It's 10 players vs ~200 bots.

So, SC is still a much more complex space. DoTA has non player bots, but they are similar to SC buildings and follow very simple rules.

You can't compare the complexity of a dota char with a unit. Sc units are very basics, they don't have xp, don't have a skill tree, don't have have 100 of possible objects and their skills don't vary so much in context.

This matters far less than you might think. Go has simpler peaces without movement and it's still much more complex than chess.

This is more of an artifact of the size of the game board than anything, I think. 9x9 go is decidedly simpler/easier than chess, and I expect chess on a 19x19 board with the number of pieces scaled proportionally (each player starting with, say, around 90 pieces) would be a lot more difficult to play/analyze than a standard go game.

Well sure, but that's the exact same issue as you see in SC2 vs DoTA. In SC you are simply dealing with a vastly more complex state.

Not in Dota.

In dota combinations open the road for brand new moves. Some items tp, some regenerates, some cancel buff, some critics, some cleave, some cut trees, some slow, some stun, some give visions, etc.

Now in SC, you have 3, 4 main builds for a given match up. You see the building, and you know where this is going.

In Dota, depending of the 10 heroes, current money and objects combination, and player skills, you may expect one build or another.

Also, a zergling or 10 zergling is pretty much the same the same to consider from the behavior point of view. The number doesn't matter that much, only the intensity of the effect. And a gling will alway do the same thing. Move. Attach. Burrow.

The same unit in Dota can have a completly different role depending of the context.

My guess is that an AI would give you a much bigger advantage in SC because they can make more APM than a human, strat or not, while on Dota at high level strat is more important on the long run.

SC 2 has ~100's of viable strategy's per race as openers, but like chess openers are just that. As the game evolves you get complex iterations and risk reward situations. Bluffing research is very much a part of high level play as min maxing requires you to force the other player to waste resources in any way possible.

For example larva are one of Zerg's most valuable resource and there are several ways of attacking that resource by killing units or simply forcing them to go more defensive.