Explosive expansion start for the AI - Suggestions appreciated.

I start out spending the initial bonus on research to give TL2 and the rest on TL1 Intel given that I start with the low bonus which I normally do. Given more research points than this I would go for sensors Ion Engines Small hulls and Cargo as they all boost the initial expansion. I would add any racial specialities to this and Physics as shields are indispensable.

Once that was spent I would go for 100% spending on research until I had around TL5 and would concentrate on expansion and colonisation to build more research outposts. Once I had TL 5 in research I tend to diversify as by that time I will have sufficient to do so but I never drop my spending on researching research below 50% until I was at TL20 as researching research creates a virtuous circle if you upgrade your facilities accordingly.

The exception to this are if I make first contact when I turn to researching Intel, Small hulls, and Basic weapons beam and pulse weapons if I have physics cannon and missiles if I don’t I also research shields and armour whilst trying to maintain some level of research on research at least 50%. If I have to go to war I would add Point defences to this but not otherwise. Whilst this happens I would continue to expand away from the first contact if possible to give me a bigger base. Diplomatically I would seek or accept any non aggression treaty I could get being aware of Knobling clauses (No research no intel)

Hmmm I never really tend to split my research until mid game (apart from the 1st turn points). I tend to do 100% x for a turn or 2 then maybe 1-10% (depending if I can get another level for that) and the rest goes to tech Y for a couple of turns.

Tech research I tend to do the same in blocks, every 10 or so turns I do a couple of turns pure tech. This is also dependant on what type of colonies I am founding if I find a breathable large or huge the 1st thing I build is a space yard and resupply, as I am building these I try and cram in as many levels of research tech as I can so when they are ready to build I have as higher level of RC I can get then leave them building and upgrade the rest and concentrate on other techs for a while.

One thing I forgot to mention is that I tend to try and get a couple of levels of intel in the 1st year or 2 of the game and build intel posts on a huge non breathable planet, and often fill up smaller planets with them. These get left on defense even when I meet other races.

As for Organics and Refining I tend to find these to be more of an issue than minerals, constructing colony ships is quite heavy on the use of these, and if I am not careful and plan ahead with them I always seem to hit that speed bump just when I meet a hostile AI race. (sods law works even in SEV) Also most of the Racial specail abilites are heavy on one or the other of these. Maybe not 1st phase but certainly before the 3rd year the AI needs to be thinking about these (or building more gathering places, I tend to try to minimize these gathering planets to focus on research).

Can you affect what types of colonies the AI builds and how often, such as if my maintence requirements for minerals is less than 33% build research or intel, if it is more than 33% of my income look for good planets to set up resource gathering, if it rises to more than 50% make this a priority.

Picking up on what Evilginger said is it possible to disable the no research and no intel options as far as the AI is concerned, at least until the AI is fixed enough to be able to handle them reasonably. Checking the AI races in my current game I found 3 of these (1 no research both sides, and 2 no intel 1 was both and the other was just 1 race doing no intel) that was out of 10 remaining races (including myself). These agreements are suicide for the AI and they seem to offer them around like candy.

I know that the AI uses the weapons tech from the racial features, but I am not sure if they use the facility bonuses (never checked) as a rule, I have occasionaly coquored planets that have used Temporal Spaceyards, but I can't seem remember seeing any of the other racial facilities. (Which is a shame as they are very powerfull and do cost quite a lot of bonus points and research points to get, in the long run I would try and make the AI use them, but if thats going to be difficult then lets concentrate on getting a good start going).

I think I understand what you have done with the planet resources (but ain't sure) You have setup a priority for minerals if a planet has above 80%, unless they already have more than a certain number of colonies producing minerals. You have done something similar for organics and radioactives, but I am not sure when they start building those.

Am I right in you basing the building of new resource gathering colonies on the amount of colonies already doing that?

If so is it pure based on the number of colonies or on the number of facilites collecting the resource? I would heavily suggest number of facilities not colonies, or better yet I would go try basing it off a % related to current maintenence, as that only ever goes up (well until you start to get hammered) say you want to be producing at the very minimum 50% above your maintence requirements, and idealy 100 - 150% more. This means you can always keep your yards producing and fleet flying. In my experience when (rarely if) you hit problems it causes serious problems.

One other thing I think is mandatory is any large or huge breathable planet (i also tend to do this on mediums as well if I don't have any bigger in the system as this is a nice back up to my military bases) should have a space yard built 1st, otherwise they take years (literaly) to build up their facilites and years again to do upgrades (upgrades imho need to be done as quickly as possible to really make them pay off), a space yard reduces that time drasticaly and allows for extra ship building when needed.

