I recall talking about it back in 2015 when farms and mines were just implemented into the game.
Well, the game is out for quite some time now and, I believe, it's time to return to this.

Simply put, the combination of mechanics leaves quite a narrow spectrum of successful strategies for pretty much any race.
Consciously ignoring force majeures like the neighboring race not attacking you despite having BFF relations with you, the metagame usually looks the same way.

1) Metal is severely limiting you in early game (as you don't have much) and late game (as there's no T3 mining tech).
2) Food does not limit the population's growth, the lack of food only culls it afterwards.

The first point discourages forge-worlds and researching factory tech.
A fully developed production world can't do anything except churning out ships and bigger better factories consume metal much faster.
Yeah, sure, you get the ship in two turns instead of six, but it's an unsustainable pace.
On the other hand, a world with two T1 factories is just fine for them heavy cruisers, especially if you have that mid-tier tech that greatly reduces the production points (but not metal) cost of ships, and it can trade or research as well.

The second point requires you to have at least some research production.
Farm tech just delays the inevitable and a late game empire usually resembles Suthlam from Tuf Voyaging.
But at least it makes sure that your most producing planets do not lose population and don't hold a grudge.
The clock is ticking from the very beginning and if you foolishly colonize at instant, your colonists will build a farm for tens of turns as you don't have a spare transport to boost the population... and then your homeworld celebrates an n+1-th million citizen.

But you'll need a lot of colonies to get you that metal, because without ships you'll be crushed inevitably.
The answer is, of course, money to rush-buy a farm, making market theory a crucial early-game technology.
Also a single market pacifies natives, unlocking these big planets for you without the need to spend 135 metal onto a tank.

Markets, mines, farms, a factory or two and labs get squeezed out.
More than that, with increasing costs, two or three labs are becoming ineffective without a factory to help build them.
And you need RPs, because even if you don't have anyone nearby, the hungerday clock is still ticking.

While extensive colonizing may be an answer, there is a better deal: science destroyers.
Yoral get a boost here, with their destroyers having three system slots, so they can house two science labs, but even a run-off-the-mill destroyer (without armor and weapons) is still good.
And they can be produced anywhere. And mass-refitted into cannon fodder in a pinch as you'll have that metal stockpiled slowly.
So, with a pair of colonies with science destroyers on repeat, you're pretty much set until you solve all your early-game problems and will tackle T2 labs tech.
And even later they can allow for a comeback if you get crushed, but manage to force-piece the enemy.

The extensive colonizing is better suited for balanced market and labs rush-buy.
You will be able to afford to rush-buy a significant fleet on-demand and not worry about ETAs or peace-time upkeeps, let alone a good manufacturing base.
Coincidentally, it makes shipyards and fleetbases irrelevant, as they only consume precious metal.

You will need to deploy a mech on a conquered planet for sure, but then you rush-buy a market there and the former territory of your enemies is happily rush-buying transports you need for the continued invasion.
After the first war your snowball may be impeded only if you don't want it for roleplay reasons.

I'm not saying the current system stemming from these two points (and exacebrated by markets and science ships) is bad.
After all, just about 1/6 of the owners of the game finished a single game and that's normal for a 4X.
The core of the game is good, and the amount of games needed to identify a regrettably single optimal strategy is probably unhealthy.

However, only the engaged audience usually buys DLCs, so it may be a good time to return to these matters.

I'm not entirely sure what point you're trying to make here - it seems to be something about there being "only one optimal strategy"? Though it's also clear that your playstyle doesn't match what I'd consider optimal play - perhaps we're optimizing different things? Here, then, is a contrasting perspective on where my observations differ from yours:

I've found having one - or occasionally even two - dedicated factory worlds to be quite useful. Sure, you can't keep them fed with metal late-game if you try to repeat-build super-dreadnaughts or dread stars. But churning out, for example, 2x Orthin Gunboat every turn is a better deal than one heavy cruiser every six turns, and not unreasonably sustainable. As for the high metal costs on space stations and shipyards - I actually find that to be a strong argument in favor of the single factory world setup, as that way you don't need to build construction facilities at more than just the one planet.

Low pop planets I tend to focus on mining; higher pop planets tend to focus on research (with one or two farms apiece if I haven't found a supergrain planet). Buying a factory is usually the first thing on making a new colony - greatly speeds up whatever else you decide you're doing there, and as techs or new population types will add more building slots, it's useful to keep around. Markets, by contrast, are mostly skippable; they're a niche utility and you don't really need them in most places. Even early-game with planets that have natives; build an extra transport and ship population over there and you won't need tanks, even if some of them end up rioting for a while.

Farming is really only severely limiting at the very start of the game, before you've had a chance to research the first upgrade. After that - I'm just not seeing this "hungerday clock" problem you're describing.

This does, however, vary somewhat depending on what sorts of planets you have available; for example, in my most recent game, most of the planets I had access to had a max population of 4-6, with only a few in the 8-12 range (and one of those had viscids!) - I had to vary my building strategies a fair bit from my usual and go for more of a distributed research environment, with four-slot planets tending towards one factory, one lab, two mines.

I'm also confused as to what circumstances would make science destroyers a "good deal" - you'd need a dozen of them to match even a small planet with a single lab, and that seems like a very high investment in production and metal that could have gone towards other things. What kinds of numbers do you typically build these in? What's the price tag like in terms of production, metal, & maintenance costs?

