Problems with Experimentors

[mem359]

I had two different games with unsatisfactory results, due to the Experimentors.
But I don't think these were intended.

1) Early in one game, I lucked out that I had a large attack and troop fleet land on the Experimentor colony. After beefing up the scanners, I was able to take them out... Unfortunately, this colony was on a bottleneck from the quarter of the galaxy I occupied, and the rest. Since the star lanes around this system were gone, there was no way to reach the other (AI) players or travel anywhere else.

It would be nice if the star lanes that were removed by the Experimentors would "regenerate" at some point after they are defeated.

(This game was a while ago, so I don't know if the map-making algorithm has changed to avoid this possibility.)

2) In my most recent game, the Experimentors had a stealth of 240. (See attached JPEG file.) Since the maximum scanner level is 210 (as far as I know, and that is only if the +10 planet special is around), there is no way to stomp them.

Either cap the planet stealth at 200, or allow some way to increase the scanner level locally. A scanner facility at their doorstep should do something. (I would have been willing to build a fractal hull with nothing but external Sensors, if it would have helped.)

[MatGB]

Re the former, yes, I agree, except there is now a patch that prevents them from putting starlanes down until after turn 200 and capturing them now triggers a victory sitrep, but I do tend to keep on playing, and it's annoying having a high level fleet stuck there. OTOH, you don't need a fleet there, no shields means just troop ships can take them.

I'll look at scripting some starlane creation for the outpost if it's captured, but it is an edge case and you have won at that point.

For the latter, ouch, I've never had that happen, bad luck, OTOH AFAIK it can only happen if they get hit by the higher end stealth granting clouds, which can only trigger if at least one empire has researched the tech taking that much stealth. That's probably why it's never happened to me, if it had I'd have probably fixed it, 3 Experimentor related patches from me in 0.4.4 all fixing bugs or making them less annoying.

Distortion Modulator can reduce the stealth by 20, an Intersteller Lighthouse in the same system can reduce by a further 30, but the latter is hard to pull off.

I suspect the best thing to do is simply reduce the outpost stealth bonus to something like 150 so that it would naturally max out at 200—you can do this in buildings.txt to rescue that game, found in the default folder, I'll get a patch together to do it for Trunk at some point soon.

(as it might be related, I came to the forum tos tart a thread on stealth and bonuses for various ships as I'm working through a balancing patch for several hulltypes that use stealth, that might be of interest)

[mem359]

Thanks for the feedback, MatGB.

In the first case, it wasn't that a large fleet was stranded at the Experimentor colony (although that is never fun). It was the particular configuration for this galaxy:

75% of the galaxy --one-starlane-- Experimentor colony --one-starlane-- 25% galaxy that I conquered

I had taken over the Experimentors around turn 100, so I was stuck doing a tech-race for the rest of the game. (As soon as I realized this, I started a different game. So be it.) As you point out, this situation was several patches ago. I just wanted to verify that the galaxy-creating algorithm checks that the Experimentors are never on a (major) starlane bottleneck.


For the second case, as you guessed the game was winding down anyways. I was black kraken stomping while working through the last of the AI colonies, when I noticed the strange situation with the Experimentors. So it wasn't crucial, but there might be some other game where a feature like this might be more important.

[MatGB]

They don't technically get created by the galaxy generation algorithm, they's, I think, spawned as part of monster generation which happens at the start of turn 1, and there is no check for that, not even sure there can be—they do check to ensure they're not too close to a an empire homeworld, beyond that I personally couldn't script anything better.

However, having the outpost place starlanes on destruction is probably a good plan, but not sure it'd guarantee prevention of the problem—I have had similar where they cut off the tail end of a spiral arm, I ended up conquering that area due to a completely different bug in which Psychogenic Domination was messed up due to something wrong in the system id settings, so you could Dominate ships that weren't anywhere near you, I took control of 1/3rd an Eaxaw fleet and a single troop ship in the bit I couldn't get to and that was enough to take a planet, build a Transformer and stargate the fleet in. Very very silly, but it was part of the prompting to get the pre-turn 200 thing fixed and committed.

And yeah, I'm going over various stealth affects at the moment, changing the outpost to a lower score would be easy and useful, it's a very rare edge case but it'd be nasty for a first time player, don't want to lose people pointlessly over that.

[Dilvish]

A graph analysis condition for checking if a set of starlanes cuts the universe like that could be made, but probably won't anytime soon.

