[pb40 spoiler] Götterdämmerung: Destruction Is The Masses' Entertainment

[Coeurva]

Re: Oracle micro. I've worked out a plan that can indeed get it by eot70 (but will have to verify in the sandbox) as long as we can get Med eoT59, Priesthood eoT62 (this might be a problem... maybe, I have an idea if it's eoT63 as well). Settler at Canoel also finishes eoT59 which is when Oracle production effectively begins.

First problem is that Canoel's stable configuration for 17h/t (phsheep, plains irrigated wheat (+6f surplus), 3 pmines) wastes the clams tile during the slowbuild turns, which is bad management. Instead, let's aim for 15h/t passing the wheat back to Rome (while building, most likely, the X6 settler there... that's a pity, but we're also likely in a race for the gold site). That means working phsheep, clams (+5f surplus), 2 pmines, 1 gmine stagnating. Two of these improvements don't exist yet, and we have a few other tasks to juggle with them.

The worker schedule is tight. We need the entire tilekit by T64, which is when Canoel will have hit size5; thus we must have one worker build the second plains mine T60-64 (incl. movement), and the worker pair assigned to X6 to go pasture the grass cow T60-61, move 66 onto grass hill T62, mine T63-64. Don't road the cow; there's no time for that.

Concurrently, we also need to hit size5 from an empty food box within these same 4t. That means we can't work the plains mine throughout size4, because that's only a 6f surplus; instead, to avoid wasting 2t working a bare grassland, we'll make a cottage on the one shared riverland between Canoel and Rungholt. At present, we have two workers next to this tile, but one of them has to leave immediately for X6 and can't be used. Instead, we'll use one worker to cottage T58 while moving one of the ivory campers 69, cottage that, cancel. T59 he moves 78 across roads and cottages the destination tile paired with the other. T60 we move one of these two onto the bare plains hill 44, to make the second plains mine by T64, and finish the cottage with the other.

Then we follow this plan. There is a 12/35h spear in the queue (was a warrior before trade connection). City centre is 2h.

T60 (produce whatever, just grab all the food + cottage) 8/28f
T61 spear 17/35 16/28f
T62 spear 26/35 22/28f (work plains mine again, pass cottage to capital)
T63 Oracle 9/150 28/28f
T64 Oracle 24/150 (swap to the 15hpt stagnating configuration)
T65 Oracle 39/150
T66 Oracle 54/150
T67 Oracle 69/150
T68 Oracle 104/150 (use the 20h chop at some point before this, should be T65 but doesn't actually matter.)
T69 whip spear, work 12hpt (size4, sheep clams pmine pmine) 68/35 => 33h overflow
T70 Oracle 13hpt (size4, work clams + grass mine + both pmines, sheep is worked by Ravenna, wheat by Rungholt, last turn has food stored so -1f doesn't starve a citizen) 150/150

T60 shouldn't be the spear since that'll waste overflowing hammers, but if I've got the invested hammers into the spear wrong somehow, it's good to have that cushion.

I'll explain the worker management in screenshots next turn if it's not clear yet (it probably isn't).

[OT4E]

Are we decided with taking Maths with Oracle. There is always an option to take Alpha and then get Maths quick with bulb builds. On the poor maps like this it can be better choice. Alpha is more expensive by itself and can boost economics further on.

The question is are we ready for mass-choppings right on the turn we get Maths and what is more important what do we gain from those chops? Or may be we take a bit time and spend those on praetorians?

[Coeurva]

Maths and Alpha cost the same in RtR iirc. But I'm not quite decided yet.

The main reason why I want to take Maths is that it leads to Calendar while Alpha doesn't, and Calendar looks like the key tech to have on this map, aside from Currency. We'll gain at least two 5-food sites (require IW as well), some 12 commerce/turn and 3 happy resources (!) from Calendar.

Our many mines make an argument to build research, but I tried that kind of hammer economy building direct wealth in PB37, and completely crashed for two reasons: (1) mines don't upgrade as well as cottages and consume more food; (2) at some point the need for units and infrastructure, which you've always put off for short-term advantages, will overwhelm your available build queues. Simply put, the lesson I drew from that was that in evaluating ROI on buildings, I had underestimated the cost of a build queue by itself (rather than any particular build in it). In single-player where you can easily conquer AI land in Classical without more than a dozen units, you don't notice this because you're never bottlenecked on queues, whereas immediate maintenance costs on Deity are so high that wealth / research builds are sometimes the only way to climb out of the hole after a necessary early landgrab, even on a small number of cities. But this is different. If we conquer someone during the Classical era in this game, we're competing for the top spot even if we're building aqueducts.

