Balance-Experiement Devserver quickplay

[DrTarDIS]

One of the things the "total overhaul idea" newspost-talkstop brought to my mind, was that we as a community and shine as a company don't have a lot of cross-talk when it comes to "how we play"

in short: a "release (latest)" dev-server with wacky-balances updated/changed regularly to explore possible "current implementation" changes would be a great asset to brainstorming, balance, and satisfaction. It's strage we have the battlemode and whatnot server options, but never really test them

Ideally this would be a "quick-play" option in the launcher. Possibly nestled in the "online play"section, and/or as an option for a disposeable gameworld mimmicking these "crazy changes" as last updated with the news-items.

Spoiler: in long

Regular changes, more communication.

Every week, completely screw ith some of the values in the config while resetting other ones to default.
all systems blocks switched to 1 solid mass, or power softcap negated, or thrust penalty reversed, or shield rates changed, storage weight/capacity, Warheads like supernovae, NPC factions with "public join" all up for game. Drastic "what does this button do?" chaos and applause and rage contained to a productive level.

keep community input on this with 2 dedicated threads in a dedicated subforum, restarted and renewed ever week.

thread 1 "what do you guys think we should try next week?"
thread 2 "this is what we played havoc with this week. What did you think?"
​

Balance, imbalance fun:

The 1M->2m Balance change caused some hooplah didn't it? The thruster boxdim one too, eh? Would have been nice to test-sample that without a dedicated "dev build" install directory...

A lot of server admins don't know what config does what. Educate them! Show the community the crazy possibilities in the larger sandbox!
​

Satisfaction

The community likes to run it's big fat mouth. They also like Being shown neat things, and are terribly full of ADHD. Cater some to that, everyone will come out better for it.​

Edit: Shameless attwhr AndyP Bench Criss Saber​

 

[Lecic]

1 week is not long enough to gain proper data on balance changes. It would need to be 2 weeks at least.

 

[DrTarDIS]

1 week is not long enough to gain proper data on balance changes. It would need to be 2 weeks at least.

Arbitrary number is arbitrary, for all I care "another week of the same! could be a damn good indicator of core-game balance changes. I'm thinking of it more as a quick-rollover test of otherwise Very-Bad-Ideas™.
Other than that, what do you think of the core idea? Any other ways to improve it?

 

[Lecic]

Arbitrary number is arbitrary,

Not exactly. 1 week really is not long enough for people to try and adjust to the new changes, especially for more drastic things like attempts at economy changes. 2 to 3 weeks is a much better timeframe for adjusting to changes and gaining proper information on how the effect the game.

As for the rest, I have no problems with a testing server running an alternate config for the devs to gather data.

 

[AndyP]

Its not really a problem that we do not offer the game modes on servers.




The battlemode server is for example running since over a year, and unless we from Schine play on it, there is barely any activity there.
I would like to see people try out this mode, and thus populate it with PvP focused ships to examine.

All except the "Event" and "Stream" server follow the dev-builds.
The "normal" one is frequently used to test config changes and try balancing.
There is some small group of players that plays there a lot and provides useful information.

Trying extreme changes like suggested by some poll options would be more for fun than yield any scientific data we could use.

Especially economy changes need full resets of the database and least 3-4 weeks to show any result, we had similar problems with the NPC factions, that they did work for a short while and then the economy became too stable to allow any player influence, but that is for now "OK" to some extend.

We are open for changes and ideas in general, as long as they follow a concept and do not turn one bad balance to the favour of another, which is actually hard to achieve.

- Andy

 

[DrTarDIS]

Its not really a problem that we do not offer the game modes on servers.
...
All except the "Event" and "Stream" server follow the dev-builds.

Just making sure you mean "release builds" not "dev builds" as one is kinda obfuscated from average players. Even then, they're so lost in the serverlist It's no wonder they don't see much activity. But that's mildly tangential to my point.

The battlemode server is for example running since over a year, and unless we from Schine play on it, there is barely any activity there.
I would like to see people try out this mode, and thus populate it with PvP focused ships to examine.

Case in point why I think it should be put in the launcher in a more visible location. "local play" "online play" ->"quickplay(battlemode)" "quickplay(Balance weekly)"

We are open for changes and ideas in general, as long as they follow a concept and do not turn one bad balance to the favour of another, which is actually hard to achieve.

- Andy

See, "one bad balance" is a kinda vague term, and it's vagueness is the point I'm trying to drive at.

