Hm...I think one thing you should look for is that a TAS should, in some way, subvert expected gameplay. This can be by doing frame-perfect glitches or sequence breaks, or by getting uncommonly close to enemies, or any number of other things.
The difficulty with rhythm games is that they are specifically "perform this sequence of button inputs as accurately as possible." Sure the TAS can do this perfectly, but there's no subversion going on here; the TAS is doing, well, exactly what you'd expect to see from perfect play.
Puzzle games, similarly, are generally "perform this sequence of actions". If the TAS just performs that sequence, even if it's going with frame perfection, there's nothing surprising for the viewer. There are puzzle games where you can break this, though, so it's not a hard-and-fast rule.
Racing games...well, it really depends on the game. If you have to repeat the same route 3 or 5 times to finish a course, then you have a lot of repetition; if the AI is hopelessly outperformed then there's not a lot of interest once you've seen one lap. Again, not all racing games are like this, but plenty of them are.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
I'm gonna agree with Nahoc here that it sounds like both runs would be qualitatively different enough to warrant being published. They amount to different path optimization goals; both paths have the potential to be entertaining and I don't think you could say that one is always going to be more entertaining than the other.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Nach: oh, don't take my post to mean I don't think this should be published. It was just an observation of how strongly nostalgia influences our decisions here.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
I'm getting mixed signals here. People complain that they don't rate because it's so many clicks away. A design change is suggested that would ostensibly make them rate more often.
Exactly how does this help reduce the traffic? Wouldn't it be the exact opposite?
Imagine that there's a little widget next to each movie that lets you rate it. When you use the widget, it sends a message to the server to update the rating for user "Warp" and movie 1701 to be 8.2, say. The server's extra load under this scenario is:
1) Generating the widgets for each movie on the page. This should be very cheap; the HTML is trivial (a single function definition and some straightforward HTML for each movie) and the extra database query needed to load the user's current ratings (so the widget remembers what you've rated already) should be a nearly-trivial join of movies vs. user ratings.
2) Handling rating requests; this doesn't even require a page to be loaded, just a database update to be performed; since there's no need for it to be done quickly, it can take as long as it likes.
You'd have to do many ratings for this to exceed the cost of doing a single rating under the existing scheme, where rating a movie requires generating two entire pages.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Changing the rating system design is not going to help that problem. It would just be fighting the symptoms rather than the root cause.
If that's the only reason to change the current design, that's a really bad one.
Reducing the number of pages that the site has to generate would reduce the load on the server, thus giving it more spare cycles to generate other pages. It's only an incremental improvement, sure, but it wouldn't be pointless.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Man, if this were an NES run without any nostalgia backing it, it'd be rejected so hard.
As someone who never played the Commander Keen games (or indeed basically any DOS games) as a kid, frankly I've found them all to be pretty dull, and the earlier games in the series are certainly no exception. Meh vote for game choice.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Thanks for the encode, Tseralith!
This was most excellent. The willingness to take damage really helps speed things up, not to mention the obviously improved precision with climbing and the slow-fall technique. Great work, Jigwally!
A few thoughts, both dealing with the level before "Rushifell":
* At the beginning you have to fly right for a long time before dropping into a pit and then flying left from there. You take damage at this point from one of the toads (which pushes you back slightly) and land, refilling your flight meter. If you'd used the slow-fall technique to extend your flight time, might you have been able to avoid the toad and need to land?
* There's a segment with falling roof tiles that lags horribly. The lag gets better once the claw shot you used to climb up there goes offscreen. Might it have been faster to put the claw shot on the left side of that area instead of the right, so it would go offscreen faster? Of course you'd have more distance to travel then.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
It's talking down the game to a tremendous degree, making it sound like something only a masochist would enjoy playing. It's a hard game, yes, but generally it's not unfairly hard, which is really what Fake Difficulty is about. The controls are tight (unless you're playing as Luigi anyway), the very first mushroom you find is poisonous (thus teaching you about the mechanic before you've sunk a lot of time in), and the puzzles are mostly about nonlinear thinking. Probably the most unfair point is the reverse warp zone, since you're being punished for finding a secret.
That said, it is very hard -- I would say that it's probably a bit harder than the second quest in Legend of Zelda (while LoZ's first quest is harder than the original SMB), though I've replayed LoZ much more recently than I have SMB:LL.
The comments about sticking elements where they don't belong is kind of strange because the original SMB (and this, which is basically a mission-pack sequel) don't have any thematic consistency anyway. Remember the fish in 8-4 in the original SMB?
Suggested alternate description:
The Lost Levels is a remake of the game released in Japan as Super Mario Bros 2. It introduces a number of new obstacles into the standard Super Mario Bros. formula, including poisonous mushrooms, wind, reverse warp zones, and of course many tricky jumps. It's difficult enough that Nintendo of America decided to completely re-theme a different game and release that as Super Mario Bros. 2 instead; the Lost Levels is the first time the Japanese sequel was released to the rest of the world.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
If they made the additional accounts they'd have to have each account make 5 posts, too, before they could vote. I suppose you could make 15 non-spammy posts in advance, and effectively have vote-capable lurker accounts, but it seems like a lot of effort to go to to influence a voting system that the judges rarely need to pay much attention to.
