[SPOILERS] swance bitten, twice shy

[williams482]

Is the displayed income the amount they will actually make this turn, or the amount they would make at 0% science? If the former, that seems fairly extraordinary.

[ljubljana]

the former! pretty cool huh.....

it's not that extraordinary if they have every city on wealth we know they have something like 1600 MFG, so i mean, they could get 1600 GPT just from that if they really wanted to. not too hard to imagine they have nothing better to do with the hammers, at this point

but what do WE do about it? build wealth to get to railroad instead of our own, actually relevant cannons?

btw spotted this in mjmd's territory



HINDU missionary? that's not their state religion.... are they spreading the others for a late culture try?

[ljubljana]

btw, for the curious, here is my basic plan for how we can take down a tank stack efficiently with horribly outdated units, thanks to the magic of AGG:

step 1: for every 6 tanks, 3 barrage 3 cannons inflicting the max of 6 hits each takes the tanks down to the following health



which is almost fully-collateraled (16.8) - we can get the rest of the way by leading with, like, one cat as well.

step 2: attack with a flanking 2 cav, which has 2/3 odds to win or retreat



after retreating, modal outcome puts the tank at 36 HP, which should be easy prey for a combat 2 + ambush tokugawa rifle. or, um..... alternatively



apparently we have pretty good odds to win with such a rifle straight-up? but i haven't been focusing on rifle production so we will not have many of these, sadly. i'd like to rectify that but sadly most of the cities close to the front are stuck on cannon duty for the forseeable future so we can hit that magic number of 1 cannon for every 2 ginger tanks

so to count it up, each 180 hammer tank takes out roughly
1/6 of a cat = 9 hammers
1/2 of a cannon = 50 hammers
1/3 of a rifle/cav = let's say it's a cav, so 40 hammers

so i believe that, with ginger spamming endless tanks and us spamming endless cannon + rifle/cav combo, they would need more than their roughly 5:3ish (just eyeballing there) edge in production to break us. now they could take cities anyways if we can't concentrate all our force in one place and they CAN..... but this is why our next tech is to be railroad if it looks like ginger is on the tank-spam path.

of course, we could also try defending in cities with city garrison draftees..... the proposition of which prompts the following almost unbelievable image



a CG3 tokurifle in a city with 60% defenses (kirishima) being attacked across a river (also kirishima) ACTUALLY BEATS a fresh tank of ginger's, straight-up  city raider promos do not help much either as the base strength of the rifle is so much lower than that of the tank. now our draftees will only be CG2 of course, but.... the unit in this example is unfortified so that should make up for it. so if we can quickly stack a bunch of draftees in the cities ginger can 1t with tanks from inside their borders, we might be able to pull off a miracle on the defensive here as well (although we do need at least two of 60% defenses / river / CG3 to get the edge, so a CG2 draftee in takayasu does still lose roughly 60-40 to a two-promo tank)

[williams482]

Now that is marvelous. Promotions and collateral damage can do some really wonderful things.

I am puzzled as to why you backed off from drafting, and with this I'm even more insistent that you keep it up. Four 110h rifles for eight pop is a good deal even if it slows the development curve somewhat, and if the units aren't actually all that needed then conventional production can be reallocated from troops to infrastructure instead.

[ljubljana]

yeah, i actually wanted to ask about that so you can evaluate my (very vague and heuristic) logic.... the reason i backed off drafting is because, as of right now, the vast majority of our cities look more or less like this



if we want to maximize military production from this city, should it be drafted? heuristically, i want to say no - drafting converts 19 + 18 = 37 food into a 110 hammer rifle at a rate of 1 food to 2.97 hammers, but with the forge in place, working a windmill converts 1 food into 5 hammers (plus gold!), and a slow-built rifle comes with one more promotion than a draftee. that extra promotion is also highly significant as it represents the third combat-line promotion that would allow us to take ambush - or alternatively, instead of the marginal value of a better rifle, slow-building could instead give us the value of letting us put our hammers into a cannon or a cav instead, if we think those are better buys by more than the promotion is worth.

that elides the time-horizon question though, and there is probably SOME level of imminence that a ginger attack could have (and maybe we've reached that point by now?) in which we should draft here, since a slow-built unit won't come out in time. but let's say we're optimizing production over the next 10 turns.... then the draft turns 37 food into 110 hammers, but working 2 more windmills instead over that time would have turned 20 food into 100 hammers of a slightly better unit. i think the hammer yields have comparable worth to us factoring in the extra promotion, and we come out way ahead on food by not drafting. so i guess for drafting to be correct here, we'd have to a) have reached the point where a draftee would reach the front in time to meet an attack but not a hand-built unit and b) be convinced that the extra units from a round of such draftees will be the difference-maker in saving cities of ours from capture, enough cities to make up for the food we'd lose.

