Post subject: TASVideos potential participation at AGDQ 2015
Moderator, Senior Ambassador, Experienced player (906)
Joined: 9/14/2008
Posts: 1014
Edit: We made it in! See the Preparation for AGDQ 2015 thread for more details. I was proud to represent TASVideos at AGDQ 2014 and a number of people both here at TASVideos and elsewhere have asked me if I would put together a TAS submission for AGDQ 2015. After discussing this with adelikat, Nach, and most importantly my wife (asassyknitter), I've agreed to organize a pitch and (if we are accepted) organize the next steps, including representing TASVideos at the marathon pending financial viability. The AGDQ 2015 submissions are open until September 3rd and we can submit up to 5 games total with a maximum of 400 characters per game with video links highly encouraged. Submissions cannot be edited so the submission has to be perfect. By September 1st I would like to write and submit a single consolidated pitch representing the best that the TASVideos community has to offer. To that end, I would like to get some feedback on ideas people have for things we can do. Here are the ideas I'm aware of so far: - We could, well, play back TAS's we know sync and talk over them like I attempted to do for Gradius. Pros: It allows games that are not marathon safe to be shown in a marathon setting and can demonstrate the usefulness of TAS tools. There is an ever-growing list of console-verified games to choose from. Cons: I discovered about 45 seconds into my commentary for Gradius that I didn't know what else to say on camera during a ~10 minute run of complete perfection (less last-minute scrambling and more rehearsing on my part could have helped there). Let's be honest, it's just not as interesting to watch a run that you know will not fail unless there's a technical issue, although good commentary and appropriate setup can help make up for this deficiency. - TheAxeMan has created a nifty script that allows the Final Fantasy RNG to be deduced based on the outcome of the first battle. Combining this with a dynamic input script would allow us to replay this otherwise challenging to console verify game on a real console. Pros: It perfectly demonstrates exactly how well games have been disassembled and what TAS tools can do. It has a "Wait, you did *what*?" factor going for it. Cons: The run will be 70 minutes long and is a Final Fantasy RPG, which has not been well received in this marathon so far (see starting on page 7 of the AGDQ 2015 submissions thread for plenty of salt about the status of FFT). We would probably have to do this concurrently with something else, i.e. start this up, allow it to run in the background or in a small corner window, then bring it back up for the final boss battle, in which case the white mage run may be more appropriate). - Weatherton can modify the layout of Mario Kart 64's Choco Mountain to allow for a four-way human race in a modified landscape. I first heard about this idea from micro500 and I think this could be extremely interesting. I suspect that a TAS would set up the level so quickly that the participants in the race wouldn't actually know what to expect and their normal strategies would be useless leading to a very humorous race. Pros: It recreates the "Wait, a bot programmed the game and now real people are playing it!" feeling from AGDQ 2014. Cons: We did Mario Kart 64 last time and it might be seen as too much of a repeat. Shaming talented speedrunners by forcing them to run into rocks might be seen as abusive so we'll have to ensure runners with the right attitude and expectations are participating. - I came up with an idea for a TASer vs. human race (Edit: We tried it out and it was a lot of fun!). Two TAS'ers will be given a set period of time (say, 4 hours) to create a TAS of a game and compete against two realtime runners in a four-way race. The idea is that the realtime runners know what the game is ahead of time and can practice, while the TAS'ers have to speed TAS (similar to speed chess, where perfection is not possible and sacrifices have to be made). At the end of the set timeperiod, the two TAS'ers submit their movie files and all four runs are started at the same time in a race. Tompa has volunteered to be one of the TAS'ers and I have a potential volunteer for a realtime speedrunner. I also have a potential game in mind which I will need to privately test out (to ensure the TAS'ers stay in the dark). Pros: If the game choice is done well it could be extremely entertaining, and whether the humans or the TASers come out ahead there's still some tension until the race completes. The TAS'ers need not be present at the marathon, although it would be a benefit if they were. We'd get to show off TWO replay boards at once. Cons: It's complicated, and if there's one thing the technical crew at AGDQ hates it's complicated ideas. I think with the right pitch and a demo video these potential objections could be overcome but it's definitely a risk. The wrong game choice would completely kill this; it might be difficult to find a run that can be TAS'ed in x hours and be close in duration to the actual human RTA time. I have to figure out how to pitch it in 400 characters. - There have been ruminations of someone creating a new Arbitrary Code Exploit (ACE) run with an even more elaborate payload than what was shown in SMW at AGDQ 2014. It'd be hard to top Masterjun's achievement but it's not impossible. Pros: The arbitrary code execution was the high point of the marathon and was the subject of some really awesome articles - the Ars Technica coverage was my favorite. If a previously unknown exploit is exposed it could recreate the same wow factor. Cons: Masterjun has indicated that he would prefer to see someone else create the payload this year. At a certain point, we just can't keep making the payload bigger than the previous year (although