RBTS9 - Final Frontier

[Mojoqmeyvam]

I say go for the Forge, and then Lu. A few battleships don't scare me, and it's good logistically.

Second Green Economy. I've never run it before, normally preferring Industrialism's trade, and I'd like to see how good it is.

[Olodune]

If we plan on going after the Avowers, I think we should plunk down a starbase in the middle of their territory, perhaps within range of their capital. With a few intercepting fighters it would be interesting to see if they can take it down. Just make sure not to place it in an asteroid field so we can get the full weapons range.

[T-hawk]

I think I'd prefer bad-Halis then Avowers. Forge hasn't done anything to us, while bad Halis was never supposed to exist. That's two votes each for Forge and Halis, leaving sooooo as the swing vote. And coincidentally, the guy who will line up our ships at the first target.

Ooh, I like the idea of a forward starbase in Avower territory. Not sure if there's a single location with squadron range to all four core systems, but three should be doable. We don't have many conships south to build lanes and the starbase, but Xanadoo can knock out some and the group by Eden's starbase will get down there eventually. We can't build it within shooting range of the capital; it's expanded to influence radius 3 so we have to build in the 4th tile, outside starbase range. And I don't know what happens if a city expands influence onto an enemy starbase tile.

Environmentalism and Green Economy is our choice, then. Democracy would make our happy cap much less tight (although sadly it takes another turn of anarchy.) Also we could take Hector as a vassal for happy, he was still offering the option last time I looked. The health cap I think it's fine to blow past and just treat each citizen on the margin as consuming 2 food. With Green, then Capricorn Hephaestus is worthwhile (and I think there's another system or two beyond it in the northwest) so have Horus Taurus do a colony ship for it.

Although we're over the size threshold for a 2-for-1 civics change, I calculate we CAN get a 3-for-2 revolution. Anyone like the idea of Cap Ship Doctrine / Democracy / Green Economy? Or should we wait one more tech to Utopia for a triple-revolt?

[Olodune]

Originally Posted by T-hawk

Although we're over the size threshold for a 2-for-1 civics change, I calculate we CAN get a 3-for-2 revolution. Anyone like the idea of Cap Ship Doctrine / Democracy / Green Economy? Or should we wait one more tech to Utopia for a triple-revolt?

I would go ahead and switch now -- then switch again when we get Utopia. Utopia is probably only a civic that we want to spend 20-40 turns in before switching back out. All other aspects being equal I think Green Economy and Utopia are two of the more interesting civic choices in FF, so let's try them out

By the way: How exactly does Utopia work? I was assuming it was essentially similar to Nutrition Facilities ... but it could be a flat +<# of planets in the system> too. The latter would be better for newer colonies.

[sooooo]

I say we go after all 3. Conquer the orange civ, take back the starbase and Exchange from the red civ, then go and bash the light blue civ.

[Olodune]

Originally Posted by sooooo

I say we go after all 3. Conquer the orange civ, take back the starbase and Exchange from the red civ, then go and bash the light blue civ.

Both fake Halis and The Forge could be considered Orange (Forge has the white lettering, while fake Halis has a pukish green color). I don't see any reason to take down the Syndicate (red) -- but with jump lanes in place the back-and-forth logistics might not be much of a problem.

[T-hawk]

If Red Syndicate is on our hit list at all, we should do that first. It'll take less force to drop a starbase than a capital star, plus our forces are right by the starbase now. Then we can build our own starbase to claim the titanium to keep going.

[T-hawk]

I had a look in CvAI.py, in functions doCityAIUpdate and doCityAIProduction. Here are some interesting facts. (For each item, add a disclaimer of "If I'm reading the code right." )

One big shortcoming is how the AI assigns pop to planets. It recalculates its population assignments to planets for every city every turn. In doing so, it seeks a food surplus of only +3 (which is not enough against FF growth costs) and prioritizes hammers after that. It never prioritizes commerce planets! That's why it researches so poorly. Compare to our Paradise that jumped on its 1-1-6 "gems" planet ASAP.

Another shortcoming is how it selects planets to build nutrition/mining/commerce facilities. For each, it selects the available planet with the largest CURRENT production of that type. This is wrong! Planet size should be the determinant here. For a food facility, the AI will pick a size-1 2-2-1 planet over a size-3 1-1-6 planet. But the facility gives only +1 food on the former and +3 food on the latter!

Mining facilities come out even worse, since the AI doesn't match up build orders with what planets it's actually working. It will build a mining facility on a 0-2-5 blue planet ahead of a 2-0-3 white planet, then happily ignore the blue planet in order to work more food.

Each time the AI selects a city build, there is a 15% chance of exiting out with no FF mod logic and building whatever the AI would build in the unmodded game. If the AI is at war, there is another check for a 40% chance to exit. This is an interesting safeguard. If both checks do not exit, then function doCityAIProduction attempts to override the standard city AI by using a weighting system.