Let's take warheads as an example:
-are their damage & radius effective? Some say "no", some say "yes, if you build them right" But actually going to various extremes in their implementation isn't looked at by more than a few interested parties, who I doubt talk to Shine very often if at all. *One week of testing at x10 damage and radius on a server, newspost and/or thread promoting it would probably give useful balance information.*

Let's take jump drives as an example:
-Some servers buff jump range. Really, a LOT of servers buff it. It just makes sense from a "load 27 sectors for (servertimeout) seconds because a chain-drive stopped in for 4 seconds" resource standpoint. Yet the official balance remains less, I'm sure for some reasoning other than "server load needs to be mitigated". *One week of testing heavier/lighter costs/ranges/times again provides useful datapoints* On this particular standpoint, I invite you to take some time running a "perpetual chain-drive ship" from the dock on any of those official servers and look at the load that pops up when the ship has-been-and-continues-to jump past the (sectortimeout) variable of the server.

Let's take the recent "too many blocks to achieve X" boil-down that puts a total rework on the table. Can you honestly say you guys have played with the blockconfig.xml in a "large multyplayer environment" to see if the desired effect can be approximated with current implementation? EG Have you considered-and-tested unbinding thrust from energy completely and rebalancing other systems from that standpoint?

The above are some rather limited examples of "you can change it now if you bother" server-side balances that could be addressed without any low-level code work.

I know most of the above, as well as others were tested favorably and unfavorably on Dominion Prime. Weather those changes were noticed by you guys is sort of the issue that leads to a "total rework of unbalanced systems" taking up your time to begin with IMHO.

In the end, I'm only pointing out that the massive community feedback loop this caused could be put to useful work if you explored "Good AND bad balance decisions" on those very servers you pointed out. The key would be Using that news&forum system to direct players towards the tests. It would also require "taking risks" on balance that would otherwise cause MASSIVE backlash because "you just broke everything, everywhere, without consulting the players."

Also, I want a big-head Dave, dangit.

Edit: points I missed first pass

There is some small group of players that plays there a lot and provides useful information.

Trying extreme changes like suggested by some poll options would be more for fun than yield any scientific data we could use.

"fun" is kind of the point of a game, no? If some "extreme change" yields fun that would otherwise be hidden behind hours of grind, is that not a beneficial change to consider?
Tangental and slightly snarky related point: How many people want to "work an unpaid job" for x-many hours before they're allowed the "fun" benefits package? Especially when those "x hours" can be wiped-clean at the drop of a hat.

As above on the "small group of players" You've built it, some are coming, but how much of the "balance tests" were actually announced to the community-at-large? For me, I've only really heard of "balance tests" on the Dev-build server. That kinda pushed any input I'd have made(on the "normal" server) out of the way, as I do not recall hearing about them. The same is most-likely true for a LOT of the player base. Do think on it.

 

[AndyP]

There are for sure a lot of changes we can make.

We did test some of them, and for sure many basic problems can be solved via blockconfig,
but only on the surface.


Examples for changes that look easy but are not:


Jump drive distance boost

Jump drives are used for travelling from A to B, so far so good, but
people jump directly from A to B, without any thoughts on the path and where it may cross.
Enemy territory can be crossed at literally no risk.
Warp-gates are just decorative, not serving any purpose with those boosts.
And yes, I have used and even built many of those chain drive ships.

The real solution would be:
Increase their power consumption by a factor of 10-25.
This would fight the chaindrives and make them useless unless you build a ship specific for that use.
(A bit like the old cloaker ships, you need to dedicate 90+% of the mass of that ship to cloaking leaving only very little payload...)

So jumping a long distance with a combat or cargo ship would be very very slow.

The result of this would be an increased value for warp-gates for long range travel.
There would be actually a reason to set up networks of them.
Players would use them, and pirates have choke points to catch transports and alike.

You would however still be able to dock your ship to a station with a huge power supply and charge like 10 jump drives while docked, and then just fire those drives in a sequence for a fast travel. So I do not dislike the sequential use of multiple drives, I just feel like a ship should not be able to recharge enough of them parallel that you can keep jumping infinitely.

However this is my personal solution for this,
some of us feel like this would be too restrictive gameplay wise,
so there is no real solution for this we all agree on.


Power Regen Proposal

Well, some aspects of that, especially the 'Focus on regeneration' can be achieved by altering the blockbehaviour configuration,
but, that will not solve problems with the general scaling of ships and components that use power.

Currently all power consumers need to go through the bottleneck of having to calculate and compare two floating point values.
While some are 0.1e/s and others can reach 100000e/s without big problems.

So if we want for example add "lights" as power consumer, to allow them flickering on power outages or low power situations, it would add thousands of calculation steps with floating point values.
Heat would then allow "norming" it from like 0.01% to 300% and save it as plain binary value. That speeds up any calculations a lot.
Keeping the 'range' in mind, it also limits the pile of things a reactor has to consider.
So while it looks easy to 'just alter a few values to achieve 90% if the proposed ideas" its in fact not helping at all with much lower level problems this is meant to fix.

