Analysis (referring to above game description):
The original question was whether Legionaries are too tough and/or too cheap. I think the answer is a qualified yes. I say qualified because steps could be taken to minimize the power that Legionaries have to dominate.
One obvious fix would be to make research go faster so that they quickly become obsolete. The problem is that even against musketmen they can put up a good fight, and muketmen cost 60 (30 for Deity AI) while Legionaries cost 10. The Legionaries will win that fight. An extreme version of early obsolescense would be that Legionaries become unbuildable at the advent of Monotherism, replaced by a superior, more defensive unit which costs 50 or 60. If Rome discovers Monotheism when they're supposed to according to history (400AD), that would pour some cold water on the Legionary plans. Of course if Rome has the Great Library, they should discover Monotheism when the Israelites and others did, well before 600 BC.
In this case it's a moot point because in the game itself, nobody even had Currency by 150AD. That's amazingly slow science for a Deity game, much slower than Earth history, even. Of course, this is all to the benefit of Legionaries.
A more simple solution would be to make the Legionaries cost more shield, something like 20.
In any case, they should cost gold to maintain. I can't believe I had an army of 1000 units scattered around the world, and I was paying no maintenance costs. In the previous version I played, where Legionaries did cost 1 gold/turn, I tried to kill them as quickly as possible, but because my empire was so vast, literally hundreds of them were en route their battle at any given time, seriously straining my treasury. But I think that's realistic. Far away wars in ancient days should be more expensive than nearby ones.
About the pre-built wonders: I really like the idea, but you should compensate for the fact that inevitably, the wonders will be concentrated in Europe. That gives a huge advantage to any agressive European power. It's just too tempting and too easy to simply sweep through Europe, one small country at a time. One simple way to reduce the temptation is to give all European cities free walls. Actually, that's a great idea. You might want to go further and designate a special European defender, who costs the same as a spearman and defends like a pikeman. Then the pikeman would be 1.4.1 and every other defensive unit would also get a bump of 1.
Alright, that's to adress the ready conquerability of Europe. Now to address intercontinental campaigns: The Greeks are the gateway to/from Asia. They should be harder to cross. I'd say the Hoplites should defend at 4, and at least some of the cities should begin with a Barracks. (Remember Sparta?) In my game, most of the Hoplites I fought were regulars.
Not to say that we want to stop the Legionaries, and none of these countermeasures would. But they would make things more difficult and more interesting.
On a general note:
Because of the slowness of scientific research, special care must be taken that events in the ancient era don't by themselves decide the course of the game. Unlike in the real world, huge ancient empires don't "rot from the inside" in Civ3. Modest gains should be possible for civilizations that traditionally thrived in the ancient era, like the Romans. But steps have to be taken to make sure those gains really are modest and not something like the conquest of India in the BC times (which is very easy on 1.92/Deity).
So though engineering the march of the Legionaries in the game is historically realistic, realism requires us to also engineer the downfall of the Roman era and the flowering of other cultures. Doing the latter is indeed hard. So, for example, what steps can we take to see to it that the Spanish develop the awesome Armada and dominate the seas for centuries? That's just one example... realistically, Spain in the game is one of those "nothing" countries, which seems unjust. They definitely need some special cheap naval units. I only bring this up because one natural reaction to what I'm saying goes like this: "So you're telling me that Legionaries kick ass... fine! They're supposed to!" True, but many other things are also supposed to happen that can't, because the game has no mechanism to dissolve the Roman empire and let others have their turn.
Of course we can't fix that limitation, but we can take steps to minimize it. One of them is to make sure that all or most the European powers hold their territory into the modern era. That requires strong defenses and able defenders. Especially with its wonders, Europe is such a rich continent that the game ends once it falls under a single power. Any steps to prevent that would surely improve the Tet mod.
Here is how I picture the improved version: Europeans should have very good attack units, but truly excellent defeder units, on par with or better than the Tibetan monk. The strong defenders will prevent any single European state from expanding too much, so their attackers do not pose much of a threat to the world, because as long as European states don't grow they will never produce the attackers in huge numbers. This contrasts with vast countries like India and China, which should have tons of crappy units. Europeans may well conquer stuff outside of Europe, but they won't make much there because of corruption. What's important is that the states in Europe keep their integrity as long as possible. For this they all need super defenders. And this I think is the best answer to the Legionaries question. I'd be happy to see them left alone if every other state in Europe had defenders as good as the Tibetan monks, and cities that come with walls. That would force players to do something more creative than the obvious and boring "roll through Europe" plan. (One thing I just thought of: can you create a new city improvement which cannot be built but may come native in the city, which adds a 50% defensive bonus on top of any other defensive bonuses from walls or whatever? If this is possible, I think every European city should have one - in addition to better European defenders. Oh, and like Temples, the defensive bonus thing should disappear if the city ever changes hands.)
OK, enough on the "European Integrity" stuff. On to some more specific observations about gameplay in 1.92/Deity.
1. The AI is certainly not shy about declaring wars, something which rarely happened in 1.8/Monarch. It seems that they have many units which they are eager to burn off. Unfortunately, they use far too few of these units to protect their cities. In my game, the Russians had about 30 spearmen running through the forest when the last Russian city was destroyed... with only 3 spearmen protecting it. And it's not like the 30 in the forest didn't have enough time to get into a city. Well, maybe that's an AI limitation you can't fix. Anyway, because of the way the AI fights wars, they usually emerge in a very weakened state because they send to many units forward, all without concentrating their attacks in a specific place. (There are exceptions to this, like the glorious Songhai suicide run. But that happens too rarely. Typically, they send 8 attackers against a city every turn. If they waited 4 turns and sent 32, may have had a chance to break through.)