the SNES Demoscene folk seem to have some good ideas). It just feels like the problem that M. Night Shyamalan has where his only trick is widely known and his popularity keeps dropping because the audience goes in knowing there will be a plot twist. Even with that pessimism I'm actually fairly optimistic that we can keep this arbitrary code execution fresh if we execute and present it correctly. Edit: Please keep in mind that the primary purpose of the Games Done Quick charity marathons is to raise money for good causes; if there are any good donation incentives please feel free to bring them up! Those are the ideas so far. I'd love to hear people's suggestions and I'll add to this list as the discussion unfolds. One other thing: I should say that I have honestly been humbled by the support I was given for AGDQ 2014 and I am extremely appreciative that there has been so much support to have me represent TASVideos again this year. I am treating this as a personal opportunity be the project manager and presenter again and I will do my best to fulfill that role. Thanks again for your support and I look forward to a lively discussion of ideas, A.C. ******
I was laid off in May 2023 and became too ill to work this year and could use support via Patreon or onetime donations as work on TASBot Re: and TASBot HD is stalled. I'm dwangoAC, TASVideos Senior Ambassador and BDFL of the TASBot community; when healthy, I post TAS content on YouTube.com/dwangoAC based on livestreams from Twitch.tv/dwangoAC.
Site Admin, Skilled player (1250)
Joined: 4/17/2010
Posts: 11473
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
I love the Mario Kart layout idea the most. It's revolutionary, and absurd, and funny. It's impudent, hence if implemented right it's going to be a lot of fun. The ACE one will only be good if "we" program something actually new. With all the hype the previous one generated, it only was putting plain and simple games into. A demoscene piece might work now, as we can still keep both ACE and human play present, just separated between the 2 runs. My old hope is also to record human input and replay it, but I dunno how to make a fun presentation out of it.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Player (146)
Joined: 7/16/2009
Posts: 686
I'm liking the TAS vs. Human idea a lot. Sure, it doesn't show to the fullest extent what TASing is about (the painstakingly picking games apart part), but it could be neat to see how TASes make up for their lack of knowledge by using not-run-safe strats. I'd imagine it could be an issue to get a good game though, as it'd have to be one that's reasonably TASable within the time frame and it'd have to be very sync stable. Still, just dropping the game name into the TASvideos IRC and seeing what kind of run comes out within 4 hours would be neat.
Experienced player (608)
Joined: 10/23/2004
Posts: 706
DwangoAC, thanks for initiating a formal submission for TASvideos and just overall being a cool guy. I'm looking forward to seeing you again at AGDQ. I thought it would be fun to give some more background on the Mario Kart Choco Mountain idea to give a feel for what the idea is and how people may react. At an annual Mario Kart 64 / Goldeneye meet that I host one of the early round matches was spiced up a bit when someone happened to hit the first boulder on its first bounce with a star. This resulted in the boulder bouncing over the lake it normally falls into and then bouncing around the level for the rest of the level: http://www.twitch.tv/theelitegaming/c/4781939 The reaction was great and many long-time Mario Kart 64 players were surprised by what happened. As it turns out, this is something that Myles Bukrim had demonstrated in a 2009 Youtube video: Link to video The day after the "Boulder Incident" that everyone loved so much, I began TASing the level to see if I could get other boulders to bounce around the level. As it turns out, there are at most three boulders that can spawn and they each spawn from a unique location. The first boulder is easy to get to bounce around the track. The other two require some significant acrobatics involving 2-4 players playing "catch" with the boulders to get their angles right (since only the terrain influences their angle). This process turns out to be difficult but possible. The result is that I can repositition all three boulders throughout the level to be bounding repeatedly and unpredictably on course sections that normally do not have boulders. Using a four-player games, I can reposition these boulders (while throwing bananas and shells for good measure) and then place all four players at the starting line to establish an unpredicable and challenging scenario for people familiar with the game. For reference, here is a picture of two re-positioned boulders:
Current Project: - Mario Kart 64
Experienced player (608)
Joined: 10/23/2004
Posts: 706
I'm not able to edit my post for some reason so here is the fixed Twitch link: http://www.twitch.tv/theelitegaming/c/4781939
Current Project: - Mario Kart 64
Emulator Coder, Player (69)
Joined: 10/4/2005
Posts: 197
I really like the Mario Kart idea, but as dwangoAC said we'll need runners who have the right attitude. However the twitch video Weatherton linked made it look really fun, so I'm less worried about that. Weatherton, can you start talking to some of the Mario Kart runners and see who will be at AGDQ and who would be interesting in participating if we decide to do this? The human vs. TAS idea is also interesting. The problem will be finding a suitable game that is interesting enough, and is sync stable to allow us to do this. I think the tension of runners racing the bot could be entertaining. Since the TASes won't be perfect the human runners might stand a chance. Would this be a game that the human runners already know and have practiced, or would this be a game that is new to everyone?