I haven’t yet played any games in god mode and had a check round on what the AI's are up to so I know that they use the weapons from the racial techs and have seen organic armour used but as most of the time I am 10+ Tl's ahead of the AI in weapons techs I don’t have much of a problem dealing with them. I would presume that the AI would use the weapons and armour as they are the most straight forward applications of the technology.

I will have to start a game with custom Aliens which use only the racial techs and see what happens and have a snoop round in the background.

Another AI diplomacy change I would recomend is to change it so that AI will be very unlikely to enter into treaties that stop them making treaties with other races, as this tends to push them into hostile realtions and war fairly quickly.

1)
set homeworlds to fastbuild

2)
build a spaceyard base or two

3)
build colony ships with sensors on them to use as coloniser scouts.

4)
use these colony ships to scout as far out as possable till they run low on suplies, then colonise the nearest good planet.

5)
build colony ships with cargo instead of sensors and supplys to colonise those planets that your colony ships have found.

6)
build frigates as needed once you meet annother race.

-----
an se5a is a ww1 fighter, it is also a car.

1) Personally I never use that simply because I forget to turn it off and 10 turns at 25% build is really harsh.

4) Different approach will have to try it, I usualy grab any breathable planets as soon as I find em

5) I find that using minimum cargo (population) works best for colonies, as migration seems to do a better job of spreading population than taking more from your homeworld. Although I do add cargo to my colony ships, thats for transporting troops to conquor hostile AI races.

6) I like to get a frigate or 2 out fairly early on before I meet any race as it gives me the ability to respond imediatly (I upgrade them often as well), I sometimes also use them to control warp points and secure choke points for myself.

Would never use fast build unless it was an genuine emergancy

se5a essentially described the SE:IV rapid expansion model and it was (is) very effective. Of course for stock SE:V you'll need to have the BSYs or it would be a very bad idea to go on emergency build.

-----

Space Empires Depot | SE:V Balance Mod

CodeListing_Interface.txt has all the functions in the main code that you can call from your AI script. Sys_AI_Empire_Stats_Get_Total_Maintenance_Resources_Expenses() and Sys_AI_Empire_Stats_Get_Colony_Resource_Production() are likely things to use to implement cordas' scheme. These two functions are documented to return structures of the "resources" data type.

CodeListing_BuiltIn.txt has utility functions for dealing with data structions. You'll probably need resources_get() to pull out the individual numbers for minerals, organics, and radioactives from the structures returned by the above function calls.

The first thing i reasearch is Sensors up to level 5. The reason is that this way it takes me just 4-5 turns to explore a solar system.

I like to research applied research first and build lots of research colonies... this way, even though I'm pretty much weaker early in the game, if I can just hold back the AI for a while (which usually isn't that hard) I end up being able to research anything I want real fast later on.

I would argue that AI ships (even armed and armoured/shielded) need to learn to retreat if they're up against clearly unfavourable odds. Especially early game where each ship loss is felt more - the AI should realise 1 Frigate doesn't stand a chance against a group of 7 say, assuming roughly equivalent techs.

Yep, especially when seven AI designed destroyers won’t usually stand a chance against a single player designed frigate already

-------------
Feel lucky earthling ?

I argee the AI has to know when ot retreat. Also, the AI needs to know when to resupply. Several times I've run across huge AI fleets that have to retreat because they are out of supply.

As far as research, I start out by pumping my research capability up several levels then sensors. After these are up I work on several things at one time, engines, hull size, intel ect.. But one other thing I do early on is learn to colonize another planet type to give me more planets to colonize early in the expansion phase.

Its faster to build new research posts than to research applied research in early game, personally I would ignore applied research until you have at least 40K research, I suggest sensors, ship size, armor, missile, projectile, small weapons, point defence, one level of engines in first turn but no more engine thereafter because there is a bug with engine upgrade for players which might affect AI, (I dont know). Set ships to use missile with maximum range strategy. Then this AI should dominate all other AIs if not players too unless auto resolve is different from tactical resolve.

Have them build at least ten troops on each planet and minimum 3 platforms with two missiles per platform.

Have them use any planet with less than 100% mineral for research 90% and intelligence 10%. When you have 40k research, then research 2 levels of applied research and upgrade after you have second level done.

Then research phyics and get meson blaster up to 7 or better and at least level one large ship,base,platform mounts and then get destroyer with optimal range strat and meson blaster. Then research shield. Build plats with two large meson blasters, one missile and PD.

etc etc

The AI is not very effective at the moment. Where can I learn to mod it ?

I agree with obliwobly about the just build more research facilities and leave researching applied research for when you have a larger amount of research to spend otherwise it can bottleneck you bad.