The weight threshold (100 minimum) is pretty low in comparison to the rest of the function. Even a single angry citizen is assigned weight 115 which will always result in overriding the standard city AI to build a happy building, unless a higher-weight override is found. It seems very common for the AI to get stuck in overgrowing the tight health and happy caps, especially considering its difficulty in managing starbases to claim resources. So it builds ever more expensive cap lifters to compensate. And the standard AI has the outlet of whipping away angry or sick citizens but that comes too late in FF.

It gives increased weight to health and happy buildings based on the number of sick and angry citizens in the city, which isn't really right. A Sports Arena restores just as much productivity to a city with 3 angry citizens as to a city with 1 angry citizen, but in the former case it will be weighted much higher. This sort of works in the standard game - a city with 3 angry citizens will in sequence build a temple, colosseum, theater - but not for FF where the prize for building one Arena is double and triple the cost to build more.

The AI (at least this function) completely ignores the Value-specific buildings! If it needs happiness, it will build the 280-hammer 7th Sports Arena over the 1st Temple of Worship! AFACT, the only way for the AI to build the value-specific buildings is if the FF AI falls through to the standard AI. It also does not build Stardocks and several other mod buildings except by fall-through.

The rules for starbase construction are very rudimentary. First, the AI doesn't dispatch a conship to a starbase location until it has already saved up enough gold. Then it weights a location mostly by the number of resources reachable by the starbase, which explains why Vlad reached out so far from his homeland for that one. It doesn't distinguish between resource types (we humans know that the cap lifters are far better.) Then there is no logic to dispatch a second conship to build lanes and resource extractors until the starbase is actually complete.

(Let me be clear that I'm not at all bashing Jon Shafer for any AI shortcomings. Final Frontier is still spectacularly well done for essentially a throw-in game mod, and I completely understand that first getting the mechanics right and functional was far more important than polishing the AI.)

[Mojoqmeyvam]

Originally Posted by T-hawk

One big shortcoming is how the AI assigns pop to planets. It recalculates its population assignments to planets for every city every turn. In doing so, it seeks a food surplus of only +3 (which is not enough against FF growth costs) and prioritizes hammers after that. It never prioritizes commerce planets! That's why it researches so poorly. Compare to our Paradise that jumped on its 1-1-6 "gems" planet ASAP.

That explains a lot. However, I'm not sure that the +3 food surplus is a bad idea, given the AI's inability to deal with busted health/happy caps. Prioritizing commerce with hammers, however, would help a ton. Skipping over the "gems" (and to a lesser extent hill silver at 0/2/5) is why we've got a monster tech edge. Oh, and a better REX plan.
Can you tell if the AI assigns all of its population every turn? The demo readings I was getting early game seemed to suggest that they weren't all assigning their pop, which is weird, and really troubling.

Originally Posted by T-hawk

The AI (at least this function) completely ignores the Value-specific buildings! If it needs happiness, it will build the 280-hammer 7th Sports Arena over the 1st Temple of Worship! AFACT, the only way for the AI to build the value-specific buildings is if the FF AI falls through to the standard AI. It also does not build Stardocks and several other mod buildings except by fall-through.

Hm. The value buildings are tough, since they all do such different things. I love the implementation of their variety, it's a shame the AI doesn't know what to do with them. Stardocks aren't great buildings to begin with, so while it would be nice if the AI knew to build them ever, it's not a huge loss. However, the AI does build the defense buildings fairly regularly, a solid choice for any AI (there's a reason Protective is an ok AI trait, but by far the worst human trait).

The Starbase AI is probably irredeemably flawed, I don't know what might make sense, except for a check of "Health + Happy should be worth twice the others".

[T-hawk]

Originally Posted by Mojoqmeyvam

Can you tell if the AI assigns all of its population every turn? The demo readings I was getting early game seemed to suggest that they weren't all assigning their pop, which is weird, and really troubling.

I figured it out. The problem is limited to game start, but it's serious.

I loaded the turn 0 savegame, worldbuildered in a great spy to infiltrate each AI, and clicked through some turns watching them. Every AI began with its population unassigned - that means 0 food and 0 hammers, doing nothing at all. New Earth starved immediately. And the AIs STAYED that way for many turns! New Earth broke out of zombie mode when it discovered Wealth and went through anarchy, which refreshed the city AI logic. Most AIs got going at turn 25 when the capital expanded borders. Poor orange Halis sat idle for 50 turns.

No wonder the player can always out-scout the AIs!

Other systems work OK when founded, it's only the capitals at game start. So it's simple to fix. Need to insert a call to CvAI::doCityAIUpdate (this is the only function that can do FF-aware AI planet assignment) at an appropriate time. A good place is probably CvFinalFrontierEvents::onBeginGameTurn, if the game turn number = 1, then loop through each AI calling CvAI::doCityAIUpdate. I don't speak Python natively but will try to figure it out sometime soon.

[timmy827]

What happens if we build a starbase near Lu before DOW? Will he send his offensive units in the area at it? If so, I'm not sure I'm in favor. I'm ok with trying it out as a secure extra basing for squadrons, but baiting an AI into throwing away tons of hammers (when they will not respond to such an obviously aggressive build while nominally at peace) seems to be crossing the exploit line.