Post subject: Re: TASVideos potential participation at AGDQ 2015
Editor, Expert player (2478)
Joined: 4/8/2005
Posts: 1573
Location: Gone for a year, just for varietyyyyyyyyy!!
dwangoAC wrote:
I came up with an idea for a TASer vs. human race.
This concept can be easily tested by choosing a game that you are not familiar with from the list of SDA runs that have no TAS. Just put a timer on for 4 hours and speed-TAS the game. Then compare the result to the existing SDA run. By the way, thanks for the nice thread. Also, this concept will be entertaining, if the TASer is shown and he constantly talks about what he is doing while making the run, essentially revealing the whole thought process.
Joined: 7/2/2007
Posts: 3960
I'm inclined to say that we should mostly try to avoid weird gimmicks. If the focus is on TASing, then let's show TASes. Here on these forums we are of course extremely familiar with TASing, which tends to mean that we're more interested in the more outré examples of the art, but in a setting like AGDQ I think we'd be better-served by simply showing a set of very high-quality TASes. The documentary thread has a bunch of discussion of what would make a good "TASVideos trailer", and I think much of that should be adaptable to an AGDQ TAS block. In general, I'd say that we should prioritize: * Runs of games that many people are familiar with and that are played in highly-unexpected ways. The SMW ACE run was an excellent example of this at AGDQ 2014. * Runs of games where the TASer (or an experienced speedrunner) can provide commentary, ideally in-person but with a voiceover or prerecorded speech if necessary. Commentary goes a long way to explaining the sometimes odd-looking decisions made in games. * Runs that stand on their own well. International Super Star Soccer Deluxe might be a good example here. But really, just run down the starred runs and select ones you like and/or have the cart for. I do think it could be interesting to show a speed-TAS vs. speedrunner competition; if nothing else it should help demonstrate how much effort a TAS takes, since I doubt that 4 hours would be enough time to make an especially optimized TAS.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Editor, Player (44)
Joined: 7/11/2010
Posts: 1029
If you're going to use prerecorded speech, find a speech synthesis program and get TASbot to do the commentary itself. I'm sure that would amuse the crowd. I'm worried that the Mario Kart 64 idea might not work so well for people who are unaware of the game. For instance, I've played a certain amount of Mario Kart (including 64, and including Choco Mountain), but I wouldn't be able to tell you where the boulders were "meant" to be at all reliably. It does cross my mind, though, that arbitrary code execution glitches could be used to modify pretty much any game that has them in this sort of way. You could "easily" add a TSR that rewrites levels in certain ways now and again. All the better if you have someone who's familiar with the original game, and maybe someone else who's familiar with the modified game and can show off a speedrun that's only possible due to TASing. That said, ESA has a "that's never happened before" counter, don't they? Making things happen that have never happened before happens all the time by accident by human skill in marathons; doing it intentionally via emulator tricks would have to be particularly spectacular to stand out. That's probably within our ability, but may take a while to plan. I'm against the "make a TAS live" method for philosophical reasons; I think that TASes are best when you have a huge, deep understanding of a game, full debug tools, and so on. With many games, there shouldn't be a need to touch the controls when making a TAS because everything is done at a higher level (e.g. look at any of MrWint's Pokémon TASes). However, I'm aware I'm in the minority on this, so it might be worthwhile disregarding my opinion. Just "here's replaying a TAS on console" is likely to be good enough, though, if it's a high-quality TAS. We should probably show off a range of things that TASes can accomplish: humanly impossible glitches; human-looking play that's to a much higher standard than anyone could manage; and total control of a game. That worked pretty well last year, and I see no real reason not to just do more of the same.
