Guys, where you find those lobby games that last till cuirrassiers? On the other hand, Cataphracts end the game prematurely too often
I could see it if someone slaps down a larger map.
When given a map of x size there's always going to be a window where y era units dominate the map. I bet you don't see a lot of knights in a dual pangaea setup, just like you wouldn't see a lot of wins using knights or cataphracts on a huge world. In practice it's less extreme but hopefully you get my point.
Ironically what has forced me to see the spec light a little bit better is the engineering 500 BC bulb gimmick and applying it on deity (Ironically, if you survive until a bit after 1k BC you actually PREFER warmonger neighbors, as they tech poorly and are ill-equipped to deal with xbow + pike + treb, and lack castles to stop the advance...feels like a cannon war).
While tooling around on g major 96 I've failed, but only because it's quick game speed. Earlier in this analysis, I discounted GPP, but you can't do that realistically in practice. In reality you can easily get 100's of GPP in multiple cities in the BCs, BEFORE multipliers. We're talking about the potential for 1000's of beakers. The problem is of course that when tech trades are off, you can't distribute them to non-bulb techs via tech sale and trade.
Even so, the results have me questioning cottages outside bur/very strong riverside site/FIN. I look at non-fin non-river cottage and it's going to take it 30 turns just to match a non-rep spec at 100% science! But that's matching in output. After that, the cottage doesn't actually outproduce the spec until it grows again and is worked long enough. How early can you place that sucker down and work it? I don't see it outside a bur cap. It's going to take 70 turns for a cottage to catch a non-rep scientist in science contribution if it's non-river non-grass. Now consider you're looking at around t80 with a new city. Do you REALLY want to be flat-out behind in raw yield until t150?
The problem doesn't stop there. It discounts bonuses like those reaching currency, CoL, etc faster in those 70 turns, making the payback time take even longer. This is BEFORE GPP!
Yes, a cottage doesn't cost food like a spec, but early game the
cap is a real bottleneck, and it is scientists that get you out often. By the time that happens, you're already flirting with the time period where cottages outside bur cap won't be decisive in time.
As Vicawoo points out, sometimes you can't spread irrigation right away, are already running 2 scientists + specials, and you need something to grow on. I can see cottages there, in bur cap (makes them 50% stronger pre multiplier and suddenly they look awesome), and when captured as villages/towns.
Most competent setups are going to run enough scientists to get the GPP, so this really becomes a question of how one is using the secondary, non-dedicated production non-capitol cities. IMO an earlier look at key early techs and maybe 1 extra great person (sometimes more) in the relevant timeframe is just stronger.
The more I think about it, the more I concur with U_Suns earlier analysis of this cottage vs farm tradeoff. Flexible production + more beakers now is going to beat more beakers later most of the time, because more beakers now allow one to reach important milestones that speed up research faster.