Open Borders don't help us here, because if the Oracle plan succeeds, we should be first by far to pop third-ring culture at a port (nobody else will have that kind of cultural output until they build Stonehenge) and thus open up the "Astro lands" which -- I'm assuming -- are the only source of ICTR on the map (which would be unusual), so we'd just donate foreign ICTR to any trading partner, without getting any in return.

Imo we have enough research power and manageable maintenance up to Currency that I think we don't need to build research. We can't mass-chop any time soon, but barracks / granary tend to come chop-assisted everywhere in any case, and free +10h on both are appreciated, even if Maths only comes into play for 2-3 cities during the time we would only have Alphabet instead.

Yeah, I want two forests per city left if possible, for 60h into forges, and then the rest forge-boosted into 1t praets per forest -- think that's the best way to make our early mining pay off. Just 1-whip the forge at 90h and repay it in less than 10t just by working 2 mines; then we'll hopefully have the strongest production base in the world.

With religion, ivory and gold, the happiness from Calendar might not even be so important; we could take MC before it. (MC with Oracle is right out because it's not on the road to Currency, which we simply need for the ICTR pseudo-Astro plan.)

The hammers that we'd invest into research builds could instead make a galley and two settlers for the "backline" islands I'm hoping for, and then it's all aboard the Currency +4c per city train.

E: In fact, by building the Oracle for Maths we're essentially doing a huge Alpha -> Maths research build with increased efficiency (and no need to spend beakers on Alpha when we don't want it).

[OT4E]

The idea was not to start building research till tanks but to invest minimum efforts and get borh Maths and Alpha by the time we were planning to get just Maths. You said we had +20 gpt on 0%, thats low and I imagine we can add 30 hammers at some short period while grassland cottage gives just +1 and it is absolutely impossible to rise research output that high at this point.

Alpha also gives discount for Currency and can fill our trade routes before we can plant those colonies. Then we will have option to close borders. Btw are you sure those islands allow to land on and settle? I know I'd better enter and look myself...

Another thought is to make Pyramids. May be Oracle, HG and Mids are too much of wonderwhoring for Rome but still worth considering. What do you think?

[Coeurva]

Mainly I'm concerned about the long-term effects on our growth curve. It's true that a cottage just improves the tile by +1c initially and the mine by +3h, but what must be compared are the actual tile yields for a citizen. A mine consumes 1 or 2 food and usually gives no commerce, so the river cottage actually comes out at +1 or +2fpt and probably +2cpt over the mine (1cpt if the mine is riverside, which most of ours won't be).

If the breakeven rate is say 50%, you gain 2gpt from a cottage then 2bpt next turn, repeat. If you work the mine for research at the same breakeven rate, you get 4bpt, but must still pay the gpt from elsewhere. As long as you can do that, the mine outperforms the cottage until (1) the cottage growth effect kicks in and (2) the accumulated food from the cottage produces another citizen, which takes forever in isolation, but adds up when comparing 3 cottages to 3 research-building plains mines instead.

The faster we go to Maths via hammers now, the more we'll pay later with less food and cottage growth. (This seems to be the trade-off for working mines in general, regardless of what they build, unless it's granaries, wboats, or fh units.) The catch is that Currency and Alpha can cushion this loss with their commerce boni that don't cost citizens.

The balancing act will be to get granaries/workers/settlers out while building research at no less than 3 cities per turn, though we can rotate duties. I was doubtful that we can do that (there's a difference between a 30hpt city building research and three 10hpt cities building research) but now I think I agree with your judgment and Alpha might be the better choice because Currency is the big enabler to fill out the land. So -- unless circumstances change somehow I'm anticipating Alpha for the tech.

The Pyramids are very useful (incredible insight from me here). Three main issues, two with counter-arguments:

(1) We'd have to settle the stone before the gold, which means we might lose that latter site. (The only other city fit to produce a settler by T70 without growth issues will be building the Oracle.) Counter-argument: the Pyramids can more than make up for that city, both in economy and happiness.
(2) Pyramids with stone cost 250h and three Maths forest chops. If we assume that we can put 120b into Maths from natural research, building Maths costs... 250h. I don't think we take Alphabet on this path. Counter-argument: A Rep specialist outperforms a mine building research / wealth even considering the cost of his enabling building. No reason to tech Alpha with Mids, but obviously we have to decide before finishing the Oracle.
(3) They're not going to let us build both the Oracle and Pyramids, are they?

I'd rather build them than either of research or the Gardens.