Editor, Skilled player (1535)
Joined: 7/9/2010
Posts: 1319
ais523 wrote:
I'm against the "make a TAS live" method for philosophical reasons; I think that TASes are best when you have a huge, deep understanding of a game, full debug tools, and so on. With many games, there shouldn't be a need to touch the controls when making a TAS because everything is done at a higher level (e.g. look at any of MrWint's Pokémon TASes). However, I'm aware I'm in the minority on this, so it might be worthwhile disregarding my opinion.
I am sure they're about to show some basic TASing techniques (using savestates, frame-advance), without getting detailed in finding glitches, game mechanics or debugging a game. There are lot of users on Twitch, where AGDQ is usually streamed, who don't have a clue how TASes are made. Also there will be many new viewers watching the stream, that never even watched a TAS, or in some cases a speedrun. Of course it's needed to say that on tasvides we generaly put more work into TASes than this, and mention other methods of creating TASes exist (Botting, TAS Editor). I'd say that most viewers would appreciate a simple, small live TASing.
Favorite animal: STOCK Gt(ROSA)26Sortm1.1(rtTA,EGFP)Nagy Grm7Tg(SMN2)89Ahmb Smn1tm1Msd Tg(SMN2*delta7)4299Ahmb Tg(tetO-SMN2,-luc)#aAhmb/J YouTube Twitch
Post subject: Thanks for the good feedback!
Moderator, Senior Ambassador, Experienced player (906)
Joined: 9/14/2008
Posts: 1014
I appreciate everyone chiming in. It seems like there are supporters for just about every idea I proposed (and a few detractors - fair enough). We still need someone other than Masterjun who wants to tackle a total control run (using ais523's definition where ACE is the Arbitrary Code Exploit and total control is the fun payload you program after you've executed said exploit). Does anyone have suggestions for good game choices to play back?
I was laid off in May 2023 and became too ill to work this year and could use support via Patreon or onetime donations as work on TASBot Re: and TASBot HD is stalled. I'm dwangoAC, TASVideos Senior Ambassador and BDFL of the TASBot community; when healthy, I post TAS content on YouTube.com/dwangoAC based on livestreams from Twitch.tv/dwangoAC.
Spikestuff
They/Them
Editor, Publisher, Expert player (2631)
Joined: 10/12/2011
Posts: 6435
Location: The land down under.
I want going to nominate: [2406] NES Kirby's Adventure "game end glitch" by MESHUGGAH, CoolKirby, Masterjun, MUGG, TASeditor & illayaya in 00:35.91 This is a good example of a FRAME WAR. But there was an issue to Console Verifying this but the last post was MESHUGGAH's. He provided a fix for the issue but I don't see any checkups with that file.
WebNations/Sabih wrote:
+fsvgm777 never censoring anything.
Disables Comments and Ratings for the YouTube account. Something better for yourself and also others.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
One idea for a total control hack would be something that seems at first glance impossible to do on the target console. I'm thinking of how demo programmers time and again seem to defy what seems possible with limited hardware in those old consoles. Or, perhaps alternatively, show some picture that seems to be of much higher quality than what the console is capable of. (There are tricks that can be used eg. on the NES to make an image look like it has much richer color palette than usual. I'm sure on the SNES this could be taken much further.) Of course I understand that making such a demo, game or even image would be a lot of work, so the suggestion may well be unrealistic in practice.
Editor, Experienced player (570)
Joined: 11/8/2010
Posts: 4035
Spikestuff wrote:
But there was an issue to Console Verifying this but the last post was MESHUGGAH's. He provided a fix for the issue but I don't see any checkups with that file.
I asked true later if it synced with his newest device and he said it desyncs on console due to less than accurate emulation. Unfortunately, it may be a while before we can verify it. I would really like to see the TASer vs. Human idea. Even watching the TASer make the TAS for 4 hours and explain the process would be interesting. As others have noted, it would also show millions of people how a TAS is made.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
CoolKirby wrote:
Even watching the TASer make the TAS for 4 hours and explain the process would be interesting. As others have noted, it would also show millions of people how a TAS is made.
