I really like the idea of situational techs ... I think you should compare average yield of tiles/starting areas, to the advantage of relative situational tech.
For instance, situational techs could be added to certain factions with a less-than-par starting location to better take advantage of their environment.
aka. Surrounded by mountains? Feel some-what limited? You have a good chance of getting Mountaineer tech (mountain movement), Terraces (limited farming/production), Cave Mushroom Farming (extensive farming), and better Windmill techs (extensive production) ....
for some example numbers, a 20% chance for a lower-tiered "mountain" tech for each mountain 1 tile away from city square. Up to max of 80%. a 15% chance for a high-tiered "mountain" tech for each mountain 1-2 tiles away from city, for a max of 75% chance.
Mountaineering could open up varying "Mountain combat" techs, some basic mountain combat being defaulted to open up with acess to mountaneering, although some "super mountain combat" or very specialty snow-skiing troops, ect, would be in the "random" techs, and certain Cave Drillers, or Mountain Climbing elephants, will have a 20-30% chance of showing up in game each, but still only accessible by the situational mountaneering tech.
I think it would be cool to have several nifty branches such as this, which are just an Extra Methyl group to add upon the complex Organic molecule which is the Faction's unique technology tree. Because each faction does have their own tree ... right?
These situational techs could be alot similar to my Environmental type model, only instead of catering the race to live in a certain environment, these situational techs could aid factions with "poor" starting locations, like Icy tundra, barren mountain ranges, vast Deserts, Thick Steamy Jungles ... with advantages in these ecosytems on resource collection and military strategy/tactics.
Who is the best at fighting in the mountain highlands? That crazy tribe that lives off mushrooms in the north. That Incan civilization ... sure did take to terraces soon after they started learning about basic-advanced farming techniques ... maybe someone just started experimenting on all of those mountains?
Man, what kind of armor are those Sand-men wearing? it seems like the Heat and Sand have no effect on them ...
Ice, snow, snow, ice ... It's snowing pretty hard ... our feet keep sinking into new-snow and permafrost ... Whats that large figure in front of us? no wait ... it seems to be an army .. but moving at a normal pace? not slowed down by the snow and ice? what? oh ... lookit Skis! Cross-country skiis?! Its possible to have ski troops! ::gets stabbed and dies::
Jungle ... so much wild-life density, but most civilizations lack the patience or respect to harvest just enough to continue the ecosystem. Techs to attune a civilization to life in the Rainforest will not make them the Industrial or Commercial giants as some nations, nor will they perhaps have enough food to compete with larger nations. Chopping down Rainforest in certain areas may prove cost effective, although being one with the rainforest will provide several advantages. One, you will still have a decent productive rate, just not AS great, Two, your science ability will far exceed your commercial ability, because of all the diverse plants to reasearch (medicines attract scientists globally to your cause, you nation is a research station for an international community). Also, enemy armies will suffer disease and poisons while travelling through the rainforest, and perhaps suffering random losses from encounters with the occasional Caiman or Anocanda (if wading through swamps) ... (wading through a river could be attacked by African River Crocodiles, that would be HEAVY losses indeed, although thats more of a Savanna thing) ... with Knowledge of the jungles ... the Jungle Nation could turtle in the depths of the Rainforest, using their biodeverse dome as a shield. Of course, Situational and Random techs will determine if they can completely successfully turtle there, or if they need a little more ambition to stay ahead. It will not always be smart to settle there, but if you do end up settling there, there will still be hope of at least some good news.
One situational (or random) tech I see for Desert Tribes would be "Cactus Orchard" ... providing a healthy supply of food n water, cactus Orchards provide where farms might not (in the desert), and also provide fresh water (clean rain and ground water) from within the Cactus. If one knows how to grow Cactus in great numbers, they need not concern themselves too highly with looking for Oasis. Im sure there will be some penalty or other for no fresh-water in the desert, wether fixed by magical (to upgrade terrain) or mundane, perhaps the penalty is inherently there.
As to getting horseback-riding early, or some-such ... having "bare-back riding" as a situational tech requiring a nearby abundance of horses might work. This tech could also provide an extra bonus to the added benefit of Saddles (since they can already ride the horses) ... perhaps allowing faster speeds for the horses with saddles, because the people will excell at not falling off their horses, and will know how to handle their weight better as to lessen the tired-ness of the horses.
perhaps that tech or some other situational horse-based tech could increase the chance of a horse-husbandry/fighting tech arriving with the completion of an agriculture-type tech, and perhaps decrease the cost slightly.
On that system, I think that techs of similar tiers should have a certain random chance of arriving as the next tech in the tech line-up, and others having a pre-requisite.
For instance, 3 tier 1 techs are in the agriculture line, and you devote 40% of your tech to agriculture so you get a tech twice as fast as if you had devoted 20%, but different techs have different speeds so your not really sure what you are going to get. And lets say that tier1-techA has a 30% chance, B has a 30% chance, and C has a 40% chance. Perhaps the chances could be determined by situational factors. Anyways, so you end up researching C, which is probably basic field-plowing, or something (while B was herb gathering and A was animal taming), and C alone can lead to techs X and Y, so now you have a 30% chance for A, a 30% chance at B has a 30% chance, X has a 20% chance and Y has a 20% chance. (or if Y was the most natural progression, X could have a 15% or 10% chance, and Y could have a 25% or 30% chance)