Re colonies: if Commodore planted coastal tiles in the fog (which we can confirm from ocean tile yields) that a coastal city can see at its second ring, but no actually usable land to go with that coast, then we get to complain about getting trolled by the mapmaker for the rest of the game. It's a win/win. -- Really, even if the land is all ice, ICTR probably would make the city pay off by itself (maybe plant it as X10 or whatever in that case).

[Coeurva]

T58 report --



Changed the tiles at Canoel to get the fifth citizen at the capital sooner by taking the wheat for a turn. X5 settler isn't actually delayed by this, and in fact we gain time to exactly produce the capital's archer before starting the X6 settler T60. Also shown: Oracle-relevant worker micro described clumsily above. Chop-whipping the granary at Venusberg has more value than a third cottage for Rungholt, thus the plan for the worker on the ivory (grass mine at V would be even more marginal until size4, which won't happen before the floodplains unlock anyhow). Half our six workers are covered by military units in that screenshot.



Through sheer magic I've managed to discern that the #2 GNP is somewhere in between 34 and 36. I'd post our contacts' graphs but the gist of it is that they're below ours, except superdeath's power.

[Coeurva]

I've directed our military units accordingly. We can have an archer and two Shock warriors (or three warriors) in Nobiskrug before that quechua can attack it. Will have to sentry the space between us, I think. If he moves adjacent to our borders I'll declare war, but no earlier than that -- no need to appear desperate here. Canoel as expected became the holy city (most populous non-capital city), which is fine by me (it's not placed to invite aggression against it).

Magic Science has three cities (plus one razed), a peace treaty with superdeath who declared on him twice (makes me wary that Zulu could come at us next, I'll probably move the 2nd archer east instead next turn), no trade connection to us, and no contact with Charriu; put 4 EP/t on him, expecting him to do the same to us.


Charriu is gaining a lot of foodhammers lately... these demos are decent, but not dominant.

It looks like we're not going to reach Priesthood by eot62, which I think will delay the Oracle by 1t unless we can grab 8h from thin air somehow. I'll try. I think it's possible by not working the clams until after T70 :/ -- do you think we can afford that? Lost commerce will be paid back at least...

[OT4E]

My inner voice says that 1 turn delay is more important than lost food or commerce.

[Coeurva]

T61 short report:

Magic Science has moved the quechua directly south, and I've retreated our warrior back into the fog (for now). Might want to explore south-westward with the non-Shock warrior later. superdeath has dumped an absurd amount of EP into MSci and recently took all EP off us to invest them into Inca. This is too obvious, makes me think we're the real target, and I've inserted archer #3 into the queue at our eastern "front city" Munsalvaesche; can always whip him if needed, costs us 3f in granary which isn't important in the grand scheme.

We need to decide whether to pursue the Pyramids with the stone city, or a landgrab with the gold city, on the start of T64, so as to realign our workforce for roading etc. Currently, I think it's better not to build them and go for the Gardens later (at 10+ cities) instead: Rep specialists compete with mines for food and outclass them, which isn't abundant here. You could view the Pyramids as upgrading research/wealth-building mines into better "tiles" lol, but is 3.5bpt (library factored in) per mine really worth ~350h? If we would otherwise have 6 4h mines building research per turn (which seems like a rather high estimate at this point in the game), that'll take roughly 30t to turn a profit, disregarding GPP (and Rep happiness) as we could always grab a library at a regular city for the same cost, if not beaker return. And that even fails to include the hammer cost of the libraries (90b lost per city).

So the Pyramids reward speccing a single city for intensive mining imo, instead of adding 2-3 mines at a lot of them, but it looks like we'll be doing the latter, which seems to be an effect that works better with the Hanging Gardens.

What do you think of building a monastery at Canoel (takes 4t) after Oracle (+ granary)? Monk + chop Agg barracks makes the first ring take 7t instead of 10t; diminishing returns, but the earlier we build these, the less they compete for better builds (forges, praetorians) later on. Canoel could instead carry a lot of the research building to Maths/Currency, though.

Priesthood eot62 might yet be possible. Maybe. It'll require some absurd tile micro but if it speeds up the Oracle by that crucial turn, all is fair. We're just a few beakers short regularly (8 or so).

[OT4E]

Pyramids are only useful for Police State. I know even with stone they cost like 8 praetorians but to evaluate the effect you must consider that right now we are not building praetorians and not ready to use them. As for potential it is great civic and awesome for Rome.

Forget about Representation, it is not for this map.

But gold city is more important however I wouldnt drop an entire idea with Pyramids for it. Even failgold doesnt seem to be absolute waste here. May be we can reach Currency a couple of turns faster then.

I dont think we have time for monastery, I'd build extra worker or some military. Somebody has to guard all those wonders, yeah?