I thought of another way to make a game hard to TAS: Make all actions have ultra-far-reaching consequences that aren't immediately obvious.
Think of a game like go, with a huge board and moves that individually do not alter it much - tactics come together over the course of many moves, and strategy lasts the whole game. The depth of the game comes from there never being an obvious next move - you have to grasp the board intuitively, rather than logically, because there's just too much of a tree of possible moves to search through.
In such a game, having rerecording control of your next move is not as useful as it is in a game where actions have obvious, instant consequences.
I'm pretty sure that a bunch of RPGs are like this.
Do I go to a weapon shop and upgrade? Do I grind for EXP here?
For example, in SF2, I pick up an upgrade in the first 1/4 of the game, that I don't use until right before the final battle, but it is of necessity to finish the game.
Sage advice from a friend of Jim: So put your tinfoil hat back in the closet, open your eyes to the truth, and realize that the government is in fact causing austismal cancer with it's 9/11 fluoride vaccinations of your water supply.
Patashu: The Difference is that in Go (and any other abstract strategy game of its ilk) the actions of the other player are variable; they have the potential to think and formulate strategy just like you do. In a video game, the response to your input is deterministic (*Ahem* Barring the recently touted exceptions that prove the rule...) Thus, console gaming has much less depth in that regard...
Now, having two TASers play a 2-player game competitively, (Each frame would be a "turn" with each revealing their input value to the other at the end)... That might have potential for depth, assuming the game in question itself had deep enough play mechanics.
I'm not convinced that each player should have to reveal input after each frame in a competitive game. It might make more sense to force your opponent to figure out what you are doing. Secrets can make things more fun. Depending on the game, a dragon punch is going to take at least a few frames to pull off, but there may or may not be any immediate visual feedback during the input frames. FWIW, I've heard that there are people out there who can pull off a dragon punch faster than some machines can pick up the input. (I'll believe it when I see it though...)
Toribash?
Most fighting games would be horribly dull in frame by frame.
Turtling and then responding to any commitment with an invincible move would be unbeatable, unless the game supports empty cancels, in which case jabs with a buffered invincible move would be unbeatable.
Maybe it doesn't qualify because of the developer's intent that people try it, but Hero Core seems like it has a great example of this otherwise: The extra difficulty mode is like a boss rush with 8 of the game's bosses on hard mode, except all of them attack simultaneously and you're at level 0 with no powerups. The bosses fill the screen with bullets, making it like a bullet hell shooter but without the tiny hitboxes, and the bosses themselves also fill up much of the screen while moving around erratically. Even under TAS conditions it's extremely difficult just to beat it at all (never mind beating it optimally), despite the gameplay being entirely skill-based (all you have to do is shoot each boss until it's dead).
I've always thought ghost N' goblins for NES is perfect for this. I takes about 5 minutes to beat and every time you beat it it gets harder with no limit that people have seen. I think the most it has ever been beaten is 4 times. So the 5th time may in fact be impossible but no one has ever tried very hard to TAS it and see if it can be done. Proving it's impossible might be hard, but doing something that's never been done before might not be.
"Great spirits have always found violent opposition from mediocrities. The latter cannot understand it when a man does not thoughtlessly submit to hereditary prejudices but honestly and courageously uses his intelligence."
I don't know if there are any actual games that would be impossible to TAS, but I know that you can certainly make impossible challenges if you really want to. I'd like to see someone beat the Final Fantasy VI All Imp, No Equipment, No Magic, No Specials, No Items, Lowest Levels, Fewest Steps Challenge (AINENMNSNILLFSC). No amount of luck manipulation can kill Kefka with those limitations in place...
Say it takes a normal player 4 hours to do a game, you could put a trigger in the game which makes the last boss only appear after 3 hours of gametime, so would normally go unnoticed.
Now imagine a TAS. If gets through the game in 50 mins, well that'd be totally pointless since you'd be forced to wait for that last boss anyway; which I'm sure any real-time run could also achieve by then. Easy way to obsolete the need for TASes.
Although this is an old thread, I like the topic.
zaphod77 wrote:
No, no, no.
The idea is to use a TAS to accomplish what's not supposed to be even possible to accomplish.
To accomplish a goal that is impossible under normal conditions, and very difficult (but still possible) when TASing.
Many Super Mario Hacks are designed to be hard even with savestates and slow mo. These do not count.
I thought of something that would basically be what you're saying awhile ago, what it was was going into Mario Party 3 multiplayer and setting three computer players to Super Hard difficulty and giving them nine stars each and setting the turns to 10, and then winning with all the stars.
Get a bunch of R.O.B.s and spare NES controllers of all kinds. Now "play" Gyromite by manipulating the R.O.B.s into reassembling the spare controllers into increasingly complex machines which themselves can be controlled by the game like R.O.B.s to build even more complex machines. Continue doing so until you have a time machine and finish in negative time.