The other alternative I can think of is that someone found an exploit in the voting system that let them record multiple votes.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
I enjoyed it. It's a different take on the concept of a TAS; more leisurely, less "must be doing something every frame or else the viewer will get bored!" I don't think this kind of approach will be valid for all games, and I could believe that eventually we'd get tired of it (as we did with the multi-run concept), but I think this is worth publishing (with the known improvements of course).
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
That's a shitton of text to read in two seconds. In fact, that's over 40 characters per second - in that time, many won't even be able to read which criminal encoded the release if the logo de-emphasizes the fact like it should. If the personalized blurb is removed and the duration is extended to three seconds, we're looking at slightly over 20 characters per second, which is barely readable.
When reading you don't actually pay much attention to each individual character. Reading this would be more like "blah blah blah Tool Assisted Speedrun blah blah URL (encoded by Joe)". So there's four important "words" in there (given that the attribution is visually de-emphasized), one of which (the URL) doesn't get parsed unless you actually care about the origin of the video. This is not hard to read.
Telling viewers to pause is not a solution.
"Things that solve the problem I'm trying to convince you exists in a way that does not achieve the result I'm arguing for are not a solution."
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
What's the reason that you can't just laser every stage? Does it take too long to manipulate it, or can it not appear on most stages, or what?
I believe the powerups for each level are set -- at least, there are several levels where there's a difficult-to-get-to central block that, once broken, releases a powerup that makes the rest of the level simpler. That's too convenient to be a matter of chance.
The laser is the most expensive powerup to actually buy, at 280MP per the notes, and the maximum MP you can get on a given level is the number of blocks on that level (which would require clearing the entire level with one shot of the ball). Since there's far fewer than 280 blocks on each level it's impossible to just buy the laser every time.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
I want to know who encodes the movies I watch. It's a nontrivial amount of work and I need to know who to thank for it!
TASVideos already has guidelines for how attention-grabbing a logo is, and what content it needs to include; that's quite enough IMO.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Nifty stuff! Clearly all Breakout needed to be interesting was bigger powerups!
Also, the splash screen before the game starts is so clearly an American addition to replace what was probably deemed an excessively cute original splash screen. ZOOM IN ON OLD MAN'S FACE!
And the boss of the desert was a sunbathing furry thing. Was she trying to get a tan? 'Cause...that's not really gonna work.
The only thing I really wondered about speedwise: against the second boss (the raccoon girl with the bubbles), wouldn't it have been faster to get hit by bubbles so the ball wouldn't have as far to go before you could hit it again? You did this sometimes but not always.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
King of Demons could be interesting. There's a Let's Play of it which is how I found out about it. Appears to be a fairly standard platformer, but after beating some of the bosses you're given the opportunity to turn into one of three different demons with varying abilities -- and the transformation sticks until you pick a new one. Using each of the basic demons once gets you access to a super form for the last levels, or you can pick the same demon several times to power it up more.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Interesting! The simultaneous ExciteBike was rather confusing though, I have to say. Also, for the individual-track ExciteBike run, it'd be amusing if you could animate the guy wrapping around vertically somehow. :)
It occurred to me that Kid Chameleon would make a good candidate for this type of video, though obviously it's not an NES game. The levels are often huge and you only see a small part of them at any given time.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
I agree, looks like it should make an excellent TAS.
Regarding the banana issue, what you might want to do is make two runs -- one that does all the normal levels (including world 11), and one that starts from a savefile to do the last world. You'd need to include a movie that generates the savefile to prove its validity, but it doesn't need to be tool-assisted.
Alternately, if you have a decent idea of how many bananas you could obtain "naturally" over the course of playing each level once, then you'd have a better idea of just how much grinding there'd be. If it's not too much then you could do the whole thing in one movie. But my main concern there is that you'd be making detours in each level to pick up bananas, and this really seems like a game where you should be aiming for fastest in-game time for each level (like with the Sonic games), for better comparison with the world record times.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Has it been proven to be possible to win the game with defeating Bowser only twice (or once for that matter)? I mean does the game check up somehow on how many times you actually have defeated him to reach the final stage? Also can you exit the basement from the inside?
Isn't that kind of like saying "Does the game check for 70 Stars before letting you past the endless staircase?"? People have managed to get past that, so I don't see why they couldn't skip the first Bowser fight in the same way.
No, it's more of a question of how the game decides when it's over. I could conceive of the game ending when you've obtained 3 keys (in which case you'd have to defeat all three Bowsers). But personally I suspect it's far more likely that the game-end logic is "attached" to the arena you fight Bowser in, which in turn must be attached to the level you reach the arena from, which is attached to a specific point in the castle. So I'd be very surprised if, on reaching the final Bowser level without having the first two keys, you'd be unable to win the game.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
This is one of those "when is the run really over?" questions. The timing we usually care about is the time from the start of the movie to when the last input is delivered, which can result in cases like this where no input is actually needed to finish the boss, so no input is given. Other examples include any% Gimmick and, Gradius 1, though those examples are less extreme.
Certainly the credits could be reached sooner, but the cost would be extending the input. At some level this is frankly a stylistic question, though.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Punch-Out certainly plays games with the clock; IIRC it slows down when you have an opponent stunned, for example. And there's absolutely a mode where your jabs are performed much faster than usual. I'm not aware of any situations in which the clock runs faster than it should, but I'm hardly an expert.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
