Testing Economic Balance Changes

A forum for chatting about in-development game features.
zolobolo
Posts: 1544
Joined: Fri Nov 25, 2016 3:49 pm

Re: Testing Economic Balance Changes

Postby zolobolo » Tue Jul 02, 2019 9:14 pm

I like the direction that market income decreases more with understaffing
Saw that markets only provide 1 trade route on level 1 - that seems like a low value but its starting tier so might be ok on the long run

The above should dampen the effect of market spam on small planets
Please consider giving markets more of a punch when they are fully staffed though so we are tempted to use them on the most valuable planet spaces

User avatar
sven
Site Admin
Posts: 1620
Joined: Sat Jan 31, 2015 10:24 pm
Location: British Columbia, Canada
Contact:

Re: Testing Economic Balance Changes

Postby sven » Tue Jul 02, 2019 9:30 pm

zolobolo wrote:- Did not attack the AI to see how far they can/will build up when left to develop - otherwise the game would have been over at around turn 100


Yeah, this is a real weakness in the current AI -- most of them will focus on colony spamming for most the early game, so if you can get a reasonable fleet early on, they're very rushable. That's less of an issue on the harder modes, because there the AIs' starting advantages cause them to effectively "skip" much of the early game. But, on Normal, it's a pretty serious weakness.

That said, what's interesting about your build is that you didn't actually need to rush to pull this off. I bet the same strategy would be fairly playable in 'Hard' or even 'Brutal'.

zolobolo
Posts: 1544
Joined: Fri Nov 25, 2016 3:49 pm

Re: Testing Economic Balance Changes

Postby zolobolo » Tue Jul 02, 2019 9:38 pm

sven wrote:This is an interesting strategy. I've been worried that destroyer swarms might be too strong in the mid game, but, if you're using them this successfully with tier 1 techs, that actually suggests that the "problem", if any, is that light ship upkeeps may just be too low across the board.

A reasonable assumption but one that will also lead down the road of large-ship-only end of the spectrum. The cocnept of aggregating benefits works both ways in the limited system. You can also take medium hull with Tier I weapons then, and when that is balanced, large hull with Tier I :)

If we say that one can only have n units of various types of elements (A good against B ok against C, B good against C but ok against A, C good against A and ok against B) that counter each other but require close to ~n units to reliably balance each other out, then the solution will likely be to take the one that allows for the extreme utilization of one single aspect (n A). Any suboptimal combination will likely be wastefull (the more elements there are the more likely this is) and you can only counter it if prepared for specifically which is not possible due to low unit count.
nA > n/2A+n/2B
nB < n/2A+n/2B
nC > n/2A+n/2B
If you have high enough count all elements that can concentrade on any location, the system gets much more unpredictable

User avatar
sven
Site Admin
Posts: 1620
Joined: Sat Jan 31, 2015 10:24 pm
Location: British Columbia, Canada
Contact:

Re: Testing Economic Balance Changes

Postby sven » Tue Jul 02, 2019 9:56 pm

zolobolo wrote:
sven wrote:This is an interesting strategy. I've been worried that destroyer swarms might be too strong in the mid game, but, if you're using them this successfully with tier 1 techs, that actually suggests that the "problem", if any, is that light ship upkeeps may just be too low across the board.

A reasonable assumption but one that will also lead down the road of large-ship-only end of the spectrum.


Well, in theory, the game could have really increased your upkeeps when you started upgrading all those CL's to fusion missiles; on the theory that the higher tech systems ought to cost more to maintain. That mass refit also got you a huge increase in attack power, so a serious upkeep consequence might have been "fair". And if I was very careful with the details of how a part-based upkeep model worked, I might be able to implement in a way that would merely reduce, rather than destroy, the late-game viability of small ships.

That said, I'm not sure if an upkeep spike following the fusion missile mass refit would have been "fun". If I really tie upkeeps to part maintenance, some players will end up in situations where they'll find themselves suddenly running huge deficits after doing a "refit all". I think that would prove unintuitive for most players, and while a few might learn to enjoy the mechanic, I suspect many others would probably just find it frustrating and confusing.