I know that this may not be mod-fixable, but it's important for overall gameplay because most of the civilizations you face quickly become mere shells of their former selves because just about all of them have gone through very costly wars. In my game, Tibet basically destroyed Persia and India's military power, to the point where you could almost walk in and take over an Indian city. Yet the Tibetans actually took very few cities because they didn't want to commit enough troops to defeat three spearmen. All that makes it too easy for the human player. I guess that's why we should play Deity.
I think I like the fact that we can't build settlers. That was one of my complaints in 1.8x - that the AI spends too many shields on settlers and not enough on defending itself. That problem has been fixed. However, some of the civilizations, especially in Africa and the Americas have as their greatest strength the capacity to expand. To compensate them, I think there should be lone barbarian settlers scattered in the uncharted lands. And I'm not talking about the way we have it now, with a stack of 5 settlers guarded by forces that can't be defeated before the industrial age. I'm serious. It pissed me off that a maxed-out army of legionaries, the world's strongest attackers, could not defeat the weakest of the barbarians, a 13 hit-point Hun. The army had 90 hit points - many of its members were elites. Now, that's just silly. I'd much prefer smaller prizes, like a single settler, for a fight that a full ancient army could actually win. Of course, the ones with the first shot would be the states who are the closest, which is as it should be. Europeans a navy or powerful expeditionary forces could also get in on it. I think this is a great idea, but the present barbarians are stupidly tough.
Somehow Arabia found a cache of attainable settlers and made a bunch of cities down the east coast of Africa, and the Vikings must have defeated some barbarians and founded Reykjavik and Helsinki. Other than that, by 150AD, no cities were built, and many were destroyed. I do understand the sense behind waiting for Astronomy before allowing colonizers, but that's little consolation if the game is long over by then. It seems to me that Astronomy should just come sooner, but also that armed expeditions into the badlands should leave a decent chance for new settlements long before the age of astronomy.
Another thing that's new in 1.9x is the pre-distribution of wonders in their historical locations. Unfortunately, the great majority of those wonders are in Europe or around the Medeterranean, including all of the crucial "happiness" wonders. That makes it too easy for an ambitious power to settle the game early on, simply by conquering the few pushover states in Europe. I can't emphasize enough why new versions of the mod must make this very hard to do, forcing the player to try something more interesting and creative, possibly involving sciences discovered after the ancient era. Hey, this is a mod after all, and the idea of special rules for Europe should be taken seriously. After all, history goes that way as well.
Because there are no wonders to build, constantly staying on the cutting edge in science becomes less important. Also, great leaders become far less important. I got at least 10 in my BC rampages. Most of them made armies. Unfortunately, the AI didn't know how to handle the new "fightable" leaders at all. In fact, though I saw several great AI leaders, all of them were used for city defense. There was an elite great leader guarding Berlin and Baghdad, for example. Now, given it's the ancient era, these leaders do a decent job as guards - but a simple army can destroy them every time. The AI was not shy about building armies in the 1.5 Tet mod (from the PTW cd), and that added some spice to the game. (I personally destroyed at least 7 full Chinese armies). The biggest problem in not cashing in the great leader for an army is that no other great leaders can be created. That's just poor strategy for the AI. I have a feeling that the problem began when the leaders got hit points. Oh, and I also got the problem of leaders created while a city is conquered, and not having the city turn into my property. This is no big deal, as you can simply select another unit, walk it into the city, and then the changeover happens.
One last thing: It seems that as version numbers of the Tet mod get higher, so does the cost of city improvements. I imagine that the idea behind it was to make the ancient era less eventful, since everyone spent a long time waiting for their temples and barracks to get built. The practical upshot of this is that rush-building becomes an indispensable strategy. That may not be bad, but we know the AI is terrible at rush-building, so in effect it's another unfair advantage to the human. It doesn't help that one citizen seems to be worth 120 shields now. The official game patches are tending in the direction of lowering the rewards for pop rushing. Maybe that's lazy of Firaxis, and the better solution would be an AI that knows how to effectively rush-build. But they didn't give us that, and the Tet mod does too much to highlight this AI weakness. I have to think that the Firaxis decision to reserve the use of rush-building for emergencies might be the wise one. If you had to sit and dilligently wait for a barracks to be built, that really would slow down the military buildup and make the ancient era less eventful (which needs to happen).
In general, this mod plays too much to AI weaknesses and not enough to its strengths. One of its strengths is quick and logical expansion, but that's out when settlers got removed. The new Tet mods basically require rampant militarism and rushing, two things the AI is terribly weak at. As a result, an average player like me can easily beat the world with ancient-era units on the highest difficulty setting. That's a bad sign. Unlike with the Europe thing, I don't have any specific recommendations about how to fix this, but I want to flag it as a worry I have.
Alright, that's enough of me talking about my worries. The reason why I dwell on them so much is because I really love the Tet mod and the inspiration behind it. If it didn't exist, Civ3 would have long been deleted from my computer. Tet's noble goal is to do a good job lining up the game and world history. Much of it works, and I want more of it to work. I can't promise to test future versions because it takes sooo long, but since I invested all this time into "testing", I thought I should write up some observations for the benefit of those who come after.