As sure:
power generation / 200
power consumption / 100
power cap (soft-limit) / 200
power capacity - stays the same
would effectively make currently working ships still work, but they could exceed their max power usage for a short time, but will be limited to 50% operation time, and need phases in combat where they need to retreat to recharge. But that was not the only point we want to fix.

And the "Feedback loop" you mentioned on this, is because players built their opinions on each others opinions what they understood from the proposal.
The actual proposal was quite vague as we (cant repeat this often enough) just wanted to throw a rough framework on the table and see what players think of it, and merge useful suggestions and points into our closer planning for this.
We actually got a lot of good things out of the feedback from people that replied on the topic
and not to each other deviating more and more from the actual proposal.

Public Balance Testing

Most of those were for trying existing ships and the effect on them, mostly for internal use and trying to reproduce bugs or play with values.
I still wont agree on buffing a system by a factor of like 100 and confirming its overpowered then, 'just for fun', this is kind of wasted time for using development time and resources, that would be something for player-servers or events.
Anyway, it is a good idea to run balance tests in a controlled environment and give them more exposure by some way.
Using main site news for those will for sure be overkill, and forums only will possibly attract only a subgroup of our playerbase.

- Andy

 

[DrTarDIS]

Anyway, it is a good idea to run balance tests in a controlled environment and give them more exposure by some way.
Using main site news for those will for sure be overkill, and forums only will possibly attract only a subgroup of our playerbase.

Just a minor philosophy point: There is no kill like overkilll, especially when underkill doesn't get the job done.

That simplification and norming by values bit seems a bit jargon-heavy in it's explanation. I halfway follow what you're saying, but there's some disconnect in how I understand the jargon you're using.

EG: Thrust&maneuvering, currently determined by thrusternumber(arbitrary complex return per block calculation) devided by 'Total mass' (multiplied by 'volume' for rotation, because space is water in starmade apparently) Could be simplified and compressed in it's calculation as i understand it

Power-dependant blocks moving over to "heat" and a discreet value doesn't jive as a simplification to me. I'd expect any benifits seen to be visible to the lighting and other visuals by putting that same binary conditional on discreet power values as they currently stand: EG if power drops below 40% have the same effects "heat above 100%" seems to propose in your example. Same effect, different name. Code hijackable from the SHP 0.9-0.5 discreet values that are checked.

If in the end Heat, power, etc overhauls is designed to "make a solid cube of systems" equaly as viable/balanced as "default pretty tradeships" The basic ratios of rotation multipliers, thrust to mass, and mass to systems might be the better area to focus. As well how those values tailor to certain archetypes.
If it's designed to "reduce number of systems blocks" (magnitude number or ratio) in relation to ship size (non-systems, magnitude number) finding a easily-balanced existing value upon witch to base dependency between the two "blocktypes" seems like the first step. It's too bad the material parser demands a positive value on SHP, as I had a brain-fart a few days ago I'd love to test: Cannons, reactors, etc as negative SHP would force and/or reward a volume of "hull" with +SHP. (over here, in spoiler tags because i cant test it right now, if you want to read an essay with examples)

Besides that though, I think Your post highlights exactly what I'm getting at:

we as a community and shine as a company don't have a lot of cross-talk when it comes to "how we play"

That's why putting those (to me) relatively derelict servers to use in such a way would be of benefit.

 

[Lecic]

So if we want for example add "lights" as power consumer, to allow them flickering on power outages or low power situations, it would add thousands of calculation steps with floating point values.
Heat would then allow "norming" it from like 0.01% to 300% and save it as plain binary value. That speeds up any calculations a lot.
Keeping the 'range' in mind, it also limits the pile of things a reactor has to consider.
So while it looks easy to 'just alter a few values to achieve 90% if the proposed ideas" its in fact not helping at all with much lower level problems this is meant to fix.

Why couldn't you just have it so lights don't actually consume power and flicker when PowerOutageStatus is True or when PowerCapacity is 0? Or make power have no <1 values, so that those are less of a problem? Also, wouldn't there still need to be what is essentially a power system in the background? A huge gun slapped onto a much smaller ship would surely fill up more of the heat bar, after all.

 

[AndyP]

Why couldn't you just have it so lights don't actually consume power and flicker when PowerOutageStatus is True or when PowerCapacity is 0? Or make power have no <1 values, so that those are less of a problem? Also, wouldn't there still need to be what is essentially a power system in the background? A huge gun slapped onto a much smaller ship would surely fill up more of the heat bar, after all.

It was more an example on the bottneck and where the game would face limits.

The light-calculation for updating all lights very fast would also be a major bottleneck.
IIRC that would be more of an issue.

So lights wont draw their power individually and only go to "off" or "flicker" when they got insufficient power,
it will for sure need some global management and use a key value.

- Andy

 

[Raisinbat]

New 1 week is not long enough to gain proper data on balance changes. It would need to be 2 weeks at least.