SgtArmyGuy
Posts: 26
Joined: Mon Jan 14, 2019 4:43 pm

Re: Testing Economic Balance Changes

Postby SgtArmyGuy » Wed Jul 03, 2019 6:01 am

sven wrote:If I really tie upkeeps to part maintenance, some players will end up in situations where they'll find themselves suddenly running huge deficits after doing a "refit all". I think that would prove unintuitive for most players, and while a few might learn to enjoy the mechanic, I suspect many others would probably just find it frustrating and confusing.


Solvable by expanding the info text of the "refit all" -button? "Refit all ships in the current fleet with a one-time cost of x metal, x coin and increasing their upkeep from y to z?"

I think it's pretty intuitive that better gear costs more coin. Just show the numbers to the player and nobody will be confused!

//Edit: As an added bonus, maybe it would be possible to transfer some of the refit costs into upkeep this way? As zolobolo and I have pointed out, refitting is too expensive in most situations under the current economic balance. It needs to be cheaper to be viable. Increase in upkeep could make the ship cost more in the long run, but the initial cost could be very minimal. I'd like this very much!

Dragar
Posts: 118
Joined: Wed Mar 21, 2018 12:20 pm

Re: Testing Economic Balance Changes

Postby Dragar » Wed Jul 03, 2019 9:09 am

You need to be a bit careful that you don't fall into the surprisingly common trap of making it cheaper to build an empty hull and then 'upgrade' than it is to build from scratch.

Uncle_Joe
Posts: 44
Joined: Wed Jul 19, 2017 10:55 pm

Re: Testing Economic Balance Changes

Postby Uncle_Joe » Wed Jul 03, 2019 4:29 pm

You also don't want to go through the Stellaris screw-up where 'naked corvettes' are better than the high-end ships due to costing a fraction in upkeep.

I don't think that is a rabbit hole worth going down. It opens up cans of worms with trying to balance between upgrading and keeping ships un-upgraded and having more of them. I think in a game like this it should always be beneficial to upgrade since it not only costs metal/money but also the research points (which is an opportunity cost since instead of building Labs you could build more Factories/Mines/Markets etc).

I'd leave it alone.

User avatar
Zoolimar
Posts: 16
Joined: Sat Jun 29, 2019 11:42 am

Re: Testing Economic Balance Changes

Postby Zoolimar » Wed Jul 03, 2019 4:46 pm

You also don't want to go through the Stellaris screw-up where 'naked corvettes' are better than the high-end ships due to costing a fraction in upkeep.

Technically it is already the case in a certain sense. Torpedo Destroyer is probably one of the best ships in the game. And most of the time it is running around without shields or top level reactors.

If we say that one can only have n units of various types of elements (A good against B ok against C, B good against C but ok against A, C good against A and ok against B) that counter each other but require close to ~n units to reliably balance each other out, then the solution will likely be to take the one that allows for the extreme utilization of one single aspect (n A). Any suboptimal combination will likely be wastefull (the more elements there are the more likely this is) and you can only counter it if prepared for specifically which is not possible due to low unit count.
nA > n/2A+n/2B
nB < n/2A+n/2B
nC > n/2A+n/2B
If you have high enough count all elements that can concentrade on any location, the system gets much more unpredictable

I'd say this is not exactly how it is in SiS due to how ship design is setup. There are almost no times when you can't fill most roles needed in the fleet, except maybe in the very beginning due to the fact that you don't have the tech yet. After you get RF lasers and Heavy Coilgun you are more or less set. Ships do have preferred roles but can be moved pretty far out of them if the need arises.

User avatar
sven
Site Admin
Posts: 1620
Joined: Sat Jan 31, 2015 10:24 pm
Location: British Columbia, Canada
Contact:

Re: Testing Economic Balance Changes

Postby sven » Wed Jul 03, 2019 4:48 pm

Uncle_Joe wrote:I don't think that is a rabbit hole worth going down. It opens up cans of worms with trying to balance between upgrading and keeping ships un-upgraded and having more of them. I think in a game like this it should always be beneficial to upgrade since it not only costs metal/money but also the research points (which is an opportunity cost since instead of building Labs you could build more Factories/Mines/Markets etc).

I'd leave it alone.