Also not sure if there is only one single viable strategy. Both of the above seem to differ from what I am usually applying and it also works quite well. Naturally there are some overlapping but still:
1. Early game build order always starts with Farm (unless low fertility)
2. Markets are not researched until around turn 100 and focus is on mines instead wherever there is no low mineral
3. Same goes for research, only 1-2 of them get placed in early game as the pop will drive research up by itself enough. If it is not enough, scout ships are equipped with science labs and rockets or PD (so that in combat they can still support combat vessels)
4. Factories are rarely built at all, and only for two reasons: 1. Warship production worlds (there get at least 2 + shipyard) 2. Economic production world (these get max 2 factories and no usually no shipyard). The later are typically large worlds that are in not in a strategy location and have low fertility and even low mineral
5. Markets only come into play once a base transport production worlds has been set up

Early game does not suffer from food shortage, but the late game does doe to very high coverall pop counts and low quality small planets get conquered where markets usually take away space from farms to settle the moral.

Altogether I am happy with the mechanic and the above strategy as it is very fun for me and recognize there are other ways for going about it.
I think both metal and food constraint is thought out and implemented well. It might not have been the exact goal but I like to think that they serve specific purpose:

1. Food is a limitation for wide empires and both for overspecializing planets as without food they quickly get in distress during blockade + much of the trading fleet is bugged down by transporting food instead of making money (I love this mechanic )
2. Metal a the limiting factor for rush-producing massive fleets, also limiting over-specialization of planets (like mentioned above, it does not make sense to have 4-5 factories though the AI sometimes does this ) and later preventing again wide empires from snowballing as they constantly have to juggle their limited amount of metal to protect their stretched-out borders. This is in 99% of the cases the reason why the AI dies out on me in mid-game and never get to super-dreads. Both the AI and me are always short on metal once capital ships are available and makes for an additional decent intensive to produce destroyers and cruisers next to them as they are more cost effective PD (even though you would have the raw producing capacity to just chim out capitals). Works perfectly well in my books if only the AI could handle this better

I'm not saying "the only viable strategy".
I'm saying "the only optimal strategy".

As in if you follow it you'll win regardless of map composition, who's around and all that.
We'll need some yardstick as to measure what is successful and what isn't.

Here, I ran a sample game on brutal, 5 opponents, normal map size, normal pace, sparse stars, rare habitability.
100th turn - http://i.imgur.com/BfdlRS6.jpg - just met the first opponent, Ashdar colonials
121th turn - http://i.imgur.com/dVP2eiJ.jpg - sure enough, they declare war on me, but I just got gunships developed and buy them on all my four planets
200th turn - http://i.imgur.com/uVHHsG7.jpg - I'm slowly expanding, while grinding away colonials' fleets; also the humans have also declared war on me, pity; Ashdar imperials to the right were content the whole game
300th turn - http://i.imgur.com/q0U0Upv.jpg - after I made some advances into colonials territory, I became a bit bogged down, so I settled for peace with them and dismantled humans; as you can see, I already won
360th turn - http://i.imgur.com/73of622.jpg - I got bored and quit

Up to the late game science destroyers still comprised a sizeable amount of my science, though in the beginning they were about 70%.
I decided not to build any military ships except gunships (because of laziness) and it worked well.

Markets aren't niche, because with an early tech they increase the amount of trade you can do, effectively doubling the income from them.
A lot of income means that I can rush-develop the whole colony, gaining RPs, metal and coins while yours is still developing.
And a large transport fleet means I can afford to evacuate a whole planet if needed.
I don't have a lot of investments into a planet, just money, and money is plenty, so I can be flexible in defense.

I do tend to have tens of science destroyers.
Their cost depends on the pace selected, but on normal it's somewhat in range of 75 PP and 45 metal.
Compared to 300-375-450 for labs it's a bargain and you don't need slots for them. Just churn them out and even if you're locked in your region of space, you don't spend turns in vain.
In the midgame or on a planet with artifacts you can afford to create a research world, but that depends on what the map generator gives you.

Metal doesn't limit rush-production of ships, because there's no penalty.
It's just that you give up the cost of the ship in metal at once instead of per-turn, but the cost is the same.

And the power of rush-buy is hard to underestimate.
While you're building, I'm already ahead.
New planets, whether by colonization or conquest are fully developed and integrated in five-seven turns max.
You have the flexibility (as money and metal go everywhere, but production capacity doesn't) to win the first war in any circumstances (as your science output is mobile and is not reliant on planets whatsoever) on any difficulty.
And then it's mopping up.

Yeah, you're optimizing for different things. Your strategy is clearly viable, and I'll have to give it a try sometime just for variety, but it's certainly nothing I'd call optimal.

Ten science destroyers, for example, costs you 750 production and 450 metal - but only gives you 20 research. Compared to that, 300 production and zero metal for a single lab on even a tiny four-max-pop planet is a bargain; set on research, that's going to give you 24 research per turn for a much smaller investment. Even if you decide, for whatever reason, to not set the planet on research and to build two labs - that's 675 production and zero metal, still notably cheaper than your science destroyers. (And once you start getting improved labs, the advantage tilts even more heavily against science ships.)

As for rush-buy - I'll use it occasionally, sure, like getting the first market built on a newly-conquered world, or the first factory on a newly-colonized planet, but after that the places can handle themselves. You may get your entire planets up and running in fewer turns, once you're up and running - but those planets are spending slots on markets, and producing less metal and research than my planets are.

Yea I think you're getting this backwards; you don't rush-buy science destroyers, you rush-buy LABS, massively more metal-efficient. On research bonus worlds the massed labs are massively(pun intended) more powerful. I usually end up getting around +100 research/turn before the first major war, and then superior metal production is not as important when your superior tech means you can win at 1:2 odds or worse. 100% Farms on fertile worlds, 100% labs on research worlds, 100% mines on on anything with 4 slots or less, 100% factories on everything else. I can't recall ever building a market, or using the science lab system.

-Will