Dominus Galaxia KickStarter is live. Plus, free beta build!

[haphazard1]

That said I may refactor the random number stuff in the future to make things fully deterministic, or even have a toggle like 1oom.

An option to control this (deterministic or not) in set up would be very good to have. The wide range of game set up choices is a very positive feature of DG.

[Zed-F]

Maybe... but don’t fall into the trap of thinking more options is always better; choice paralysis is a real thing, and more options means those options have to be tested, supported, and maintained, which can take resources away from other priorities. The feature already has an on/off switch and sometimes it’s better to plant a stake in the ground and say ”this is the best implementation I can reasonably do of feature X, use it or don’t.”

I tend to prefer adding options that allow players to control aspects of the game that can’t otherwise be controlled (e.g. the handicapping system I mentioned earlier) rather than increasingly fine tuning of controls for something that already has one.

[Jeff Graw]

Maybe... but don’t fall into the trap of thinking more options is always better; choice paralysis is a real thing, and more options means those options have to be tested, supported, and maintained, which can take resources away from other priorities.

With the former, that's why most "choices" are only exposed in data. So, inside of Variables.txt there are a number of interesting things you can change like min and max number of random techs per rung, the size of each rung, the penalty for travelling outside of star-lanes, the distance between stars, and so on. These could all be exposed in-game, but that would be overwhelming so some curation is necessary.

Another cool thing that you can control is the animation speed of space battles. This will be exposed in-game though after I make it so certain background animations maintain a constant speed.

The later is certainly true. For the time being, I'm planning to spend the majority of my time finishing off and polishing the core, necessary, tasks as opposed to going on new excursions.

[RefSteel]

It’s very possible to get crazy powerful results out of the RNG when designing ships using prototyping.

Wait ... what?

If I build a ship with 2 neutron guns, and the RNG decides I get to have 3 instead

Does this mean...?

Don’t like the design you made due to RNG screwing it up?

...

So it's not just a prototyping cost/delay; it's not just that I'm forced to use default designs to start the game without paying a penalty; it's that I don't ever get to use any ship I design at any point in the game unless I luck into a "no changes" result occasionally?  It would actually be amusing to me if the most efficient way to upgrade your ships' combat capabilities is often to ignore techs that unlock new weapons, shields, and battle computers (which in DG are side research projects, not ways up the tree) at a given rung (or even to skip the fields entirely, depending on your expected victory horizon) in favor of running multiple prototypes in parallel to fish for powerful random mods.  (This "strategy" has the advantage, among others, of effectively collapsing the "research cool new thing" and "wait for cool new thing's prototype" times and costs into one!)  Fortunately, the whole prototyping system can be turned off; that's good.  Probably there are players who will like it because sometimes they get a bonus out of it and, as Nero Wolfe once said, "We are all vainer of our luck than of our merits."  (In "The Rubber Band," by Rex Stout)  So if I play DG, I can just click the "Please, no" box for this one, and those who like it can enjoy it.  But I do want to illustrate:

News bulletin:

The new "Greased Lightning" supersonic ramjet fighter commissioned by the United States of Dominus Terra is finally ready for production.  Designed by Notorion Aerospace Industries as the USDT's answer to the Red Bear Republic's MiG 974,573 fighter-interceptor, the "Greased Lightning" will be the first USDT fighter capable of matching the latest MiG's top speed of Mach 1.8 under combat conditions, and will carry up to 700 lbs of ordnance on wingtip weapons mounts.  According to a spokesman for Notorion, "Well, after submitting our design bid, we brought in some consultants from the partnership of Mike Rancid, Tom Nefarious, and Jacqueline Guilty, and the pRNG sent the designs back to us with some minor changes.  The Greased Lightning's top combat speed is actually just Mach 1.2, but technically that's still supersonic, even if the new MiGs can fly circles around it, and now it can carry up to half a ton of ordnance, with a third hardpoint underneath the fuselage!  The pRNG said it was way cooler this way, so we went ahead and tooled up all our factories to build it this way instead, and didn't bother to tell anybody about it until now that it's production-ready, because why would they need to know, really?"

An officer of the USDT Air Force who wished to remain anonymous had this to say:  "We were going to have some Greased Lightnings built after hearing the news, but then we realized our existing long-range reconnaissance aircraft - although they famously turned out to be next to useless for long-range reconnaissance - are capable of carrying more than enough munitions to destroy our greatest national enemies:  The corporate offices of Notorion Aerospace Industries and the pRNG."