Yeah, I'm reaching the same conclusion.

I do think there's probably still room for a bit more balancing between the different hulls (as evidenced by yesterday's patch notes), but that's really a separate collection of issues.

User avatar
sven
Site Admin
Posts: 1620
Joined: Sat Jan 31, 2015 10:24 pm
Location: British Columbia, Canada
Contact:

Re: Testing Economic Balance Changes

Postby sven » Wed Jul 03, 2019 5:31 pm

SgtArmyGuy wrote:As zolobolo and I have pointed out, refitting is too expensive in most situations under the current economic balance. It needs to be cheaper to be viable. Increase in upkeep could make the ship cost more in the long run, but the initial cost could be very minimal. I'd like this very much!


Dragar wrote:You need to be a bit careful that you don't fall into the surprisingly common trap of making it cheaper to build an empty hull and then 'upgrade' than it is to build from scratch.


The way things are balanced right now, refit costs are in a kinda messy place. One delicate detail here is how "labor costs" are converted to "coin costs" when you do a rush buy or an instant refit. Right now, rush buying incomplete production converts coins to labor at 3 to 1, but if you instant refit a ship, you convert coins to labor at 2 to 1. In some senses that's probably wrong; because it does mean there are perverse incentives to do strange things involving building and then refitting empty hulls.

In practice, no one seems to be "exploiting" this mechanic, but, it's worth noting that there is a sense in which I've already pushed refit costs lower than I probably should have.

That said, there is still some wiggle room to adjust the numbers here. Right now, refit costs are actually defined as a weighted combination of two different numbers -- the first number is based on an approximation of how much work it would actually take to change the ship from one design to another, which the second number is simply the difference in the cost of building one design vs. building the other design. The second number is less "realistic", but, often far lower than the first.

Arguably, what I should do is increase the coins to labor exchange rate used in refitting to match the rush buy formula, but then change the refit formula so it gives more weight to the "simple cost difference" cost estimation. I'm not sure if it would really end up making the refit costs noticeably lower in practice, but, it would probably be more "correct" than the current math.

User avatar
sven
Site Admin
Posts: 1620
Joined: Sat Jan 31, 2015 10:24 pm
Location: British Columbia, Canada
Contact:

Re: Testing Economic Balance Changes

Postby sven » Wed Jul 03, 2019 5:52 pm

sven wrote:Arguably, what I should do is increase the coins to labor exchange rate used in refitting to match the rush buy formula, but then change the refit formula so it gives more weight to the "simple cost difference" cost estimation. I'm not sure if it would really end up making the refit costs noticeably lower in practice, but, it would probably be more "correct" than the current math.


Yeah, after a little consideration, I've decided the numbers here really should be cleaned up. r38143 includes a diff that does exactly this. In the cases I've looked at, it ends up decreasing instant refit costs fairly significantly, while "manual" refit costs are decreased even more significantly. I'm a little concerned that I may have overdone things with some of the weighting formula changes -- there's a chance I may roll these back a bit after doing some playtesting.

zolobolo
Posts: 1544
Joined: Fri Nov 25, 2016 3:49 pm

Re: Testing Economic Balance Changes

Postby zolobolo » Wed Jul 03, 2019 9:26 pm

New game with normal difficulty, oval small galaxy with default settings:
- Upkeep of Ashar Light Cruiser is not higher then that of Escort Carrier which is fair even if a bit strange (due to size difference and later being much furhter in the tech tree). I do imagine the Ashdar as having the Liht Cruiser as an endgame tech left over from the great war so it fits in with that concept as most what they produce is weaker in the end I like this concept and roll with that :)
- Ashdar Missile Cruiser if finally usefull: either with missiles or torps, both give it enough of a punch to be worth the time to build it!
- Reduced the number of mines I build and placed research labs and markets instead. First leads to faster tech progress, later to me being able to produce ships again which is strange: I do not actually fele the limit (which I do not mind at all) due to this, but the AI has some issues:
- AI does not seem to be able to build ships but its not clear why. Thely mostly seem to have the money and metal
- Especially in early game (first 100 turns) they lack the number of ships necessary to defend their territory: it is very easy to attack them in a chosen location and after they loose a coupel fo ships they are effectively crippled
- Upgrade of AI ships seems ot be working well: I do not see outdated ships hanging around for long
- Refit is crazy cheap in the beginning but quite expensive later so I would say its ok
- Same goes to mid-game: they tend to form a strike fleet with all their ships leaving theit entire empire defensless
- Still: I think they should not offer up colonies: getting a fully developed colony (or even 3) is a huge advantage to the player as no time is spent amassing tanks, and no pops/improvements are destroyed due to bombing
- Ashdar Fleet Carrier is ok, it seems to fit with the overall tectics of the race better then its tanky past form. It did not get another hangar, but the price reduction must have helped as it is on par with the mid sized carrier
- Planetary Defenses are powerfull as expected and can hold a planet realtively well into the mid-game on their own withouth starbase

