City Comparison by Race

[Anthony]

I wrote a build order toy that decides what a city should build next, ignoring unit production, using various weightings for the resources a city can produce; for high men, in a location with max population 16 and taxes set to 1.5g, the build order was:
Builder's Hall, Granary, Smith, Marketplace, Farmers' Market, Shrine, Library, Sawmill, Temple, Foresters' Guild, Miners' Guild, Sages' Guild, Animists' Guild, University, Bank, Mechanicians' Guild, Parthenon, Ship Wrights' Guild, Ship Yard, Merchants' Guild, Alchemists' Guild, Cathedral, Wizards' Guild (because the logic checks current population and unrest levels, order may vary slightly by race. The code will attempt to buy, though its threshold for spending money means mostly it only applies to builders hall, granary, marketplace, and farmers' market). Anyway, it might be interesting to look at cumulative statistics after 5, 10, 15, and 20 years, by race (columns: racename, turn(total production), population, production, gold, mana, research, unrest, current build target).
5 years:
Ba 60(340): 6800(+100)pop, 8p, 81(+10)$, 0(+0)m, 0(+0)r, 0u: Farmers Market(44)
Be 60(340): 5730(+90)pop, 6p, 19(+7)$, 65(+2)m, 0(+0)r, 0u: Farmers Market(88)
DE 60(240): 4530(+70)pop, 5p, 82(+3)$, 130(+4)m, 0(+0)r, 0u: Marketplace(80)
Dr 60(240): 5150(+80)pop, 6p, 96(+4)$, 58(+2)m, 0(+0)r, 0u: Marketplace(54)
Dw 60(340): 4710(+70)pop, 7p, 104(+15)$, 0(+0)m, 0(+0)r, 0u: Farmers Market(58)
Gn 60(240): 5150(+80)pop, 6p, 96(+4)$, 0(+0)m, 0(+0)r, 0u: Marketplace(54)
Hf 60(340): 5810(+90)pop, 7p, 109(+7)$, 0(+0)m, 0(+0)r, 0u: Farmers Market(100)
HE 60(240): 4530(+70)pop, 5p, 82(+3)$, 49(+2)m, 0(+0)r, 0u: Marketplace(80)
HM 60(340): 5730(+90)pop, 6p, 19(+7)$, 0(+0)m, 0(+0)r, 0u: Farmers Market(88)
Kl 60(240): 5150(+80)pop, 8p, 67(+4)$, 0(+0)m, 0(+0)r, 0u: Marketplace(22)
Lz 60(340): 6270(+90)pop, 8p, 54(+10)$, 0(+0)m, 0(+0)r, 0u: Farmers Market(64)
No 60(240): 5150(+80)pop, 6p, 110(+5)$, 0(+0)m, 0(+0)r, 0u: Marketplace(54)
Oc 60(340): 5730(+90)pop, 6p, 19(+7)$, 0(+0)m, 0(+0)r, 0u: Farmers Market(88)
Tr 60(240): 4530(+70)pop, 5p, 82(+3)$, 0(+0)m, 0(+0)r, 0u: Marketplace(80)
10 years:
Ba 120(1700): 14300(+110)pop, 39p, 659(+13)$, 102(+3)m, 90(+2)r, 1u: Parthenon(154)
Be 120(1300): 12270(+100)pop, 28p, 506(+14)$, 383(+9)m, 76(+2)r, 0u: Miners Guild(104)
DE 120(1000): 9970(+90)pop, 24p, 259(+8)$, 549(+10)m, 68(+2)r, 0u: Temple(56)
Dr 120(1300): 11180(+90)pop, 28p, 388(+11)$, 316(+8)m, 66(+2)r, 0u: Miners Guild(194)
Dw 120(1670): 10690(+90)pop, 48p, 1290(+50)$, 91(+6)m, 98(+5)r, 0u: Trade Goods(1)
Gn 120(1300): 11180(+90)pop, 28p, 388(+11)$, 46(+3)m, 66(+2)r, 0u: Miners Guild(194)
Hf 120(1670): 12400(+100)pop, 42p, 501(+9)$, 67(+3)m, 68(+5)r, 0u: Alchemists Guild(250)
HE 120(1000): 9970(+90)pop, 24p, 259(+8)$, 255(+5)m, 68(+2)r, 0u: Temple(56)
HM 120(1300): 12270(+100)pop, 28p, 506(+14)$, 79(+3)m, 76(+2)r, 0u: Miners Guild(104)