[Olodune]

T-Hawk's code reading explains a lot ... I have a post a while back wondering why the Demo screen is "broken" in FF -- really it looks like it is just because the AI has unassigned citizens at the start!!!

That kind of handicap is quite difficult to overcome (ie Deity in standard Civ for us humans). I will respond more coherently tomorrow.

@ Timmy -- an aggressive SB towards the Avowers could be an exploit if the AI handles it poorly, but isn't it worth a try?

[Mojoqmeyvam]

Also, in many ways we're testing the mod. So let's throw an aggressive SB down there, and see how Lu reacts. My suspicion is he won't throw that much at it, but I could be proven wrong.

[T-hawk]

For a food facility, the AI will pick a size-1 2-2-1 planet over a size-3 1-1-6 planet. But the facility gives only +1 food on the former and +3 food on the latter!

Not quite correct, the latter planet is 3 food total (1 x 3 size) while the former is 2 food total (2 x 1 size) so the AI will correctly pick the latter. So it's not quite so bad since planet size is considered indirectly. The worst case is picking a size-2 white planet (4 food) over a size-3 planet with 1 food (3 total), either gems or 1-2-3 orange.

Originally Posted by Olodune

By the way: How exactly does Utopia work? I was assuming it was essentially similar to Nutrition Facilities ... but it could be a flat +<# of planets in the system> too. The latter would be better for newer colonies.

Yes, Utopia is basically Nutrition Facilities, every laborer produces +1 food. Pretty powerful, and neat that FF provides a civic to interact with food mechanics, something that standard Civ doesn't have.

I'm all in favor of an aggressive starbase at the Avowers. If we start labeling every FF AI hole as an exploit, that list will become a mile light-year long.

Have we decided what to do with civics? If I were soloing, I'd wait the extra handful of turns to Utopia, then triple-revolt to Green/Utopia/Democracy for 2 turns of anarchy. If we adopt Green now, then later Utopia/Democracy/Cap Ship Doctrine (or equivalently the other order, Green/Democracy/Cap now, Utopia by itself later), the net effect is one turn of anarchy for a Doctrine switch and a few extra turns of Green.

That's assuming we want Democracy at all, but with Green/Utopia exploding growth I think we'll need the happy (it's as Representation.)

sooooo <---- UP NOW
Cyneheard / Mojo <---- On Deck

[Olodune]

Originally Posted by T-hawk

Need to insert a call to CvAI::doCityAIUpdate (this is the only function that can do FF-aware AI planet assignment) at an appropriate time. A good place is probably CvFinalFrontierEvents:: onBeginGameTurn, if the game turn number = 1, then loop through each AI calling CvAI::doCityAIUpdate. I don't speak Python natively but will try to figure it out sometime soon.

I like the idea of doing some code edits to try and improve the FF AI -- at least with the low hanging fruit edits. FF is missing many of the features that cause the standard AI to struggle a bit (city placement, civ-wide-unique building optimization (HE/NE etc), specialists, worker-turn optimization), so the AI could be quite competitive.

[shadyforce]

I always wondered why AIs always start so slowly. I assumed it was just because they prioritised getting 5+ PDS in the capital before thinking about expanding.

[Olodune]

Originally Posted by T-hawk

... (or equivalently the other order, Green/Democracy/Cap now, Utopia by itself later), the net effect is one turn of anarchy for a Doctrine switch and a few extra turns of Green.

^^ This has my vote. The doctrine switch would be at the best possible time (3 BBs in the queue). Since Green changes the micro substantially I think switching to it asap will help with building choice optimization -- eg Our largest planets can now be size 5(!), so the value of a 3rd mining/nutrition facility could be worth it in some cases.

[sooooo]

Sorry, not much time to report this time.

My set was mainly moving our ships towards the "fake" Halis orange civ and getting our warp lanes down there too. I build construction ships, capital shipyards and battleships.

We finished environmentalism and did the triple-revolt suggested.

I built a colony ship for Capricon Hephaestus but the fake Halis got there first! The colony ship is still awaiting orders.

All the ship manoevering and warp-lane building can get rather tedious - you seem to need a lot of construction ships to be able to move your ships around at any decent pace.

[timmy827]

OK, if we're attacking Halis get started as soon as the jump lane is finished. Also, maybe we should start on Lu w/o the starbase? I'm fine with the rest of the team voting against me, but it hasn't been started yet...if we are doing this next player needs to forget the warp lane being built by the southern conship and hurry it to the starbase spot immediately.

I think our builds should be adjusted to a few less battleships and few more carriers/squadrons.

[Mojoqmeyvam]

I'll look at the save and play sometime tomorrow, likely early afternoon. Anyone have any advice/plans? Ideally, I'd like to take Aestoria this round, and leave the next person prepared for dropping Lu. Assuming that's what our plans are.

Edit: Good points, Timmy. I'll start the SB ASAP on my turn, and let extra Con ships build down that ways.