I would advise you never to build in any substantive changes just to prevent save-scumming: Trying to prevent it is pointless; if people want to save scum, either they'll find a way to do it, as Zed-F explains for this case, or they'll enjoy the game less; either way, there's no benefit for those of us who never do it.

I'm loving this in-depth discussion, guys. Even (and often especially) the critical parts!

I'm glad to hear it!

We may have a minor difference in philosophy here, or at least priority. I actually somewhat agree with you that the sliding scale of costs makes guesstimating the optimal point more difficult, although we probably disagree on extent.

The reason that costs scale as they do is to avoid the all-too-obvious inflexion points that step functions create.

Yes, it's a difference in philosophy:  Something this discussion made me realize (somewhat to my surprise, actually) is that I believe obvious inflection points are a good thing for this type of game.  The question, "How many factories should I build before they stop being worthwhile?" is just a math problem, and not very interesting strategically.  As far as possible, I want my decisions to be based things like, "Do I want to get [Important Project X] out ASAP, or do I want a better manufacturing base first so I'll be better able to react to emergencies?  Do I want to concentrate as much early production as I can at one location, or spread out more production around the empire?"  Of course, each answer will usually be, "A little bit of both," and the real decision comes down to how important each is relative to the others.  The optimal means of achieving those relative priorities will come down to a math problem anyway, but I believe that it's better for a game when the factors in that problem are transparent and easy to understand.  This is actually one of MoO's weaknesses out of the box (the pop growth formula is anything but transparent until you've read up on it a bit or learned good rules of thumb from the forums here, and pop growth is ultimately a major driving force for the overall strength of your empire in the game) but it looks to me as if the same weakness is present and amplified in DG with factories and waste effects.  Philosophical differences aside though, it may be easier to get a feel for when you want to switch projects (given the priority on which you've decided) than I think.

(Also, given another big philosophical difference, it may not matter:  For instance, it may be optimal to just set all the ratios as you want them for each planet ASAP and never switch projects outside of emergencies, depending on the severity of the game's various switching penalties.)

In practice it's usually the opposite. Travel outside the starlane network exists, but happens at 1/4 speed. This makes for a painfully slow trip, but you can bypass static defences at choke-points. Without starlanes super high warp speeds are often marginal and superfluous. With starlanes, super high warp speeds begin to change the nature of how you move your fleets.

I must have communicated poorly; this is more or less what I was saying - except for the last sentence, which can't possibly be true.  The upgrade I mentioned, from warp 6 to warp 7, has no impact on transit time for the majority of journeys of 40 parsecs or less, and add only one turn to the others!  (And even those are mostly cases of 6 or 7 turn journeys becoming 5 or 6 turn ones...)  You only get a two-turn bonus on a few special trip lengths, all of them 49(!) parsecs or more.  The round-about routes required to follow starlanes and the greater average size of DG galaxies mean there will be more cases of longer journeys, and of course if there's a star you want to bypass so you can't use the starlanes and have to eat the x4 penalty, that makes for even longer trips, but that doesn't change the basic situation.  Of course the difference between Warp 7 and Warp 2 is huge; high-warp engines certainly have value - my point is that once you have high warp engines, an incremental increase in speed isn't very exciting, which means that when you start at Warp 2, there are fewer exciting upgrades before you reach the point where each upgrade has little or no effect on transit times.

[Jeff Graw]

I would advise you never to build in any substantive changes just to prevent save-scumming:  Trying to prevent it is pointless; if people want to save scum, either they'll find a way to do it, as Zed-F explains for this case, or they'll enjoy the game less; either way, there's no benefit for those of us who never do it.

I think that's a false dichotomy. There are players who will always save-scum, and players who never will, but almost everyone is in-between. In general, I think for most people save-scumming reduces fun but is a constant temptation. If you can increase the cost and/or decrease the benefit of save-scumming, then that moves the threshold at which those individuals will decide to save-scum. I believe that the x-turn delay + saving that RNG seed accomplishes that for the majority in the case of ship crafting, although obviously I'm projecting my own disposition somewhat... as in with this implementation *I* don't feel tempted to save-scum. But that's laregely unavoidable.

Philosophical differences aside though, it may be easier to get a feel for when you want to switch projects (given the priority on which you've decided) than I think.

I hope so. There's an easy way to find out

depending on the severity of the game's various switching penalties.)