Kl 120(1100): 11280(+90)pop, 54p, 839(+37)$, 17(+1)m, 40(+2)r, 0u: Trade Goods(1)
Lz 120(700): 13190(+100)pop, 20p, 921(+30)$, 101(+3)m, 84(+2)r, 0u: Trade Goods(1)
No 120(1300): 11080(+90)pop, 28p, 551(+16)$, 47(+3)m, 66(+2)r, 0u: Miners Guild(200)
Oc 120(1300): 12270(+100)pop, 28p, 506(+14)$, 79(+3)m, 76(+2)r, 0u: Miners Guild(104)
Tr 120(1000): 9970(+90)pop, 24p, 259(+8)$, 27(+1)m, 68(+2)r, 0u: Temple(56)
15 years:
Ba 180(2070): 19930(+80)pop, 54p, 2676(+45)$, 591(+9)m, 369(+5)r, 1u: Trade Goods(1)
Be 180(4400): 17400(+70)pop, 72p, 1118(+9)$, 1042(+17)m, 527(+10)r, 0u: Cathedral(239)
DE 180(3050): 14660(+70)pop, 62p, 826(+7)$, 1438(+17)m, 471(+10)r, 0u: Ship Wrights Gui(38)
Dr 180(3900): 16050(+70)pop, 54p, 1093(+31)$, 873(+14)m, 533(+10)r, 0u: Parthenon(400)
Dw 180(1670): 15190(+60)pop, 63p, 5140(+76)$, 451(+6)m, 398(+5)r, 1u: Trade Goods(1)
Gn 180(1300): 16050(+70)pop, 44p, 2218(+40)$, 226(+3)m, 186(+2)r, 1u: Trade Goods(1)
Hf 180(2870): 17490(+70)pop, 58p, 1747(+40)$, 656(+13)m, 368(+5)r, 0u: Trade Goods(1)
HE 180(3050): 14660(+70)pop, 62p, 826(+7)$, 773(+10)m, 471(+10)r, 0u: Ship Wrights Gui(38)
HM 180(4250): 17400(+70)pop, 72p, 1118(+9)$, 295(+6)m, 527(+10)r, 0u: Merchants Guild(106)
Kl 180(1100): 16120(+70)pop, 69p, 3610(+57)$, 77(+1)m, 160(+2)r, 0u: Trade Goods(1)
Lz 180(700): 18590(+80)pop, 22p, 2945(+39)$, 281(+3)m, 204(+2)r, 2u: Trade Goods(1)
No 180(3850): 15970(+70)pop, 62p, 1613(+16)$, 227(+3)m, 525(+10)r, 0u: Merchants Guild(352)
Oc 180(4250): 17400(+70)pop, 72p, 1118(+9)$, 295(+6)m, 527(+10)r, 0u: Merchants Guild(106)
Tr 180(2580): 14660(+70)pop, 38p, 887(+24)$, 376(+10)m, 188(+2)r, 0u: Trade Goods(1)
20 years:
Ba 240(2070): 21050(+0)pop, 54p, 5396(+46)$, 1131(+9)m, 669(+5)r, 2u: Trade Goods(1)
Be 240(5400): 21050(+0)pop, 85p, 3313(+55)$, 2231(+20)m, 1463(+18)r, 0u: Trade Goods(1)
DE 240(5500): 18100(+50)pop, 75p, 2596(+64)$, 2739(+24)m, 1215(+18)r, 0u: Trade Goods(1)
Dr 240(5700): 19660(+50)pop, 65p, 3526(+63)$, 1986(+19)m, 1293(+18)r, 0u: Trade Goods(1)
Dw 240(1670): 18500(+50)pop, 65p, 9980(+86)$, 811(+6)m, 698(+5)r, 2u: Trade Goods(1)
Gn 240(1300): 19660(+50)pop, 50p, 4786(+48)$, 406(+3)m, 306(+2)r, 2u: Trade Goods(1)
Hf 240(2870): 21010(+0)pop, 65p, 4489(+51)$, 1436(+13)m, 668(+5)r, 1u: Trade Goods(1)
HE 240(5100): 18100(+50)pop, 70p, 2742(+59)$, 1461(+12)m, 1263(+18)r, 1u: Trade Goods(1)
HM 240(6300): 21050(+0)pop, 85p, 4140(+78)$, 899(+10)m, 1367(+18)r, 0u: Trade Goods(1)
Kl 240(1100): 19710(+50)pop, 84p, 7455(+71)$, 137(+1)m, 280(+2)r, 0u: Trade Goods(1)
Lz 240(700): 21030(+0)pop, 23p, 5485(+44)$, 461(+3)m, 324(+2)r, 3u: Trade Goods(1)
No 240(5300): 19600(+50)pop, 80p, 5163(+85)$, 811(+13)m, 1125(+10)r, 0u: Trade Goods(1)
Oc 240(6300): 21050(+0)pop, 85p, 4140(+78)$, 899(+10)m, 1367(+18)r, 0u: Trade Goods(1)
Tr 240(2580): 18100(+50)pop, 46p, 2681(+38)$, 976(+10)m, 308(+2)r, 0u: Trade Goods(1)