so are we there yet? i didn't think so 10 turns ago when i slowed down the drafting, and indeed, since we haven't been attacked in that time i feel like the above analysis suggests that i was probably right to so think..... but with industrial coming for ginger in probably 3-4 turns, i'm guessing they are around 8 turns from being able to throw a 40ish unit tank stack at us, so perhaps we have reached that point now? but throwing away food like that just feels so short-termist, in a situation where our yields are so far behind the leaders that we really can ill afford to make those kinds of sacrifices unless we are really damn confident that an attack is coming when we think it is

so that logic is kinda vague and feelingsy, right? let's be a little more quantitative - consider nishikigi



well, consider a nishikigi that is size 8, and therefore maxed out on tiles, and has one more desert hill windmill to work. should we draft THAT city? drafting gets us a 110 hammer rifle, but takes the city from size 8 at, i think, 17/36 food stored to 17/32 food at size 6, with a +2 food surplus. so it will take 8 turns to regrow to size 7, in which time we miss out on 8 * 6 * 1.25 = 60 hammers from not working two desert hill windmills, and we will then be at 17/34 food. then it takes 17 turns to grow back to size 8, and we miss out on 17 * 3 * 1.25 = 63.75 hammers from the one windmill we're not working relative to if we had never drafted. so in terms of raw hammers, not drafting DOES still come out ahead.... but only barely, by 13.75 hammers, and not until something like 20 turns after the draft action is taken. on the other hand we also lose commerce from drafting, 8 * 2 + 17 = 34 of it, and we're getting a slightly worse unit.

so by drafting, we are paying 34 gold, 13.75 hammers, and a promotion to get a rifleman unit now instead of whenever natural production would make one - is that worth doing? well, i hadn't done the math when i made the decision, but since nishikigi is not actually making a rifleman right now, i thought that it was, and i did draft this city last turn. but if we were talking about a city like a food-capped version of myogiryu, where we'd be coming off PLAINS hill windmills to draft, i'd be much less sure. that city would lose 50 hammers and 34 gold over 25 turns by drafting, which is starting to approach the value of a whole extra unit that we give up just to accelerate the one unit we do get. of course it's still right to do that en masse if we will otherwise lose cities that we could have held with the extra draftees, but how confident are we of that, really?

of course, things will look completely different once we reach the point where our cities are tile-capped and can draft off of specialists instead of tiles that give a better food->yields conversion rate than drafting. that is where i am trying to get by power-growing like this, but we are not quite there yet. the closest we have to such a city is gondar on the border with america, which i think will grow onto its first specialist after this next turn? so we are not far from the point where mass-drafting will be obviously correct, and we will still get there eventually if we start mass-drafting now anyways, but i'm not sure if that's what we should be doing....?

[ljubljana]

ok, one more try: what about high-food surplus cities such as



this city neighbors low-food myogiryu to the west and national epic abi to the north, so that unworked plains windmill is the last tile it will ever grow onto. it's at +7 food surplus now with 32/42 in the box, and makes 30 hpt base after the forge (but before the PRO market boost). in terms of just raw hammer count, if we don't draft, it will make 30 hammers for the next 2 turns, then 35 hammers forever afterward. but if we DO draft this turn, we go to 32/38 at size 9 with +9 food surplus, and then we have

- 1 turn at size 9 making 20 hpt, grow to size 10 with 22/40 food and +8 surplus
- 3 turns at size 10 making 25 hpt, grow to size 11 with 26/42 food and +7 surplus
- 3 turns at size 11 making 30 hpt, grow to size 12

so in those 7 turns we lose 10, 5, 10, 10, 10, 5, 5, and 5 hammers relative to baseline = 60 hammers, so drafting's better? well, maybe, but that does ignore the significant commerce we'd lose by not hitting the tile cap and working specialists.... then again, we just said that specialists are bad tiles and the point of this setup is to hit a steady-state where we draft off as many of them as possible to convert the food invested therein into hammers instead. so if drafting off specialists that we are currently working is in the "clearly correct" bucket, perhaps drafting off a city that is close to the food cap, so as to convert FUTURE turns working specialists into hammers, is still in the "pretty good" bucket? or at least in the "better than drafting off of pure plains windmill turns at myogiryu" bucket? is.... is that right?