There isn't any switching penalty for colony management... at least not yet.

I mean, colony management is something that sucks in every 4X game. None of it is particularly strategic because of how internal it tends to be, so you're basically left with the choice of non-strategic, less of a math puzzle (more difficult to screw up), scales well, but boring vs. non-strategic, more of a math puzzle (easier to screw up), scales terribly, but is interesting.

And although you, I, and everyone here prefers the former to the later... the former still really sucks. Colony management is my white whale. I've been playing around with ideas in my head about how to do it better, and actually make it strategic, for years now. Nothing I'm ready to share yet. But if I have the opportunity, I'd like to try and rethink colony management in general in a future version of DG.

my point is that once you have high warp engines, an incremental increase in speed isn't very exciting, which means that when you start at Warp 2, there are fewer exciting upgrades before you reach the point where each upgrade has little or no effect on transit times.

The difference is that in DG you can continue researching higher and higher end warp drives because the tree doesn't end (although you could easily mod it to work similar to MoO 1). So if you play long enough you do start to organically switch from primarily using lanes to primarily using direct travel. Theoretically, a direct path with super high end warp could take less time than using starlanes, since you need 1 turn at each stop when you travel through the starlane network.

[Zed-F]

I would advise you never to build in any substantive changes just to prevent save-scumming:  Trying to prevent it is pointless; if people want to save scum, either they'll find a way to do it, as Zed-F explains for this case, or they'll enjoy the game less; either way, there's no benefit for those of us who never do it.

I think that's a false dichotomy. There are players who will always save-scum, and players who never will, but almost everyone is in-between. In general, I think for most people save-scumming reduces fun but is a constant temptation. If you can increase the cost and/or decrease the benefit of save-scumming, then that moves the threshold at which those individuals will decide to save-scum.

I agree with Jeff on this point. In most cases it's probably not hard for the developer to reduce the cost-benefit of save scumming below the threshold where it's tempting to do, it just requires a little foresight about how the system might be able to be abused and how tempting it might be to abuse it in particular ways. I've pointed out that it might be easier to abuse save-scumming in conjunction with prototyping than Jeff realized; now it's up to him to choose whether to spend a bit of time re-examining his solution to see if it can be improved in this respect.

It would actually be amusing to me if the most efficient way to upgrade your ships' combat capabilities is often to ignore techs that unlock new weapons, shields, and battle computers (which in DG are side research projects, not ways up the tree) at a given rung (or even to skip the fields entirely, depending on your expected victory horizon) in favor of running multiple prototypes in parallel to fish for powerful random mods.

Well no, it works on a percentage bonus/penalty basis for each numerical attribute, so you do need something to start from. But you can pay extra at design time to shift the odds of getting bonuses, or save money on a design at greater risk of getting penalties. Same for the amount of time it takes for the design to finalize. But your proposal would be funny, if also horrifying.


On another topic, I thought I would mention some of the spy/diplomacy issues I encountered in my most recent game. I'll begin with the recap (briefly) and then expand with some other observations. Mostly I think these are just bugs / feature requests, not so much design problems.

• Diplomatic relations (love nub value) aren't preserved across saves. Definitely a problem if I want to make friends.
  
• No way to easily tell spies not to constantly propose theft/arson against neighbors you want to maintain good relations with, if you want passive intel gathering. (I guess just having treaties will give some level of intel though...)
  
• No feedback on what effect defensive spy spending has had. I did notice in my most recent game that detection spending was more effective than prevention spending, but I don't yet have a good feel for whether the cost was worth the benefit. Some kind of estimate of enemy agent penetrations that I could correlate to spending would be helpful.
  
• Closing the embassy with the Xygob did not prevent them from asking for trade agreements, map trades, etc.
  
• I was sure I had something else, but now I can't remember it. :/

Also on another topic, I noticed that in an earlier build of the game there were more difficulty levels; were these removed as redundant or just not in the beta build? What is the level where the AI gets neither bonuses nor penalties, and how big are the AI bonuses/penalties above/below that? Or does the AI level just affect how much AI logic is exercised, and the bonuses are specified entirely separately? I guess I'd just like a bit of an exposition on how the AI difficulty settings work.

[Jeff Graw]

Diplomatic relations (love nub value) aren't preserved across saves. Definitely a problem if I want to make friends.

Fixed in more recent versions.