[Vezzra]

I just wanted to verify that the galaxy-creating algorithm checks that the Experimentors are never on a (major) starlane bottleneck.

As Mat already pointed out, the Experimentors aren't placed by the universe generation functions, but by content scripts. Universe generation is already partly scriptable, but not enough to move Experimentor creation completely into it. Once that's possible, I'd suggest doing so, so issues like that can be dealt with (and it also allows way more flexibility to set up something like the Experimentors).

[MatGB]

Given they are meant to be a victory condition, they should be part of the generation script and options I think, but making them so isn't high on the list I'd want people able to do the coding to do, it's already mostly fixed, I think I can code a few starlane generation things when the outpost is destroyed, I'll give it a go when I've finished my current stealth messing.

[mem359]

I recently had the Experimentor system generate a new star lane after it was conquered, so that coding work is appreciated.

I found a good example of the bottleneck problem, as shown in the attached image. (Experimentors on Rauros.)
Fortunately, it was only an AI that got trapped.
(I understand this is low priority problem; this is just for documentation.)

[MatGB]

As Mat already pointed out, the Experimentors aren't placed by the universe generation functions, but by content scripts. Universe generation is already partly scriptable, but not enough to move Experimentor creation completely into it. Once that's possible, I'd suggest doing so, so issues like that can be dealt with (and it also allows way more flexibility to set up something like the Experimentors).

My suggestion would be to stop treating them as a 'monster' that spawns, creates the outpost building and populates the planet with experimentors, and instead have them created alongside the AI homeworlds, preferably before the AI, roughly in the middle of the map and with distances set in the swame way as the AIs are.

If they're in the middle, they'd be far less likely to create detached clusters as pictured and also, while it'd mean experienced players know to expect them and roughly where, that's not necessarily a bad thing and it would reduce the chances of them spawning at the extreme opposite side of a large galaxy and wiping out the nearby systems before the player(s) have even found them, etc.

However, I'm still learning to read Python, let alone manage to script in it, so I'll leave that idea to those that know that code better'n me (and read it/test it to improve my skills with it).

@Mem, that is a good example, thank you, haven't tried to replicate it but I'll assume it works roughly as is (there's an element of randomness above the seed so it mmight not replicate exactly but it's good to know—Irregular 2 means a very recent test build, I assume #7311?)

[mem359]

@Mem, that is a good example, thank you, haven't tried to replicate it but I'll assume it works roughly as is (there's an element of randomness above the seed so it mmight not replicate exactly but it's good to know—Irregular 2 means a very recent test build, I assume #7311?)

Ah yes, it was the Mac build 7311 that Vezzra posted yesterday.

(I'm never quite certain about the "randomness above the seed", but I figured it couldn't hurt to list what I used to generate the galaxy.)

[MatGB]

No, it's definitely helpful, but the seed affects universe generation, monster spawns are done after that and while they're normally similar, they're not always the same. And beyond that I'm not sure why or what is and isn't affected, I just know what I've observed.

[asm64]

Hello. I want to request some things about balancing. So if we start game with all low options but about 200-300 star systems, grow in the game process going too slow because planets are placed very rarely. But Experimentor outpost always appears about 200 turns. I propose to bind to the total number of planets (without gas giants and asteroid belts) and the total number of star systems. Something like:

experimentor_appear_turn=k*(starsystems/planets)

[Vezzra]

Hello. I want to request some things about balancing. So if we start game with all low options but about 200-300 star systems, grow in the game process going too slow because planets are placed very rarely. But Experimentor outpost always appears about 200 turns. I propose to bind to the total number of planets (without gas giants and asteroid belts) and the total number of star systems. Something like:

experimentor_appear_turn=k*(starsystems/planets)

Sounds reasonable. What do the others think?

[Mitten.O]

I think the eventual actual difficulty settings will let players do this manually, if they want. Seems better than arbitrarily scaling it. Those wanting a challenge may want to try an early experimentors, low density -combo.

[Vezzra]

But unless we get that, we could put some simple formula in as a stop-gap solution (it's just a few lines in the content scripts after all, no major production). When I want to play on a low planet map, and also want some monsters, that doesn't mean I also want the significantly increased challenge of the experimentors showing up after the same number of turns as they would on a high planet map. However, right now I can't have that. I only get far more monster fun I'm asking for, or no wandering space critters at all.

Once we put in some more setup options to fine tune these things, we would of course take out the formula.