After recent talk in Thinker mod thread I was seeking a challenge, because I found AI is not very strong on Transcend:
topic: AI can't fight even when they are "on paper" strongest faction and initiate war
I got feedback from mod creator that my game settings are not the most suitable for strongest AI:
> topic: AI can't fight even when they are "on paper" strongest faction and initiate war
..
Dry maps are generally worse for AI performance too. If you want more challenge, try also using blind research, otherwise with the default tech tree it might be too easy to rush some techs.
..
Generally the AIs might not have the time to build up during the first 80 turns if the game settings are unfavourable for them. You can try changing conquer_priority=150 or something so the AI will build more units. Sane values for that setting are somewhere in the range 50..250.
..
The AI already has some priorities which cause it to increase combat unit production if it loses bases, but it's probably not going to help much on smallest map sizes.
so in this new game it is:
standard map
no landmarks
weak/rare/dense
168 years for AI to grow
a few minor thinker.ini changes in linked post above, but i left default conquer_priority
I don't generally like standard maps because they are really "huge maps" - you will operate at around 100-150 bases which is insane grind; but I wanted a challenge so I decided I will observe AI (observer mode), don't play and just take over one of the AIs after some time; my original plan was to play just 10 turns to see how far I can get playing a faction that is much weaker than top1; I thought a performance metric of "+5 bases conquered per turn for 5 years" would be sufficient to decide I am winning and AI will not have chance, since they already play their "best strats" and can't adapt; in such game each turn can take 1-2h so I thought if I can't get this performance metric in 10 years I will just pass, cause I dont like to grind too much
I observed a few games (mostly keep enter pressed so you can skip quickly to particular year) and decided this game would be a good candidate because Hive AI turned out to be really strong: around 2220 they pretty much waged war against everyone else simultaneously, they conquered univ and angels and in 2268 they were quickly progressing vs remaining opponents morgan,gaia, lal (sparta was at just few bases at this point)
meanwhile on south morgan was fighting sparta for like ~100-130 years which stifled his growth a lot, but he did manage to win leaving sparta with just a few bases; this was with the help of lal; morgan and lal were allies most/all game IIRC; note sparta was boosted: +2 effic; start with 60EC; starting techs: doc:mobility and B1:industrial base - and they still lost to morgan (and lal); also i applied nerf for univ for this game: -1ind
in 2268 the situation looks like this:
bases/pop/mins/energy:
morgan:
52/262/728/1060
hive:
164/850/1893/1333
lal:
60/324/764/883
note if u dont know: thinker mod is exposing these stats so you dont need to count it
as you see hive compared to morgan is clearly >3x in bases and pop while 2,6x in minerals + they have cloning vats and a ton of other good SPs; additionally since they are already in full-war-mode they have a huge army and roughly ~80% of their bases keeps producing army units!! u can check urself attached save
hive main army units:
14-1-x:
22 shard copter
24 shard penetrator
55 shard tanks
5 shard squad 13/3r
11 shard artillery
45+ various probes
3 nukes + 4 in prod
morgan:
6 shard copter
4 shard needlejet
3 shard interceptor
13 drop shard squad 14/<5>/1
9 various probes
this is pretty ridiculous army for hive; also consider most map is covered with mag tubes ! so its perfect scenario for rapid blitzkrieg war
note that for the last few years until 2268 I set diplomacy settings to permanent pact between gaia+morgan+lal to make them stronger; I was actually initially planning to keep it this way when I start playing because I thought I likely have no chance anyway vs such huge hive faction and it would be cooler to play this kind of "world war" with just 2 sides; but after all I dropped this "permanent pact" setting in 2268 already to make it more realistic for me
on paper morgan is top3 but I pick him cause I like morgan faction more than lal and also he has better potential - better SPs already built
first year - 2268
so I started at the end of the turn after all moves by morgan by auditing his bases and adjusting production:
1) AI invests in expensive but low-efficiency units like eg
144 cost units: 1-8-3 probes or 14-<5>-1 drop squads
they dont make smart cheap units like 1-1-1drops or 4-1-10anti-air
or just 14-1-1drops
2) AI doesnt build any satellites - this is probably top1 biggest AI mistake in late game; hive has 29 Aerospace complex, morgan has cloudbase; in the other game i saw AI was making satellites in same version of thinker 3.0
3) AI has ton of various formers, eg morgan ~60, but no clean ones ?
generally AI use no clean units?
I decided on main strategic directions:
a) mass air units needler/copter + some drop-pods - especially 1-1-1 which u can drop + upgrade to defender the same turn if u need to
b) keep peace with everyone except hive
c) mass satellites
d) drop all research, 90% slider to eco - in total war game scenario like this it makes no sense to make any research cause it's too easy to steal tech; so research does not give any advantage (besides every faction is already at 70%+ tech researched); it's better to let enemy research for you (hive had huge tech output) and steal; I actually made mistake of keeping 10% to tech for some reason
since I didn't have perfectly consolidated territory (bases of lal and sparta within my territory/landmass) I had to move through lal territory to get land units to hive; that's why I prioritized connecting mag tubes to lal and hive and here is where the problems started
...