Spies

Yup, spying still needs a bit more love. I would have done it before entering KS, but ran out of time. During the KS period, I'm trying to limit changes to things that aren't likely to break anything.

Closing the embassy with the Xygob did not prevent them from asking for trade agreements, map trades, etc.

Good catch! Will be fixed in the next version. Although I'll only stop the AI from making proposals in order to prevent abuse (eg. can't close embassy to prevent them declaring war)

Also on another topic, I noticed that in an earlier build of the game there were more difficulty levels; were these removed as redundant or just not in the beta build? What is the level where the AI gets neither bonuses nor penalties, and how big are the AI bonuses/penalties above/below that? Or does the AI level just affect how much AI logic is exercised, and the bonuses are specified entirely separately? I guess I'd just like a bit of an exposition on how the AI difficulty settings work.

There's still a bit of subtle non-production cheating here and there in the KS build, which can mostly be disabled in Variables.txt... although I'd like to get rid of that altogether for the release version.

AI skill level is controlled via a value between 0 and 1.

There are two modes: Adaptive and static.

Adaptive will scale each AI's skill level throughout the game depending on how well that empire is doing, and how well the player is doing. Basically, the idea is that the "challenging" preset should be challenging regardless of your actual skillset. There's a toggle that enables AIs to continue scaling beyond that 0...1 range via production bonuses that can be toggled on or off. An interesting side effect here is that you typically end up with more evenly matched AI opponents, since the ones who are losing will scale up while the ones who are winning will scale down.

Static just sets all the AIs to a specific level, whether that's 0.2, or 0.6, or whatever. You can assign production bonuses separately to this. You can play with fully skilled AIs with half production, or half skilled AIs with doubled production, or any other combination.

[kjn]

I mean, colony management is something that sucks in every 4X game. None of it is particularly strategic because of how internal it tends to be, so you're basically left with the choice of non-strategic, less of a math puzzle (more difficult to screw up), scales well, but boring vs. non-strategic, more of a math puzzle (easier to screw up), scales terribly, but is interesting.

And although you, I, and everyone here prefers the former to the later... the former still really sucks. Colony management is my white whale. I've been playing around with ideas in my head about how to do it better, and actually make it strategic, for years now. Nothing I'm ready to share yet. But if I have the opportunity, I'd like to try and rethink colony management in general in a future version of DG.

I think a key here might be meaningful choices. Most of the choices regarding colonies in MOO regards to how fast a colony will develop, but the actual end results are always the same. You have the ultra poor/poor/average/rich/ultra rich trait, together with size, but you have no as a player to shape how the world will be different from any other world.

SotS suffered from that problem as well: all the differences between worlds were basically built in. Late-game you could specialise worlds using stations, and planning out a station network with (partial) specialisations held some interesting strategic choices, but by the point you had the tech to build stations the game was usually won anyway.

Such a system would approach the Civ style of cities, though could be a bit more abstracted (and skip the citizen management aspect). Like that every world can provide some revenue and maintain small ships, but would need some kinds of infrastructure to provide even basic research, maintain larger ships, or build ships. To some degree this can even be abstracted away, e.g. by having a mechanism whereby a world gets more infrastructure to do X simply by doing X.

Note that such a system can easily be overdone or opaque if things are made too abstract, but some specialisation of worlds with input from players can provide for plenty of meaningful choices.

[Zed-F]

I don't really agree kjn. In games I have played where entire planets specialize in one thing or another, this tends to feel extremely artificial. I don't actually think it's a problem that entire worlds tend to be generalists; not only does this cut down on tedious busywork of colony management and put the emphasis more on the spaceships, where it belongs, but it's also more appropriate for a game played at this scale. I say leave city management and specialization to other games that have cities.

[kjn]

We already have city management in MOO, through population management and factory and colonist growth. And while I agree that I think most worlds should be somewhat generalist, I think there also should be room for some specialisation. The choice to build the fleet in one world or another should have some impact in how the empire develops.

Now, I certainly don't want to turn a MOO-inspired game into Civ. The goal should rather be to where some worlds are better in some aspects than in others, while not making stuff too complex.

MOO attempted a system which rewarded steady research via the triple bonus system, which turned out to be opaque and opened for gaming. But what if the game instead rewarded the player by giving some bonus for funnelling research in a specific field through specific worlds?

Of course, all of this need to be carefully designed, but I think meaningful and long-term choices are a big plus in a 4X game.