Given how much repetition is involved in making a TAS, wouldn't it get pretty boring pretty fast?
Site Admin, Skilled player (1250)
Joined: 4/17/2010
Posts: 11473
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
Warp wrote:
One idea for a total control hack would be something that seems at first glance impossible to do on the target console. I'm thinking of how demo programmers time and again seem to defy what seems possible with limited hardware in those old consoles. Or, perhaps alternatively, show some picture that seems to be of much higher quality than what the console is capable of. (There are tricks that can be used eg. on the NES to make an image look like it has much richer color palette than usual. I'm sure on the SNES this could be taken much further.) Of course I understand that making such a demo, game or even image would be a lot of work, so the suggestion may well be unrealistic in practice.
http://wiki.nesdev.com/w/index.php/Full_palette_demo Though, doesn't look like we have an NES game that would allow such deep total control.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Moderator, Senior Ambassador, Experienced player (906)
Joined: 9/14/2008
Posts: 1014
Warp wrote:
Given how much repetition is involved in making a TAS, wouldn't it get pretty boring pretty fast?
The four-hour TAS'ing time would be off-camera to allow the competing TAS'ers to fully concentrate on their work. There would be an initial announcement followed by the TAS'ers going to their respective corners to independently make the best TAS they can. At the end of the block of games the TAS'ers would play back their (hopefully finished) runs competing against each other and two other human players on the live stream. Oddly, the realtime speedrunners I've spoken to are extremely confident that they will come out on top, so it'll be interesting to see what the outcome is. I think there's enough interest in this idea that we should try a dry run. Can I get two volunteers? I'll pick a two-player race of appropriate length from a previous marathon which is on an easily TASable platform and the two volunteers will be given a set amount of time to complete a TAS of the game. The TAS'ers can reference any and all material they like but the TAS must be new material. After the time limit expires we'll post the results of the race and hopefully someone can do some kind of comparison encode. If the two volunteers are not available at the same time to do this "live" I think we can do the dry run on the honor system, i.e. allow the two runners to work on their own schedule but with a timer running. The purpose here is to ensure the concept is viable so that should be fine for now. So, any volunteers for this? Thanks for everyone's support!
I was laid off in May 2023 and became too ill to work this year and could use support via Patreon or onetime donations as work on TASBot Re: and TASBot HD is stalled. I'm dwangoAC, TASVideos Senior Ambassador and BDFL of the TASBot community; when healthy, I post TAS content on YouTube.com/dwangoAC based on livestreams from Twitch.tv/dwangoAC.
Eszik
He/Him
Joined: 2/9/2014
Posts: 163
feos wrote:
My old hope is also to record human input and replay it, but I dunno how to make a fun presentation out of it.
We might replay the SMB WR‚ as it was played on FCEUX. It's not too long and it would be a kind of joke, like "SMB WR was tied at AGDQ!"
I problably made mistakes, sorry for my bad English, I'm French :v
Experienced player (608)
Joined: 10/23/2004
Posts: 706
dwangoAC, would the TASers know which console the game will be on first? Adjusting to new tools takes some time.
Current Project: - Mario Kart 64
Moderator, Senior Ambassador, Experienced player (906)
Joined: 9/14/2008
Posts: 1014
Weatherton wrote:
dwangoAC, would the TASers know which console the game will be on first? Adjusting to new tools takes some time.
I'm almost positive that we'll be using NES as there are multiple emulators that support it. I don't see any obscure or hard to TAS platforms being in the running for exactly the reasons you mentioned.
I was laid off in May 2023 and became too ill to work this year and could use support via Patreon or onetime donations as work on TASBot Re: and TASBot HD is stalled. I'm dwangoAC, TASVideos Senior Ambassador and BDFL of the TASBot community; when healthy, I post TAS content on YouTube.com/dwangoAC based on livestreams from Twitch.tv/dwangoAC.
Spikestuff
They/Them
Editor, Publisher, Expert player (2631)
Joined: 10/12/2011
Posts: 6435
Location: The land down under.
Eszik wrote:
feos wrote:
My old hope is also to record human input and replay it, but I dunno how to make a fun presentation out of it.
We might replay the SMB WR‚ as it was played on FCEUX. It's not too long and it would be a kind of joke, like "SMB WR was tied at AGDQ!"