I have to defend the Yoral torp destroyer though strange that it might sound: First the number of torps matches the art which is reason anough to have 8 of them :), but there is another reason and that ties into the other obviously OP tactics like Tinkers+missile only config, Ashdar Light Cruiser, Boarding in general, Phidi mercs, or Yoral everything: I begin to see these in a broader concept similarly to 40k

There are OP concepts, but there are so many and enought combination of them to balance each other out. Yes there are huge differences in power level but every faction has access to at least one of these, and if not, they can steal/get it form the others. Even the larger ship numbers reduce the effect of these tactics in a way

Ashbery76
Posts: 99
Joined: Tue Feb 03, 2015 9:04 pm

Re: Testing Economic Balance Changes

Postby Ashbery76 » Wed Jul 03, 2019 9:56 pm

My opinion on 4x economies is that it should always feel like you never have enough.All 4x games only really have this early game.They all end up with massive bloat.

Dragar
Posts: 118
Joined: Wed Mar 21, 2018 12:20 pm

Re: Testing Economic Balance Changes

Postby Dragar » Wed Jul 03, 2019 9:57 pm

sven wrote:In practice, no one seems to be "exploiting" this mechanic, but, it's worth noting that there is a sense in which I've already pushed refit costs lower than I probably should have.


Probably because coin to labour conversion is pretty lousy! But I think you are right to tidy the maths up before it becomes an issue.

User avatar
sven
Site Admin
Posts: 1620
Joined: Sat Jan 31, 2015 10:24 pm
Location: British Columbia, Canada
Contact:

Re: Testing Economic Balance Changes

Postby sven » Wed Jul 03, 2019 9:57 pm

zolobolo wrote:- Same goes to mid-game: they tend to form a strike fleet with all their ships leaving theit entire empire defensless


I've been watching AI vs. AI galaxies play out, and what I'm seeing is that while this behavior happens sometimes, it certainly doesn't happen all the time. And as long as it's not a consistent behavior, I'm ok with leaving it as is. Particularly if the AI is not actively under attack, I think it should be willing to take the occasional "gambit" and send everything that it's got into enemy territory seeking a decisive engagement.

These kinds of huge attacks may not always be strategically sound, but, I think they can be fun to play against.

zolobolo wrote:- Still: I think they should not offer up colonies: getting a fully developed colony (or even 3) is a huge advantage to the player as no time is spent amassing tanks, and no pops/improvements are destroyed due to bombing


The flip side of having an AI that likes to fight big decisive battles is that there need to be reasonable behavior in the case that it loses those big decisive battles. The AI usually doesn't have many resources on hand, particularly if it's just lost a war, so the main things it really could "surrender" in return for peace would be territory or technologies. In a strictly strategic sense, I agree that it's often a bad move for the AI to hand over territory; losing a few colonies to bombing is often far better than seeing those same colonies in the hands of your opponent. But in a world-building sense, I think it makes sense that an alien empire would be willing to surrender some of its territory if it thinks that it's lost the capacity to effectively defend itself.

Getting the balance right of exactly how "desperate" the AI should be for peace at any moment is a bit tricky -- and I do think it sometimes offers too much or too little territory in exchange for peace. That's an area of the code I'd like to further refine; though the formula are already pretty complex, and it's not obvious to me how to improve them.


Return to “Testing”

Who is online

Users browsing this forum: No registered users and 72 guests