I'd like to thank everyone who has created or helped create a TAS. You've all provided me with at least a few hundred hours of enjoyment.
After reading all of the replies to this post, I find it very hard to choose even 10 or 15 favorites. There are just so many wonderful runs.
I think all of my favorite runs have already been listed above - my favorites are the same as most others' favorites.
But I'd like to mention two more runs I enjoy that haven't been mentioned yet. I am not sure why I enjoy these videos, but I do: Atlantis no Nazo and Strider (Genesis).
Genisto's Bionic Commando also deserves a special mention. Not only did the TAS itself stun me when it was first released, but it also showed me what TASes were capable of. Prior to that, TASes either looked like ordinary human play (but with no mistakes) or showed off the power of rerecording in an obvious way (like the Gradius playaround or 99 lives in Mario 3). The play in Bionic Commando looked completely inhuman yet also purposeful.
I play Blackjack professionally (counting cards etc.).
Three pro gamblers on one small forum that's not a gambling forum? What are the odds? :)
I'm glad to hear that card counting is still viable. I don't do much myself due to my focus on the sports marketplace and poker, but I do notice more and more continuous shuffling machines lately.
I do'nt know why these people think this is talent...
It is a different sort of talent than the sort of talent required to play quickly through a game without using save states, slow motion, or other tools. Think, for example, of music and mathematics. Some are talented at one, some at the other, some at neither, and a lucky few at both, but both require different skills.
If you would like to see what sort of talent is required, try making one of these videos yourself, preferably without watching the already-existing videos of the same games. Afterwards, be astonished when your time is much slower than those of the movies already published on this site despite the fact that you thought you did everything perfectly and fixed all of your mistakes.
But still, if the creators created a save state point shortly before the dragon, and the dragon went the wrong way, their actions getting back to the dragon on their next attempt would still be identical, frame by frame when trying again right? And since the dragon's randomness is only pseudorandom, wouldn't it behave exactly the same way each time they tried again? Or would the walkthrough makers have to delay a split second in order to changeup the dragon's actions?
Thanks
You are right that identical, frame-by-frame actions will always result in the same pseudorandom behavior.
However, delay is not the only way to produce a different outcome. Pressing different buttons also works in many games. I'm not sure exactly how the pseudorandomness works in Castlevania, though (it differs from game to game).
Some of my favorite artists are Bone Thugs -n- Harmony, Joe Budden, Outkast, and Nas.
Heh. A few minutes ago, I was just listening to "I Am The One" for the first time, and I thought to myself "I wonder if this guy is a Bone Thugs fan".
As for me, I like pretty much every genre of music out there. I listen to plenty of metal, rap, classical, country, electronica, and anything in between.
But lately, I've mostly been listening to so-called "intelligent dance music" (which seems to be mostly be undanceable, and I'm also not sure what's intelligent about it). I've probably listened to more Boards of Canada than any other artist over the last 5 years or so.
LOL Don't hold your breath on this :) Once I find some place to host more files, I'll link to a few more songs I've done if you're interested (or anyone else, for that matter).
Please do. I think I've listened to "As I Write" about 20 times already.
There is a very wide range of talent in this game.
Some people, at age 8, after less than one year of experience, are better than some 40-year-old adults who have been playing their whole lives.
The videos on this site are created by playing the games in slow motion (often one frame at a time).
Many of the moves you see (like climbing the walls quickly) can't be duplicated by a human playing at full speed because they require you to press the buttons and the D-pad many times per second.
JXQ, I am very impressed. I can't seem to stop listening to your song.
About poker: I remember there was some discussion about poker in an old thread, and some others here did play. But not many people play for a living, so I'm always interested when I find another one of us.
I didn't know I'd find someone else on this board that does this for a living. Very cool :)
I used to play only poker for a living, but now I divide my time between poker and sports betting.
I spend a lot of time playing games in general (video games, board games - especially chess - but occasionally other games, like Go).
I studied computer science and finished my degree several years ago, but I've mostly only done programming as a hobby. I've always ended up stumbling upon other ways to make money (like poker and sports), so I have only spent a few months doing professional software development. It's not nearly as much fun for me as doing it as a hobby :)
Other than that, I go for a lot of walks, listen to a lot of music, and sleep a lot.
It is very interesting to read everyone's answers in this thread, even though they're mostly the geeky answers I expected :) Keep it up, everyone.
