There's a lot in this post - collected thoughts from a few test games. Please do feel free to tear things apart
Exploration is hugely important. We need to know if we have neighbours, where resources are, whether we can circumnavigate pre-Optics, whether we'll need a big navy for barb defence or invasions, how much tech that navy will need behind it, getting trade routes from meeting AIs. We need (at least) an early workboat that finds at least one good expansion site and then goes off into the unknown to meet people and maybe circumnavigate.
We are
-poor, so early Slavery is probably essential - and in that context, at small-to-medium populations, 1
> 1
(except that rate of supply is limited). With a granary, 1
> 2
.
My first thoughts were to tech Fishing, Mining, BW, Sailing (in that order) while building four workboats (three to work, one to explore; the third boat gets the first whip and arrives to work the third clam perfectly as we regrow to size 3) then lighthouse (sailing research finishes just before whip-time, perfect), worker, galley, settler (about T80). The clams and land improvements (farm, mine) are each only worth +2
or +2
, but the clams are probably slightly better value in the first place (2
2
vs 3
) so it seems right to prioritise them over a worker.
I tried a worker-first idea Grifftavian suggested, but he ran out of useful work too soon for my liking. The only directly useful tasks are the farm, the road on the farm, the chop+mine on the plains hill, and the chop on the grassland. That's 28 turns worth (or so). Building four workboats plus galley (4*45 + 75
) will take quite a few turns... say 90
from 2 whips still leaves 15 turns of 5
/turn working plains mine (which won't be available immediately, and might conflict with working clams). Until the galley is out, our worker is not getting much value (and it doesn't explode in value even then).
Also, to stay healthy at size 5 (which is our pre-monument happy-cap), we need 2 forests and 2 health resources connected, or 3 resources. So there is some merit to not chopping the two grassland forests until we've got a trade going, or our second city out (hopefully with a different seafood, and then the chopping is into an Oracle or something). This gives an early worker even less to do... We can do speculative mines to try to pop gems, but if we don't work them, that'll suck.
My best Oracle time on my own maps is 95 turns: teching Fishing, Mining, BW, Sailing, Myst, Medi, PH, Writing and a bit of Alpha, building WB, WB, WB(whip), WB, lighthouse(whip), worker, galley (whip 2 pop), settler (finished 1 turn after PH), oracle (whip 4 pop). The worker put the farm up, a few turns on building the PFH mine, then galleyed across to the small island to get that mine up, then back to finish the PFH mine (now that the chop will go to the Oracle), then 1 side chop for the Oracle. Second city had its seafood worked (by the fourth WB, which had done 25 turns exploring in the meantime) and a replacement WB-explorer already whipped out. Paris was size 4 and Orleans size 1. I do have to waste 10
while teching Fishing at the start. I think there's no useful way around that. 10 turns on a worker before switching to WB gets that WB out 2 turns slower than using the first 11 turns to grow to size 2. I might try a worker/WB/worker, WB, WB, WB approach some time.
To get T95 I had to work those grassland forests quite a bit while building the Oracle. While (preparing for) building a wonder, 1
>1
because the straight whip is downvalued for a wonder. I can do a similar T96 Oracle working more of the food squares that is about 11
+
and 20
better (once both scenarios have worked T96). This is a noticeably better future springboard, but I once had one of my scenario's two Industrious AIs build an Oracle at T94. Stonehenge often comes out about T87. I do not think we have time for the Oracle to do cutesy stuff whipping small things maximizing overflow to the Oracle, because we have to spend so much time building 4 WBs, lighthouse, galley, settler before that Oracle. My Oracles are getting built less than 15 turns after the previous whip, which means there's not enough gap for playing around.
Later on in the early game, I have been noticing that the cheap +3
lighthouse is nice, but we actually can't sustain working our four food resource tiles and staying happy between whips (unless we go polywarrior-HR). We just don't have enough profitable-enough
tiles. Options include
- growing the city unhappy (accepts 2 lost per extra population; which I did while building my Oracles)
- not working (say) the grassland corn (trades 5 for about 13 on a grassland mine, for a comparable loss to the above)
- turning on avoid growth wastes at least as much as the foregoing, whether you rearrange the tiles worked or not
- arranging the timing for multi-population whips (but going below size 4 is probably always wrong except for a wonder, because we give up working a good tile), and
- building many settlers and workers
So, combining 1 and 5, to get the best value out of our geography, I think the capital will have to spend lots of time at the happy-cap (size 5 or 6 with monument) building settlers and workers, with a brief growth period (building warrior/galley/workboat) before each whip-
expires, timing things so that we minimize the turns spent one size larger (when we'll be unhappy and maybe unhealthy), and do whips of more than one population only when we have a granary.
Medium term, if we haven't skittled an early AI, we proabably want HR and either CoL specialists (e.g. at size 11 we can run 7 scientists in the capital... with one-two side cities of merchants we can run an economy that works OK while continuing to expand... better with Pyramids, or Confu Pacificism), or HR+Colossus and spam population on coast tiles for the commerce.
Way down the track, the capital can become a pretty good Globe Theatre drafting center. It has enough food for 2 of our UU musketmen every 3 turns. That synergizes with our early capital-driven REX if we settle sites with about 2 food resources that can draft regularly later. Bureaucracy looks underwhelmingly good unless we capture an AI capital or something. Need to plan for 6 theatres and Drama, though.
Assuming we are semi-isolated or distant from the AIs, we'll need a serious navy. Galleys and triremes don't cut it against the millions of caravels Renaissance-era Emperor AIs tend to have.
What to Oracle? CoL or MC seem the two candidates to me. On my first play of Grifftavian's save, I Oracled CoL but got dogpiled at T150 by four barb galleys. So on a replay I Oracled MC, got four triremes up (about what I needed to avoid losing working boats) and saw no action at all.
MC is worth more beakers, at least. A general plan of Oracle MC, whip forge in capital (and/or side city?) and hope for GE-Pyramids while teching CoL-Philo to run a mega-SE (Rep, Caste, Paci) would be a very good use of our
supply. We judge whether to put NE in the capital at the time. Meanwhile, the side cities prepare a pile of workshops (+1
Caste, and later Guilds and Chemistry near our tech paths) for crushing the world with drafted UU musketmen and Trebs or Cannons. Caste is great for conquering with a draft, because we run poly-artists until we have 10% culture, and then draft the conquered cities into oblivion (well, size 5 anyway!).
Civil Service and Philo are prereqs for Nationalism for drafting, and Gunpowder is a long way away on another track. A first thought is to do Paper-Edu-Lib(bulb Nat)-Gunpowder, however experience in my test games suggests we need to have arranged for Optics and Drama also (caravels to soak damage in the invasion fleet, and Drama to get 6 theatres for the Globe Theatre). If so, since Paper, Edu and Lib of themselves do us no good for conquering, if we've been able to (say) trade away Philo for Machinery, then now Guilds-Nat-Gunpowder is probably cheaper (and then a Lib bulb for Steel may not be out of the question) and Optics is a quick diversion beforehand (so the navy is mostly ready before Gunpowder and/or Engineering is teched). If we need Astro to get to the AIs, then Paper-Edu-Lib(Astro)-build Observatory (and maybe uni)-Nat-Gunpowder-backfill is probably best.
We also need to consider carefully whether to trade with the GLH-owning AI. If they're a decent AI, then if we trade techs with them we may make a monster that we can't tech down in time to conquer along with the rest of the world. If we trade with them, they have to be first target, I think.