Lizardmen, because they max out so fast, do surprisingly well (I suspect my build logic is insufficiently interested in trade goods), and no attempt is made to account for terrain bonuses, including road/shore/river gold, though nomads get their bonus. Overall, though, unless you're looking at 20% or higher racial unrest, building colonies of one of the four cost-60 settlers (Ba, HM, No, Oc) looks pretty consistently like your best choice, assuming you have any options.

[I like Serena]

Nice analysis. I am still wondering how different races perform with various build orders.

Btw, I usually set my tax higher than you apparently did (2 +/- 0.5 gp/turn).
That's because support cities only have to generate resources so I can buy off production in the important cities that are building fighting stacks.
Furthermore, it increases the chances on heroes, useful artifacts, and mercenaries (their offering is skipped if you do not have enough money).
As a result it seems to be better to build a shrine before or after the market place, depending on race and terrain bonuses.

What is your reason for building the Alchemist's guild so late?

Have you considered building housing?
There was a thread where someone analyzed the best build order to generate maximum growth.
His conclusion was that it pays off to build housing up to pop 8 after farmers' market but before a shrine. He did cast gaia's blessing on the town as soon as possible.

Since I am currently working on automatic building queues, I'm considering to try an automatic series of turns on a map with sufficient cities.
Then redo it with a different build order queue.

Oh, and happy new year!!

[Anthony]

Nice analysis. I am still wondering how different races perform with various build orders.

Btw, I usually set my tax higher than you apparently did (2 +/- 0.5 gp/turn).

2G gets very annoying when you've got racial unrest modifiers, and it only makes around a 10% difference even with no modifiers.

As a result it seems to be better to build a shrine before or after the market place, depending on race and terrain bonuses.

The main reason for the marketplace is that the farmers' market is very high value. If you change the weighting of the population growth modifier, the next building after the granary is either the Library or the Shrine.

What is your reason for building the Alchemist's guild so late?

The way the build simulator works is that it looks at the city's production with and without the building (using current population+1, as a crude way of representing growth during the build process), gives that a point score (current weighting is 5 for gold, 8 for production or research, 10 for power, and 1 per point of population growth rate; food just automatically adjusts number of workers and has no value), and divides by build cost. +3m/-3g for 250 isn't very good; the alchemists' guild is only really valuable if you're building units. I suspect the way I value production may be off, I probably shouldn't be building a mechanicians' guild (production is not directly useful, it just lets you build other things).

Have you considered building housing?

I don't have a current method of weighting that.

Oh, and happy new year!!

You too.

[Anthony]

Crunching some numbers, +1 population (at tax rate 1.5g) gives you +0.5 farmers, +0.2 workers, +0.3 rebels, +0.65 base production, +1.05 base gold. With a marketplace, that's 1.58 base gold. At the value scheme above, that's worth +5 (from production) +8 (from gold) = 13. By comparison, a library (+2 research, -1 gold) is worth 11; thus, +1 population is worth about 70 (the library is a very efficient building). For a pop-5 village with granary, farmers' market, and 2food taken up by a garrison, production is 9 and it takes 7 turns to build a library, so if we can get a net of +1 population from 9 turns of housing, it's worthwhile. With 4 workers/1 farmer, building housing is +95% pop growth rate, so if our base pop growth rate is at least 120 we achieve our desired rate. For high men, pop growth rate is (maxpop-curpop+1)*5+50, so we need (maxpop-curpop) of 13, or a pop max of 18. I guess that's sometimes going to be worthwhile.

