10.0s] 20652/s - 206519 total - 389 hits (13707 max) - 88 (22.6%) crits - 8 targets
[Summary] Damage [10.0s] 19616/s - 196162 total - 417 hits (13109 max) - 124 (29.7%) crits - 8 targets
[Summary] Damage [10.0s] 21765/s - 217654 total - 415 hits (10776 max) - 111 (26.7%) crits - 8 targets
[Summary] Damage [10.0s] 19255/s - 192550 total - 409 hits (10143 max) - 114 (27.9%) crits - 8 targets
[Summary] Damage [10.0s] 21455/s - 214550 total - 416 hits (11245 max) - 109 (26.2%) crits - 8 targets
[Summary] Damage [10.0s] 18091/s - 180911 total - 425 hits (11451 max) - 117 (27.5%) crits - 8 targets
[Summary] Damage [10.0s] 19107/s - 191071 total - 383 hits (12027 max) - 86 (22.5%) crits - 8 targets
[Summary] Damage [10.0s] 17238/s - 172379 total - 389 hits (11533 max) - 106 (27.2%) crits - 8 targets
[Summary] Damage [10.0s] 20225/s - 202250 total - 399 hits (10377 max) - 112 (28.1%) crits - 8 targets
[Summary] Damage [10.0s] 18410/s - 184096 total - 390 hits (13896 max) - 99 (25.4%) crits - 8 targets
[Summary] Damage [10.0s] 18513/s - 185125 total - 415 hits (10990 max) - 116 (28.0%) crits - 8 targets
[Summary] Damage [10.0s] 24076/s - 240755 total - 413 hits (14070 max) - 114 (27.6%) crits - 8 targets
[Summary] Damage [10.0s] 17634/s - 176337 total - 391 hits (10865 max) - 110 (28.1%) crits - 8 targets

You can repeat that kind of test on PC Test (or Live) with a full strength character and compare your average results (Mean and Median). That's where we can have multiple test points to see how well the Tunso Test results scale to full strength. Again, I suspect they might scale slightly differently for each powerset but we'll see. Let's prove it. For the damage range, I prefer to keep my first two cold 10s samples. For calculating means and medians, I throw them out.

Here was Tunso's thread about conducting a what I love to call a "Tunso Test". I'm not sure if he'll love the name or not.

https://forums.station.sony.com/dcuo/index.php?threads/testing-tips-dos-and-donts.256329/

Get all your T1 crit innates. Don't get any T2 innates. Don't get any WM innates. Ideally, don't unlock WM unless you already tested without it and then later want the comparison of testing with it. No mods go in your red/yellow/blue slots. You can have Max Damage white mod, but no other white mods.

He mentioned what your might/precision will be. He quoted general values. Spec'd per instructions, your stats will more precisely be Might:8255 and Precision: 2796. Again, no league hall proficiencies nor generator mods. He also mentioned spending around 100SP. It probably will be less. Otherwise any more will just be unlocking weapon combos for testing different weapons. Don't test midrange until after you've tested your very best attempt at pointblank melee and full range.

Really try to use a non-WM combo first and test WM later. You may be surprised. I think of Nature as a 12s rotation powerset. So I know certain loadouts fundamentally mean powers coming off cooldowns at certain times. I pick my weapon combos based on what does great damage per Shiny's thread, but I also look for stuff that has just the right amount of animation time. I find the ones Tunso listed tend to time out perfectly to keep up a rotation where various powers with 12s cooldowns are used for clipping at interleaved points int he rotation. It's almost as if he designed them to be that way.

It's really only ideal to do a Tunso Test on the PC Test server. You're best off making a new toon for it. You probably need to create a league hall also. Keep in mind that league halls cost 1 prestige on PC Test and also that having base items placed in a lair will provide max prestige on a toon. So feel free to put a couple decorations in your lair, walk out, create your league of one, and then walk in.