Man i can't even finish a ship in that time

The real solution would be:
Increase their power consumption by a factor of 10-25.
This would fight the chaindrives and make them useless unless you build a ship specific for that use.
(A bit like the old cloaker ships, you need to dedicate 90+% of the mass of that ship to cloaking leaving only very little payload...)

So jumping a long distance with a combat or cargo ship would be very very slow.

The result of this would be an increased value for warp-gates for long range travel.
There would be actually a reason to set up networks of them.
Players would use them, and pirates have choke points to catch transports and alike.

Omg yes please do this

 

[DrTarDIS]

Man i can't even finish a ship in that time

I Somewhat envision an "example starter package" would be part of the test. Something that emulates already having spent your 10 hours of infrastructure building.
A "normal" server starting package is usually "some blocks, a helmet, a light, and a handgun" like you're some pioneer in the wild west with your six-shooter, looking for the gold rush.
A "test" package on a survival server would include one or more "starter ships" and "starter stations" that showcased some ideal "you've got it made" goalpost you should pass before the balance tests really mattered to you.
a :test" package on battlemode would be "ideal archtype ships for various team roles" that you choose from round-to-round.
In both cases players making/retrofitting existing blueprints of their own design would be "bonus data" outside the envisioned context, but still useful.

The key points are to encourage communication both ways, while putting more of the "find what's broken" burden on a wider playerbase.

 

[Lecic]

I Somewhat envision an "example starter package" would be part of the test. Something that emulates already having spent your 10 hours of infrastructure building.
A "normal" server starting package is usually "some blocks, a helmet, a light, and a handgun" like you're some pioneer in the wild west with your six-shooter, looking for the gold rush.
A "test" package on a survival server would include one or more "starter ships" and "starter stations" that showcased some ideal "you've got it made" goalpost you should pass before the balance tests really mattered to you.
a :test" package on battlemode would be "ideal archtype ships for various team roles" that you choose from round-to-round.
In both cases players making/retrofitting existing blueprints of their own design would be "bonus data" outside the envisioned context, but still useful.

The key points are to encourage communication both ways, while putting more of the "find what's broken" burden on a wider playerbase.

Resources aren't enough. You would need time to build actual ships for use under a new config change...

 

[DrTarDIS]

Resources aren't enough. You would need time to build actual ships for use under a new config change...

A "test" package on a survival server would include one or more "starter ships" and "starter stations" that showcased some ideal "you've got it made" goalpost you should pass before the balance tests really mattered to you.

I'm not sure where these two points differ?
Edit: Making the implied explicit: What do you think would not be covered by having some standard ships(plural) package that get minorly changed week-to-week to reflect the direct the devs WANT the game to progress in?

 

[Raisinbat]

I'm not sure where these two points differ?
Edit: Making the implied explicit: What do you think would not be covered by having some standard ships(plural) package that get minorly changed week-to-week to reflect the direct the devs WANT the game to progress in?

because you have to build ships for the new config, or are you suggesting devs should waste their time making ships for us?

I started thinking of values that would warrant a rebuild but i think what DOESN'T require a rebuild...? If you change any damage value, gotta rework all dem guns... Change any power value and gotta redo the power... I guess shields don't really warrant much change?

 

[Nickizzy]

Case in point why I think it should be put in the launcher in a more visible location. "local play" "online play" ->"quickplay(battlemode)" "quickplay(Balance weekly)"

Most of all, I agree with this. I never even heard of these servers before. The problem with server lists is that it's super hard to make any one stand out, and you end up with hundreds of servers that you can't figure out. I just find a server with a large player count usually, which really promotes the LvD and LazyGamers servers that have been around a while and pushes back new ones or ones that empty out a lot.

 

[DrTarDIS]

because you have to build ships for the new config, or are you suggesting devs should waste their time making ships for us?

I started thinking of values that would warrant a rebuild but i think what DOESN'T require a rebuild...? If you change any damage value, gotta rework all dem guns... Change any power value and gotta redo the power... I guess shields don't really warrant much change?

What doesn't: anything not screwed with in "this week's change"
I suppose I wasn't clear on one assumption: The devs have some "ideal concept" of a ship that every change and balance is designed to work for/towards.
If I ASSuME that that is indeed the case, a ship-or-station designed to be "ideally balanced in it's role" for each role type (miner, transport, escort, defence, attack etc) should persist between changes to config with very minor reworks.
Essentially: The NPC ships serve their role function. They would be a very good starting point for "starter packages." Add in a couple "edge of the envelope" classes (sacrifice all else for speed{or survivability, or damage output, etc} while still maintaining the "ideal cosmetic ratio" or whatever) for the supplied blueprints should minimise time spent placing clusters of cubes and maximise time spent "testing the buckets-and-shovels in the sandbox"