[Anthony]

Oh, comparison with different tax rates: at tax rate 1g my cities start losing money, which my logic isn't smart enough to cope with.
Tax rate 1.0g (my logic isn't smart about negative cash flow)
HM 60(240): 5730(+90)pop, 6p, 48(+2)$, 0(+0)m, 0(+0)r, 0u: Marketplace(32)
HM 120(1520): 12080(+100)pop, 37p, 89(+2)$, 18(+1)m, 88(+5)r, 0u: University(226)
HM 180(4500): 17270(+70)pop, 72p, -78(+4)$, 218(+6)m, 653(+10)r, 0u: Parthenon(256)
HM 240(6300): 21000(+0)pop, 85p, 1265(+47)$, 833(+10)m, 1501(+18)r, 0u: Trade Goods(1)
Tax rate 1.5g
HM 60(340): 5730(+90)pop, 6p, 19(+7)$, 0(+0)m, 0(+0)r, 0u: Farmers Market(88)
HM 120(1300): 12270(+100)pop, 28p, 506(+14)$, 79(+3)m, 76(+2)r, 0u: Miners Guild(104)
HM 180(4250): 17400(+70)pop, 72p, 1118(+9)$, 295(+6)m, 527(+10)r, 0u: Merchants Guild(106)
HM 240(6300): 21050(+0)pop, 85p, 4140(+78)$, 899(+10)m, 1367(+18)r, 0u: Trade Goods(1)
Tax rate 2.0g
HM 60(340): 5650(+90)pop, 6p, 64(+7)$, 1(+1)m, 0(+0)r, 0u: Marketplace(94)
HM 120(1300): 11940(+100)pop, 25p, 751(+18)$, 111(+3)m, 40(+2)r, 1u: Miners Guild(300)
HM 180(3250): 17160(+70)pop, 50p, 1777(+28)$, 360(+6)m, 285(+10)r, 2: Oracle(170)
HM 240(6800): 21030(+0)pop, 80p, 4523(+100)$, 839(+10)m, 909(+18)r, 1u: Trade Goods(1)
So yes, it arguably comes out better with tax rate 2.0 (you get more gold, less power and research). Now, with a 10% racial unrest:
Tax rate 1.5g
HM 60(240): 5650(+90)pop, 4p, 106(+3)$, 0(+0)m, 0(+0)r, 1u: Marketplace(57)
HM 120(1000): 11830(+100)pop, 18p, 361(+10)$, 71(+3)m, 26(+2)r, 1u: Foresters Guild(110)
HM 180(3350): 17080(+70)pop, 54p, 762(+13)$, 311(+6)m, 247(+10)r, 1u: Mechanicians Gui(346)
HM 240(6800): 20960(+60)pop, 85p, 2220(+68)$, 795(+10)m, 895(+18)r, 0u: Trade Goods(1)
Tax rate 2.0g
HM 60(100): 5080(+60)pop, 2p, 184(+7)$, 0(+0)m, 0(+0)r, 1u: Granary(40)
HM 120(640): 9790(+110)pop, 11p, 393(+14)$, 27(+1)m, 0(+0)r, 2u: Temple(134)
HM 180(1400): 15550(+80)pop, 20p, 1321(+16)$, 175(+3)m, 76(+2)r, 5u: Parthenon(120)
HM 240(3350): 19810(+60)pop, 52p, 2297(+25)$, 517(+6)m, 361(+10)r, 3u: Ship Wrights Gui(100)
I think that's overall worse (the lower population is due to shrine before farmers' market); gold is higher, but power and research is significantly worse.

[I like Serena]

Interesting!

I consider research in early-game not particularly useful, since there are no interesting spells to be had.
So building libraries just slows me down in the phase of the game where it is most important to quickly grow and spit out relatively strong units to conquer cities.
I only build libraries in mid-game, and from then on I put them early in my build order.

I usually never build mechanicians guilds or wizard's towers. They take too long. It's more useful to have other buildings a little bit earlier when they still make a difference.

It does look pretty bad to tax 2.0 gp with racial unrest.
You did build a garrison with 10% racial unrest?

Anyway, in my opinion it's less important what unrest sensitive cities do, but more important what the actually useful cities do - typically your capital.
If your capital (or another city) can produce a little bit faster to get to effective units more quickly, it's worth it if other cities are struggling. That is, until you get to the point where you only want to score points by increasing population (when the game is already won).
As a rule I increase the tax to where gold production doesn't increase anymore or goes down, and then set it one step lower, leaving me with a reasonable production.
That also means I may lower the tax rate if I have many unrest sensitive cities.
Of course I'll increase it again if I can cast Just Cause, or when my garrisons, shrines and farmers' markets are mostly completed.

[Anthony]

Interesting!