Seconding this. Link to video http://dehacked.2y.net/microstorage.php/info/493401020/smb-114.fm2
WebNations/Sabih wrote:
+fsvgm777 never censoring anything.
Disables Comments and Ratings for the YouTube account. Something better for yourself and also others.
Editor, Player (69)
Joined: 1/18/2008
Posts: 663
Eszik wrote:
feos wrote:
My old hope is also to record human input and replay it, but I dunno how to make a fun presentation out of it.
We might replay the SMB WR‚ as it was played on FCEUX. It's not too long and it would be a kind of joke, like "SMB WR was tied at AGDQ!"
If any race I think this would be it - 3v1 to see who gets closest to ideal :)
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Post subject: Less than a week left!
Moderator, Senior Ambassador, Experienced player (906)
Joined: 9/14/2008
Posts: 1014
This is just a friendly reminder that there is less than a week left before I need to submit the TASVideos response for AGDQ 2015. It's time to start thinking about donation incentives - if I submit the 5 ideas we have so far, are there any good incentives people can donate toward? Last time out of the ~$12K we raised during the TAS block less than $500 came from donation incentives, which is something I'd like to improve on this time around. As always, the Games Done Quick marathons are for charity so anything we can do to help raise money for a good cause would be awesome. Thanks for those who have contacted me in IRC about game suggestions. Keep em' coming, there's gotta be others out there we haven't thought of yet. Thanks!
I was laid off in May 2023 and became too ill to work this year and could use support via Patreon or onetime donations as work on TASBot Re: and TASBot HD is stalled. I'm dwangoAC, TASVideos Senior Ambassador and BDFL of the TASBot community; when healthy, I post TAS content on YouTube.com/dwangoAC based on livestreams from Twitch.tv/dwangoAC.
Experienced player (608)
Joined: 10/23/2004
Posts: 706
DwangoAC and micro500 requested that I create a demo of the Choco Mountain Gauntlet concept. Here it is: Link to video
Current Project: - Mario Kart 64
Joined: 8/28/2014
Posts: 7
While the Choco Mountain Gauntlet concept is technically interesting, I'm not sure how well it might be received due to the track not really changing all that much. Having a few boulders and bananas around doesn't seem like a dramatic difference from a player's perspective. A 70-minute FF run would definitely be too long to focus on, but I like the idea of letting it run unattended. About the human vs TAS race... I have misgivings. I think races are difficult to plan because you need them to be close in order to be most interesting. If either party pulls far ahead, it's not very compelling. That's one reason there are so few of them at each event. You need people who are guaranteed to be similar in performance. With all the variables involved here, we can't make that guarantee. In my opinion, any ACE will be cool, even if it doesn't top the previous year's in terms of complexity. I might be disappointed if there wasn't one. As for playing back recordings of runs, I like those, but of course it's difficult to find the right balance of visually impressive execution and glitches that can't be reliably performed in real time, versus length. Commentary helps a lot. First, to get a few out of the way that aren't on the console-verified list... I sort of wanted to recommend Metroid 2, if it were console verifiable. At 25 minutes, it's a bit long for people who aren't familiar with it to watch Samus wander the corrupted caves of SR388. It's a shame there isn't a verified run of Super Metroid, because the movement execution in that is always a joy to watch, for me. It's really the perfect choice. I'd jump at any Castlevanias, were they on the list. Especially post-SotN ones. Then, of those on the list... Marble Madness is quick and cool. I like how the camera is usually far behind. I expected RBI Baseball to be boring, but it's hilarious. I definitely vote for that one. I like the NES Mega Man "game end glitch". No, I love it. It's almost too short, but that makes it a good transition between other games. Mega Man 2 is pretty long, but with all the glitches and oddities, I could see including it. Super Mario Bros. "warps" and Super Mario Bros 2 "warps" are solid choices. Super Mario Bros 3 "game end glitch" is gold. Great choice. I do not recommend the "warps" one, because it's 50% autoscrollers that bore the hell out of me. Tetris "fastest 999999" is enjoyable. It's not so long as to wear out its welcome. Dr. Mario has an ... interesting ending. I could see including it. A lot of the rest are well made, but simply not that compelling to watch in a marathon setting; I mean, I could see the room sitting there in silence as they play. What is fun about a TAS varies by person, though, so I don't know everything.