Also something that is HUGELY overlooked... Switch to monolith once your base economy takes off... scrap all others and use monolith (slowly upgrade that is hehe.) because at the highest level a monolith gives 12,900 of EACH vs 10,000 of ONE... then build and USE the material converter (or whatever it's called) the AI should be able to easily calculate and use that device to great effect in combination with monoliths.

YES monoliths are maint heavy but well worth it if you also use value improvement plants (or nature shrines hehe)

also the use of robotoid plants and such give huge boosts... using the organic tech facilities is very useful. (even though they state different all 3 will work together cumulatively or so it seems to me) or at least the effects they have that vary work in any case it's well worth having 1 of each in a system since they are systemwide.

fate/nature/time shrines are absolutely awesome as well... nature especially since it caps out early with 22% improvement while the value improvement plant is only 7.5% and local

The AI doesn't make enough use of systemwide facilities IMO.

also a research policy that puts like 1-5% in a critical weapon type and 1-5% in sensors and leaves them there til they are capped (or you get to where you want to max em fast etc.) allows you to continually improve them while doing other things because those are CHEAP techs they don't need 100% except for the initial push just to make them available and useful.

the huge key the AI lacks is getting it's research off the ground... it easily gets more resources than it can ever effectively use but it never gets it's research to any really useful level...

Lategame I find shields and armor worthless... Range counts for a LOT more at that stage because most of the high tech weapons can blow through them too easily... HOWEVER this might be changed in 1.17 since I am still playing through my first game in this patch and I am finding some good balance fixes such as electrical discharge caps at half what it used too... Still a single heavy fighter can have 12 guns that do 592 damage vs the 1730 shield points of a maxed phased shield lol. anyway... fighter weapons are as potent as normal ship weapons (sans mount) at max tech it seems but just very short ranged.

Anyway... Hope some of this might prove useful.

I disagree the best way to get the AI of to a good start is to research applied research whilst building and upgrading as many Research centres as possible. I would find an AI who just built lots of tl1 centres a walk over as all I have to do is find the one colony with a star port on it and glass that to cut off all that research from the empire, which would rapidly reduce the AI Empire to only those few Research centres in the home system crippling it. Such a tactic would also cause the AI to expand faster than it could protect its holdings again not a wise thing to do.

Hadn't thought about the vunerability of space ports, as I habbitualy take natural merchant in my racial traits (its just too handy).

The AI should be instructed to build - recylce - build weapons platforms on its space port worlds, and also build sat and mine defences for these worlds. Yes stationing ships will help as well, but realisticly ships are probably better used elsewhere rather than being parked on these planets.

Of course the AI needs to have a formula to upgrade facs and as far as I am aware all mod builders know this and are working to implement it (or need repeated introductions to Mr Baseball bat until they do).

The debate about when to research Applied research is going to go on and on, so my suggestion is that some AI races should be told to do it on a constant basis from the begining of the game, and others to do it in lumps. Not every AI should have to approach this in the same manner, as long as they consider the implications of this.

The AI could also do with being given better instructions on how to spend its 1st turn RP. If it wants to get a theoritcal science at the start of the game its probably best to do it now in 1 turn rather than doing XX% of it and then the rest over the next Y turns. I would make it spend at least 20% on sensors.

However it’s spent it needs to be spent in such a way as the AI gains something from it immediately. I will generally allocate it all to one area then reduce it to the minimum needed to gain a TL in a turn then do the same with the next most desirable tech and so on that way you get something out of that initial boost which can be used straight away.

Is it possible to put a halt at warp gate comand in, then wait till the end of the turn and jump all ships waiting to go through the warp gate to jump together?

I know using the fucntions to check idle ships its possible to do once per sector which would make the search to see if all ships have moved as far as they can easier to do.

If you could do this it would help a little, I have seen the AI send upto 4 or 5 ships through a gate one after another, when individualy they get killed (with barely denting my organic armour, but together would at least stand a fighting chance or do me some real damage. Add that onto a decent ship build and it would become a reasonable fight.

P.S. Go enjoy your beer, and have one on me

good work I look forward to fighting this AI

Mmm... I wonder what this AI will do if I turn on 100% Vision such that the AI is guaranteed to know where to attack and where to avoid. Congratulations on your progress!

Incidentally, this AI is tuned for stock right? Ah well, Balance Mod is going to re-vamp a few things anyway so I'll wait until that's settled and this AI gets a BM version.

Uuu... I always make my own tweaks to any mod I play, so I guess I'll have to learn up AI tweaking too to make sure the AI keeps up with my changes. Looking forward to reading your readme.txt