I should note, the Tunso Testing Tips thread disavows midrange as a thing. I break down what they're thinking in terms of Pure Melee and Pure Range. I then describe how you can use ranged weapon combos plus any melee powers (the 12.5m ones, not the 7m ones) to create the phenomenon known as midrange. That shouldn't be news to you, players have invented the concept of midrange and been founding their rotations on the idea for a while. I illustrate how that emerged with pictures in the tweet below.

https://twitter.com/SoreDCUO/status/621743624368300034
(Don't forget to scroll down and look at all 3 tweets...4 pictures)

Sore, I dunno how you do it...





RESPECT

Awesome job Sore....maybe they should hire you to do the testing!

Good thread, Sore.

Good thread. Don't the new definition of testing make parser testing no good. Don't the devs have to make to powers work either 25 or 7m? I just think we are testing a myth, and unless the devs make the powers match there definition, testing is useless

The Tunso Tests use the parser too. I suspect changing all 12.5m abilities to 7m would be met with player frustration and anger. It's certainly one option for next steps. There may be other less controversial options. Maybe. The purpose of the tests now are to (1) ensure melee and range numbers are consistent and then (2) crisply articulate how much of the melee-level damage is being preserved at midrange. It makes it more clear whether the issue is damage figures or mechanics. It may lead to focusing on specific abilities.

Sore can you produce the same results in test with Ice, HL, Fire, Mental, etc...We need charts to compare powers equally across the board...

Proximity requirements are idiotic anyway. I can understand having them IF and ONLY IF there were animations that matched that area of effect. As it stands, I have played the game for 5 years and still couldn't tell you what the hell 10 meters or 3 meters looks like on a fictional DCUO floor.

I realize that doesn't contribute to the conversation, other than to affirm that yes, most all of us test at "mid range" because most of believe that the idea that our damage is constantly shifting based on where we are standing is really dumb.

I think we need to open this up to everyone. We will see who steps up. I only feel qualified doing Nature tests. The two Ice tests were thrown together by Radium.

That's why we need a range indicator on the targeting reticule. Could be color. Currently, we have gray for out of range and red for in range. Perhaps green for 12.5 m or closer?

I was doing some tunso testing on my fire alt lay night. I didn't record any data because I got annoyed but ill tell you it's bad.

My fire dps is 119cr and unmodded. White mods boost was expired 90 SP specced but using nothing but burning determination and fire burst I was doing 20ks 10s on 8 dummies. When I did fiery weapon and inferno I was getting 22k-24K DPS where's my sorc main is doing 14k fully modded 124.

ON PS I cant really meet the qualifications of an acceptable test

You are not doing Tunso test as you need to be CR 126, otherwise your test results are null and void.

We might of lost a good portion of our testers due to how DBG handled the last set of testing. Lots of tester have been turned off since Quantum as they did not listen than. Players were stating Quantum was too strong, mostly due to the Max Damage glitch, but it took until GU47 to see it in action.

I dont like the 10 second intervals for Sorcery and Earth.
The pets aren't properly gauged. (Im speaking long term)
You have to use Despawns

Fury and Crystal alone can put out 12k if they go berserk.
usually they gauge around 7-8k.
That may seem well but your powers are weak in comparison along with your weapon choice needing to be HB Brawling or Two Handed

I really like this idea and I believe it has been mentioned before. And since we functionally do have a mid-range, my preference would be:

Gray = Out of range
Green = greater than 12.5 m
Yellow = 7 m - 12.5 m
Red = less than 7 m

Good thread, well written and insightful and I really dont want to derail it but I just cant see one logical reason why anyone should cater to the kind of testing described in Tunsos post. Melee (point blank) is NOT VIABLE, full range is not effective in most cases.
This "Tunso Tests" are based on artificial conditions that bear no reality check.
Equaling everything < 25m is just wrong in every possible way but explains why after all the "balancing" there is still no real balance.
Im not going into other details of this kind of testing which I also consider "problematic" to say the least as you already provided really good "workarounds"/insights for most of them.
Imo its on the devs to adapt their testing not the other way around as there is no reason to violate comparability any further when the cat is finally out of the bag. Also the shifting of the burden of proof from Devs to community isnt acceptable at all.