I consider research in early-game not particularly useful, since there are no interesting spells to be had.

Not sure which period you think of as 'early game' (Summon Hero is nice to have fairly early), and if you have low numbers of books there may be cheap commons you didn't start out with and want to research (of the research cost 20 spells, bless, warp wood, and skeletons are easily worthwhile).

So building libraries just slows me down in the phase of the game where it is most important to quickly grow and spit out relatively strong units to conquer cities.

Unless you've got Heroism, I'm pretty unimpressed with capturing cities with regular units, but I'm likely to produce settlers and engineers after the farmers' market.

[I like Serena]

Not sure which period you think of as 'early game' (Summon Hero is nice to have fairly early), and if you have low numbers of books there may be cheap commons you didn't start out with and want to research (of the research cost 20 spells, bless, warp wood, and skeletons are easily worthwhile).

Unless you've got Heroism, I'm pretty unimpressed with capturing cities with regular units, but I'm likely to produce settlers and engineers after the farmers' market.

Here's my view. Open for debate of course.

Early game: spearmen with heroism, pack of hell hounds, gang of ghouls, dwarves, doom drakes. Support with phantom warriors.
Late early to mid game: army of longbowmen, slingers, wolf riders, pikemen, hammerhands.

Mid game: heroes. Preferably trade Summon Hero with an opponent, which is quicker than researching it yourself.
Prayer if you can research or trade it.

Late game: champions, paladins, adamantium hammerhands, stag beetles.

After the game: sky drakes, death knights, colossi, great drakes.

I try to avoid producing settlers. Too slow. Better to produce longbowmen or slingers a little bit earlier.

[Anthony]

Here's my view. Open for debate of course.

Early game: spearmen with heroism, pack of hell hounds, gang of ghouls, dwarves, doom drakes. Support with phantom warriors.

Well, depends on race. I'd go with swordsmen for Halfling, Lizardmen, and Troll, and for High Elves, High Men, Nomads, and Orcs their early game units might as well not exist, just build immediately towards more advanced stuff (for High Elves that's longbowmen, for High Men that's Pikemen; for Nomads there's some virtue to Horsebowmen. Orcs should just give up and produce resources).

I try to avoid producing settlers. Too slow. Better to produce longbowmen or slingers a little bit earlier.

It blocks the AI from colonizing, which it otherwise likes to do (particularly with Insecticide), and sites near minerals or Wild Game pay off quite fast (strangely, Wild Game actually produces food when you have an outpost, whereas minerals don't produce anything until you have a hamlet).

[I like Serena]

Here's my view. Open for debate of course.

Early game: spearmen with heroism, pack of hell hounds, gang of ghouls, dwarves, doom drakes. Support with phantom warriors.

Well, depends on race. I'd go with swordsmen for Halfling, Lizardmen, and Troll, and for High Elves, High Men, Nomads, and Orcs their early game units might as well not exist, just build immediately towards more advanced stuff (for High Elves that's longbowmen, for High Men that's Pikemen; for Nomads there's some virtue to Horsebowmen. Orcs should just give up and produce resources).

I try to avoid producing settlers. Too slow. Better to produce longbowmen or slingers a little bit earlier.

It blocks the AI from colonizing, which it otherwise likes to do (particularly with Insecticide), and sites near minerals or Wild Game pay off quite fast (strangely, Wild Game actually produces food when you have an outpost, whereas minerals don't produce anything until you have a hamlet).

If you march out spearmen of High Elves, High Men, Nomads, or Orcs, even before a marketplace, with heroism, they can take most cities. You can leave your capital empty for a couple of turns to achieve it.
Furthermore, you can produce twice as many spearmen as you can produce swordsmen for a lower upkeep per unit.
And as elite there's not that much of a difference, assuming you have more spearmen figures than swordsmen figures.
Only works if you can quickly locate a city to conquer. But if there is, it gives you a nice boost.

Admittedly, it's a fairly small window of time that you can do this.
Otherwise you're stuck with those races until you can produce longbowmen or pikemen.
And yes, with orcs the only way to go is with spells and let them produce the power to cast them (unless you can conquer a useful city in those early turns).
Orcs have to be combined with a spell strategy to be feasible.
So you might as well produce 1 or 2 settlers after farmers' market to boost power, hopefully before an opponent or neutrals start to invade. Either way, you're handicapped with orcs.

And yes, if I find bonus tiles nearby like mithril, preferably together with wildgame, I immediatly go for a settler - possibly even before a marketplace.