undefined

points

[-]

Starcraft might be more complex in absolute terms (not sure about that - discrete combinatorial problems can be genuinely more complex can continuous ones, from an algorithm point of view, because solutions are harder to come by)

But chess theory, the human activity of analyzing chess, is hugely more complex than whatever human players have analyzed about the game of starcraft

What I mean is, perhaps the best neural networks that play starcraft are as complex as chess neural networks, and this complexity is irreducible, but starcraft players haven't developed as much theory in comparison

by zemvpferreira4 hours ago|

prev|

[-]

Simple is definitely the wrong term. Chess's simplicity is where its inherent difficulty comes from. It's paradoxically much more difficult to optimize than more complex games like SC.

by summa_tech2 hours ago|

prev|

[-]

Chess presents a comparatively small and, importantly, discrete set of choices at any moment. So it feels like it's a solvable logic puzzle: like it should be possible, at any moment, to make an optimal move. You can predict if you lose 2 or if you lose 3 pieces because of your next move, and you're expected to use this knowledge. The strategy of chess is about perfection in every turn.

RTSes present continuous, large choice spaces. So it doesn't really feel like as much of a logic puzzle, and perfection does not appear to be within ones grasp at every moment. Whether you'll lose 4 or 6 of the T2 fighter-bombers is not relevant. The strategy of RTSes is strategy of big plans and high level abstraction.

by chongli49 minutes ago|

parent|

[-]

The strategy of RTSes is strategy of big plans and high level abstraction

That's not true in all RTSes. Take StarCraft, for example, and there are plenty of games on record that were decided not just by 1 unit, but by 1 attack from 1 unit. There are Zerg players, for example, who have developed a reputation for creating havoc after getting a single zergling (the smallest and cheapest attacking unit) into their opponent's base. A single shot from a Protoss reaver can mean the difference between taking minimal damage and losing half of your workers (and subsequently the game).

by summa_tech3 minutes ago|

parent|

[-]

Sure, StarCraft is kind of a hybrid when you think about it. The guaranteed-hit model, the extremely simplified low vs high ground approach to determine if a shot is possible, etc. are pretty deterministic. But in more complicated situations it's still a lot less predictable than chess. Even the examples you're giving can only happen probabilistically outside very early game.

But I'm thinking about TA-style games, the topic of this discussion, which pride themselves on large armies. Though, to your point, early game of Supreme Commander is also quite chess-like, because of how restricted the set of opportunities is.

by kbelder2 hours ago|

prev|

[-]

Perhaps simpler in the relevant choices presented to the player? The fact that a specific meta can be found, and victory requires using the meta, means that many important choices have been removed from the player.

Chess surely has a meta, but it's been honed so the meta is a huge number of significantly different paths. It's a balancing issue. Give Starcraft another few centuries of play and maybe it'll be the same.

by annasnna4 hours ago|

prev|

[-]

I don’t know but I can imagine many of the levers in games cancel each other out or don’t turn out to be useful while in chess every variable is orthogonal. It’s all important. Complicated versus complex, like how untangling Christmas lights is time-consuming and gnarly but it is not complex fundamentally.

That said, I don’t know if it is true in those cases.