in years 2270 lal made "break pact and sudden attack" and killed a few of my units; idk why but i suspect he didnt like terraforming on his territory ? does it work like that it angers AIs regardless of what changes you make (whether positive or negative) ? cause by that time i probably constructed like 4-5 tubes
so the same year I fight back and immediately take over one city which was closest to me on south and had no defence; immediately he offers cease fire , i reject, and then he offer 728e to keep peace ! i took it and IIRC we make treaty or even pact
around this time my next strategic moves were the following:
e) wait till hive research super tensil solids, steal it and insta build space elevator
f) since now i can drop anywhere which is kinda ultimate war tech I'll focus all drops to steal base with cloning vats, then switch to power/thought control for +4 morale with no disadvantage (CV removes them) and then drop everywhere around hive territory focusing on bases with weak defense
around 2271-2 I recalled how sparta was fighting morgan all game, while now I see they have 2 bases near me with no defense, so I just broke pact with them and took them over
quickly sparta offered to sign servitude pact with me and I also extracted 1 more base from them through blackmail; BUT i forgot they were in pact with lal, who intervene in their defense
so now I have a few more turns fighting lal and hive which was dangerous ! I took over few bases from lal but left one of my main core bases without any def, so lal took over it but then he immediately offer peace and treaty and offer to give back this base as an incentive, so we sign it - i still wanted to stick to my original plan to fighting only vs hive (though multiple fronts vs hive) as i was afraid fighting vs lal+hive would be too much; ok, we have treaty with lal, but he doesnt want pact any more and i see he is creeping his units on my territory, 1-2 years later I demand withdrawal and he pronounce vendetta, well that make sense
so now I lost any trust in him and decided im gonna fight him for the rest of game!! what helped me was that lal was stupid enough to also declare vendetta on gaia - so gaia was suffering the most vs hive on north and lal on south - gaia lost a few huge cities to lal; idk but i feel like AI is dropping their character in late game and just waging total wars ? i mean i'd expect lal to keep peace with gaia, but maybe they didnt like gaia running fundy ?
meanwhile vs hive and overall my situation is pretty good as I leveraged their war vs gaia and took some cities in this area:
year/bases/pop/mins/energy:
2273/91/?/1600/1700 hive: 146/805/1618/1154
so after 5 years im already +39 bases, which is much better than my plan of +5/turn!!
my focus is to stir maximum damage to hive so I use another strat:
g) take over many bases in given area even if it means leaving some bases with no defense units; when conquering base immediately sell some facility
so this strat means at beginning of each round I'm losing like 5-10 bases per turn (besides losing ~20 battles
), but it also leads to more bases being conquered so overall +net base count surplus is bigger
additionally I made a few sea bases, but likely made mistake of not building any drop colony pods
after 5 years of this game i felt pretty confident i can't lose, my e,f) strategic moves were successfully executed in 2275 and additionally since i fought vs lal i also prioritized taking over his base with cyborg factory; this meant im now at +5morale and making insta elite units
I focused army composition mostly on drops:
14-1-1 drops
1-1-1 drop clean
I also made some 6-1-3 but they were mostly not not useful; likely would be better to make some 1-1-3 + more 14-1-1 ans 6-1-3 is not good for fighting anyway so 1-1-3 would only be good for taking over bases
and:
14-1-1 soporiphic gas - which gives -2morale AI; this is best unit vs lal especially since im connected with tubes with him
do you think I could make any better units at this stage ?
since i put 90% slider to eco I'm at ~+2k e/turn which I use all to insta build units everywhere, especially in cities exposed to enemy (but likely save for the next turn) + insta build satellites
I kept some tracking of my progress, so here is the overall stats:
year/bases/pop/mins/energy/ +base count:
2268/52/262/728/1060
2271/74/338/1252/1349 +22
2272/86 +12
2273/91/./1600/1700 +5 hive: 146/805/1618/1154
2274/111/408/1814/2026 +20
2275/112/475/1811/1780 +1
2276/123/518/2041/1895 +11
2277/142/602/2332/2178 +19
2278/182/717/2986/2853 +40
some more thought on late game:
4) AI makes planet busters and just parks them in the city ? makes no sense; looks like an epic waste on AI side
5) AI sometimes leaves bases with like only 6+ air units eg 14-1-14, no real defenders like 1-8-1 so its easy to kill all units in such base with single copter
6) hive had a ton of probes but didnt steal will to power tech ? strange; they did use probes a lot to buy bases
7) hive AI didnt utilize their army really well, eg:
they had 55 shard tanks in 2268 but a ton was just parked as police - that is huge waste; with this kind of army they likely could blitzkrieg gaia in a few years ! why not just leave only proper 1-1-1police in bases? same they parked some air units in bases on north that were completely save
8 ) its really good to purposely make eco damage as with +4 planet single random shard copter (which i build everywhere anyway) will kill the whole chiron army easily and u rake in big bucks + broken squares doesn't even matter as fungus give 2-1-1 - i guess i should have utilized this more; am i right that fungus pop never kill boreholes ?
2278:
attached:
AI observer standard map hive with cloning vats - last AI save
morgan vs unsurpassed hive with cloning vats, 2269 - my 1st unit movement turn after already making changes to city production at the end of prev year
morgan vs unsurpassed hive with cloning vats, 2278 - my last turn
New SMAC quizzes available.