ps shout-out to these three cities, which are all resourceless and were named some variant of "miserable filler" when pinned, and, jesus, look at them now.... biology is wild, yall

[ljubljana]

anyway

249

which i definitely will not have time to finish in one session since it looks like i'll be making some significant drafting decisions....

anyways just another normal builder turn WAIT omg that's right it's great person time!! ok, we are really hoping for the minority chance of a non-merchant GP here so abi can run up to FOUR merchant specialists on top of its current compliment to accelerate the last, and most obnoxious, GP without creating a chance of a conflict. we have just one each of temple and courthouse there so if we pop merchant and then an artist or scientist from abi's current GP, we'll have to generate the last one via 2x artist/scientist + 1x engi, spy, and prophet which is only 5 specialists, which is really not enough to get there anytime soon. so who's it gonna be?



*silent fist pumps*

industrial in 3, and now they are not making gold so i suspect THAT represents their true timing. so priority 1 is making sure we can actually hit steel in 2, which we are a bit off of right now and will need a few wealth and research builds to actualize



here is a more accurate representation of the degree to which i am levee-brained at the moment

[williams482]

I find your math extremely persuasive here. If drafting off windmills is a marginal choice purely on production grounds and a massive loser on commerce grounds, then we probably shouldn't be doing it unless conditions are urgent (and I agree with you that they do not appear to be right now).

I think your plan to grow onto and then draft off of specialists is the stronger approach here, especially because the deferral hopefully allows us to turn some of those pops spent on drafted rifles into drafted infantry, a much better deal. Not that we're all that close to Assembly Line, but you gotta have hope. Specialists are dramatically lower yield than those farms and windmills and that makes a big difference.

[ljubljana]

well, i think you did convince me to do a bit more of it starting a bit earlier than i was otherwise planning to - if we think we could be getting tanked in 8 turns, that actually is pretty urgent, especially if our draftees need full fortify bonuses to have good odds against a tank... i am going to try not to draft off of the myogiryus of the world but we have a decent number of tomokazes and nishikigis as well, where it seems more viable

the other thing is that in many places we are about to start growing onto coast tiles instead of specialists.... i am confident that a coast tile is better than a specialist, but is it better by enough that it is worth working, or do we draft off of them whenever possible? i have a lot of trouble really assessing the value of tiles that are not of the form "turn food into large amounts of some other resource" but are instead more like "receive, for free, some very minor benefit". i guess one basis for comparison is to a (100 hammer PRO) market, right? growing onto 2 coasts will give us 4 gold (most of our coastal cities do not have multiplier buildings...), which is like a market that makes 8 gold at 100% commerce since our breakeven is around 50%, so a market in a city making 40 raw commerce. and then i could "un-build" that market by drafting and turning it into a (somewhat less valuable than a hand-built 110 hammer) draftee rifleman instead.

tbh, that hypothetical market is a lot better than some of the ones i've actually made - i just built one in tomokaze which, as you can see, is a 23 commerce city (though part of the motivation for that one was to open up non-blank second specialist slots and more happy to draft off of). and we're still running wealth and research builds enough of the time (hell, i'm running some right now) that it's not all that unreasonable to rate a 2/0/2 coast tile as similar in value to a 2/2/0 tile, if working the coast means i DON'T have to build an equivalent amount of wealth somewhere.... so does that mean we're supposed to actually work these coast tiles when we could be turning them into rifles? if so that is contrary to ALL of my accumulated (and possibly misguided) civ intuition

oh and re specialists being super weak: they are so weak that they are even weaker than the specialist that produces 3 hammers, aka turning one of our plains farms into a watermill..... i expect i will want to do some of this at some point, except we really don't have the worker labor for it. we have enough unimproved tiles still that we're starting to run into them at some of our faster-growing cities (ura is threatening to grow onto a bare desert hill this turn....) and are about to gain access to another enormous worker sink in rails. do we actually need to make MORE workers, this late in the game???

[williams482]

Coast tiles are better than specialists because they allow growth to continue, and growth at that level of marginal returns is most valuable to permit further drafting. So without having run numbers, my gut check is to draft off of those water tiles.

I don't think that case for 1c = 1h actually works very well, if nothing else because wealth and research builds are generally targeted, small scale things around the edges to get stuff over the line or to convert marginal production into a more immediately useful yield (I assume; you aren't regularly building wealth in cities that could be 6 turning cavs, right?).