I recall accidentally stumbling upon the FFL2 power cycle trick many years ago on a family vacation. I thought I'd found something new, but alas, someone else had already published it :)
I believe I was able to use it to get and complete the same battles repeatedly (thus ensuring a power-up after every battle), but I'm not totally sure about that. Give it a try. If it doesn't work (or it's not useful), sorry for wasting your time :)
Edit: Now that I think about it, way back then, I may have just noticed that I could get a power-up after every battle as long as I power-cycle and make the same movement. I certainly remember trying to get repeated power-ups after saving to work, but I'm unsure if I succeeded. I wish I remembered if I did.
Well done. I didn't think I'd enjoy this video as much as the Sonic 2 and 3 videos, but I did. The time spent avoiding the drowning death was well worth it.
I'm also amused that the items and enemies in GH3 are placed in such a way that you can bounce through most of the level like that.
1) chess art? that sounds interesting where can i find that?
You won't find it. It's a bit difficult to explain, but let me give you an example of some "art". Hopefully you are familiar with ICS-style chess servers. (like FICS, ICC, chess.net, etc.)
My program logs in to the chess server and goes in to examine mode. First, it clears the board. Then, it puts a white knight on a8, a black knight on b8, a white knight on c8, and so on, until the whole board is filled with 64 knights of alternating colors. Then, it removes the a8 knight, removes the b8 knight, and so on, until the board is empty again.
It does all of this at lightning speed, so what the viewer sees is a fast-moving wave of alternating color coming from the top of the board down to the bottom. It looks a lot like water waves coming at you on a beach.
A few people thought this "knight wave" had something of a hypnotic effect. They stared at it for 10+ minutes straight.
And now you can see that the reason you won't find this art is that it ceases to exist once the program logs out.
Hopefully my description was clear enough. Anyway, there are a lot of creative things you can do using chess server software.
how about a 2 player TAS? maybe input a classic chess match. chessmasters would appreciate the game itself, others would be amazed at the pieces moving around the board lighting fast?
i dunno, that seems the most promising
I explored this idea about 6 or 7 years ago on the chess servers. I wrote a program that connected to the servers and sent arbitrary input. I had it set up board positions and play moves at lightning speed. It would play through a whole game in a few seconds. I also used it to create some "chess art". I think the chess servers are a better medium for this sort of TAS than Chessmaster/Battle Chess for two reasons:
1) Chess software is quite slow compared to chess servers (you can play out a whole game in as little as 1-2 seconds on a chess server, but each move takes around a second in the video games)
2) Chess software is generally limited to playing out games and some limited board setup features. On chess servers, there are more features to exploit, like a more robust board setup feature, the ability to add commentary, the ability to move back and forth an arbitrary number of moves, the ability to highlight pieces/squares for users of some interfaces, and probably some other things I'm forgetting right now.
The technique of making simplifying assumptions that are likely (but not guaranteed) to hold up and drastically reduce the size of the search space is a common one.
Based on my experience with game-playing computer programs, I'd say that you're right - the first attempts at "solving" video games will only solve segments of the games. Actually, I'd say that Bisqwit has already started along this path with his brilliant Rockman-playing program.
If you had a solution produced by such a program, you couldn't say that you have proven that your solution is the fastest. However, you could say that you've proven your solution is the fastest assuming that all of your |'s are really |'s (which they probably are, but you don't know for sure). And that's good enough for me (most of the time :))
You may be able to prove that some of your |'s are really |'s without a complete brute-force start-to-finish playthrough. You may be able to produce proof that your |'s are really |'s by looking at the code.
Of course, as Boco pointed out, this whole discussion may be nothing more than an interesting theoretical discussion, because the amount of computer power required to solve even small segments is enormous. Many people think that even the game of chess won't ever be solved by computers, and chess only has somewhere between 10^43 and 10^50 legal positions.
I'm optimistic, though, and I really hope to see perfect chess and perfect TASes someday :)
Another problem with the Seal of Quality was that the 5 game limitation was easily circumvented. I once read that Konami created the "Ultra" brand in order to be able to publish more than 5 games per year.
Hmm. I think part of the problem description is that you aren't supposed to refer to previous answers. So references to "the previous largest number in the thread" and such would be no good. That was clever, though :)
I like all of your answers. This forum has some interesting and intelligent people.
The recent discussion reminds me of another interesting problem I saw a few years ago.
Write the largest natural number you can. Any notation is acceptable as long as:
1) It is easily understood (this is subjective and depends on your readership, but I think posters in this thread will be reasonable)
2) The representation of your number (and any function definitions, explanations, etc.) must not exceed a certain number of characters. I don't remember the character limit on the original problem, but let's make it 500 characters.