Does he want to do speedruns, or TASes? While both require a deep understanding of the game in question, speedruns require more gaming skill, while TASes require more planning and understanding of how computers and programs work. Speedruns are basically playing the game normally and trying to do the best you possibly can; TASes take a much more systematic approach with a lot of trial and error, "playing" one frame at a time and trying every plausible method of tackling each obstacle in the game so that the best method is always used.
Best of luck to you. I neither TAS nor speedrun, as I have entirely too many hobbies already.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Your second theory sounds more plausible to me than the first, seeing as the PS2 (which uses a modified PSX controller, right?) has dual analog sticks and I don't think anyone's ever really had trouble with it for at least simple things like camera control.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
From my understanding of the DMCA (and I am not a lawyer), while YouTube is required to take the content down on receipt of the DMCA takedown notice, they are also required to put it back up if you contest that notice. After that it's between you and the sender of the notice to determine who's on the right side, and YouTube's only further involvement is to take the thing down again if you proved to be in the wrong.
Of course, it's unlikely that any individual is going to want to go up against a bench of corporate lawyers.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
I watched the first few minutes, and there is waaaay too high a plot to gameplay ratio, especially considering that the plot itself is incomprehensible to me. Voting meh for game choice.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Ah, the old item-get delay. Ehh. I'm sure there will be a 100% run that will be able to show off the Ice Spreader et al, so no worries about missing out on them for the any% run. Otherwise it sounds mostly like there are some things you'd want to do to avoid cutscenes, and the Omega Pirate, neither of which I think are huge dealbreakers, so I'd still say go for realtime.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
The problem I have with these videos is that they're so ridiculously hard that even the most skilled of players resorts to bombspam and dies shortly after running out of those bombs. It makes it difficult to appreciate the boss's bullet patterns because they keep getting wiped clear by explosions. It'd also be nice to see how the patterns are "meant" to be solved -- surely there is a way to pass the pattern without ever using a bomb, yes?
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Not being a part of the Prime speedrunning scene, I'd say start from a fresh file. My eye's more on the entertainment factor than on the timing factor, and IMO leaving the cutscenes in would, for Prime, make a more entertaining movie. Maybe not for Prime 3, which drags on rather longer...
I agree with you on version choice. As for realtime vs. gametime, generally I favor realtime but I don't know what the tradeoffs would be in this case. Could you give some more details?
Also, I'm looking forward to seeing this!
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
That sounds to me like solving the hard AI problem, personally. :) Teaching a computer to understand source code at anything above a syntactical level would be a monumental task.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
You'd need to do something like a full walkabout of the TV + console + device to make it clear that there are no other video inputs to the TV than the console. Even then you could be running a custom computer installed inside the NES itself (I've seen an NES case used as a small-form-factor computer case before) or just doing some post-processing trickery to make the game show up on the TV screen with appropriate visual effects.
In short, haters gonna hate.
Well, I did try Super Mario Bros 3. I discovered the console polls the controller twice a frame in that game, so I first had to adjust for that. I still had bad luck trying to get it to run. Sometimes it wouldn't even make it into the first level, sometimes it would make it halfway through the first level, and sometimes it failed somewhere in between. I'm not entirely sure what the problem is, I may give it a try again soon.
I tried a few other games like Marble Madness, and didn't have too much luck with them either. Are there other games that are as well emulated or lacking a RNG like Super Mario Bros that I might have better luck with? I only went with SMB because people were talking about it being the most likely candidate for this to work.
There's puzzle games like Adventures of Lolo, but the semi-turn-based nature of those games means that they're less visually impressive (being basically a highspeed walkthrough of the puzzles).
EDIT: holy crud! Why's my text so big in the first section?
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
He describes some of the limitations of this approach in the article: if the game has any dependence on uninitialized memory states (e.g. as a source of randomness) or on complex timing, then the TAS probably won't sync.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
You shouldn't need step 1 since you can iterate the process. It's more like
Clear all memory search
Damage the boss
Search for values that have decreased
Damage the boss
Search the results of the prior search for values that have decreased again
Iterate until you only have one result
Thus while your position might decrease in some of those searches, since it doesn't decrease in all of them it'll get eliminated at some point. The difficulty comes in guessing the number of bytes allocated to the boss's health and determining when you have actually damaged it.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
You should try to find the boss HP in memory to verify that. With other Forces I'd agree with your assessment, but for the Cyclone Force I wouldn't be surprised if Capsule Laser, or just battering the boss with the force while firing hyper shots, would be stronger.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
A few ad-hoc tests shows that Google won't return results much bigger than around 1.7 * 10^308, or about 2^1024. I guess they aren't willing to dedicate more than a kibibit to numeric accuracy.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
I could definitely see that working. A nice, controlled environment should keep the search path fairly tidy. Basically you'd be coding an AI for the game that can abuse savestates for maximum carnage. Good luck!
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
I think we'd certainly find it very interesting, but any remotely successful TAS bot is going to be incredibly game-specific, and I'm not just talking about things like memory addresses.
But hey, feel free to prove me wrong. :)
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
In general, number of credits is a concern for any game where adding credits extends a play timer that can also be extended by taking actions in-game (e.g. Gauntlet). There, extra credits would allow you to save time by not having to detour to pick up health to keep the timer from running out, but that could easily lend itself to simply ignoring all damage sources since you can buy them out without any time sacrifice.
Not that I'd expect a Gauntlet TAS to be submitted here since the game's endless, but Magic Sword operates on similar principles and does have an end. I'm not certain in that case if you can die just from not adding money for long enough, but I'm pretty sure you can get your health low enough to have the low-health beep going continuously, which would itself be pretty poor entertainment-wise.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.