Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
Rock on! Great run!
I noticed one thing about this movie, though. I played the published fm2 file, and it synched all the way through, but the emulator lagged for a few moments every 256 frames or so... even when fastforwarding. I wonder why.
Also, I've always wondered why some of the lava dragons spit fire, when others don't.
I'm still convinced there are many more screen scrolling glitches in this game to be abused. If you look closely at this movie, you can see a moment where the room scrolls both horizontally and vertically! Granted, it's severely corrupted, but still... it shows the engine is capable of that.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
I've been playing Braid almost since it came out on the 360, and I don't even own a 360. I also got it the day it came out for PC. I still haven't beaten the Speedrun Mode or gotten any of the secret stars.
Truly brilliant level design, and the innovation and implementation of its gimmick mechanic rivals that of Portal. The beautiful use of licensed music doesn't hurt either.
I'd recommend this game to anybody. TASers would especially enjoy it, they'd probably master the rewinding a lot faster than most other people. But they'd still get hung up on a lot of the puzzles.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
Unrelated to the TAS improvement, but I've just made a video that some might find entertaining.
http://www.youtube.com/watch?v=vXkIxp8dlyY
This movie aims to pop a maximum combo on each level, then eat all the fruit. The current movie goes up to Level 20.
I'm stuck on Level 35 at the moment, but I'll post another movie once I beat level 40.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
There are so many for StarCraft it's not even funny. Too bad we can't really implement them, except maybe in SC64, but who'd want to play that?
Absolute Victory: On the last Protoss mission of the original campaign, destroy every last zerg before killing the Overmind.
Splat: Destroy 8 friendly units with one badly-placed Spider Mine.
Get Out Fast: Defeat Terran 10 "The Hammer Falls" of the classic campaign in less than 10 minutes.
Feel the Rush: Beat all 30 levels of the classic campaign on the highest speed, without saving. (Not necessarily in the same sitting, though)
Tech Snatcher T: As a Terran player, steal 6 or more enemy add-ons in a campaign level.
Tech Snatcher Z: As a Zerg player, infest 6 or more enemy command centers in a campaign level.
Suicide Squad: As a Zerg player, kill at least 20 enemy buildings and 50 enemy units with Infested Terrans on a campaign level.
Sunken Rush: As a Zerg player, beat a campaign level without building any units besides drones. (Overlords don't count.)
Tactical Strike: As a Terran player, launch a nuke that kills at least 30 enemy units in a campaign level.
Fate Averter: In a campaign level, save many units or buildings from a scripted attack meant to destroy them.
Show Me The Money: Acquire 20,000 or more minerals and gas in a campaign level.
I Have a Job To Do: In a non-installation campaign level, destroy over 50% of destroyed enemy units (buildings count) using heroes that will fail the mission if they die.
Max Production: Create 200 worker units.
Quick Insertion: Beat Terran 6 "Norad II" of the original Terran missions in less than 10 minutes, and without destroying any enemy defensive colonies.
I could go on all day...
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
gia wrote:
I was thinking, this is a tas community, what if there aren't only achievements for play skill but also tas achievements. That means the objectives are quite difficult to execute realtime and a such are meant to require a movie file to accomplish them...
I actually thought of this myself, and spent some time analyzing the idea, but I don't think the idea holds up. For one thing, it is WAY too easy to cheat, so much so that it almost defeats the point. Just share a movie file.
For another thing, it's too specialized an occupation. It seems that it would only apply to this TAS community. Everyone else would just kinda throw up their hands and have no idea what actually to do to accomplish the tasks. So, why not just make the movie and post it on YouTube or something to show off? That's an even better way to get recognition and satisfaction for your work.
For another thing, if you're going to TAS a game, why not just TAS it, and try to get a movie accepted to this site? Why would you want to aim for something like "Beat the game at least 5 minutes slower than the published TAS" when you could just go all out and compete with the movie?
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
Hah. Here's my unfortunate method.
1) Find a game that's never been done before so I don't have to worry about tough competition or wrapping my mind around fifty tricks discovered by other people, also a game that I enjoy or find amusing in some way.
2) Begin TASing, all goes well.
3) Partway through the movie, find that a certain level requires 40 times more patience or technical competency than the rest of the game put together
4) Try said part about 3,000 times before giving up
5) Wait 6 months
6) Randomly try again, and by some stroke of utter luck, get it done perfectly.
7) Submit movie and enjoy reviews, feedback, and publication.
8) 30 minutes after movie is published, find a huge improvement that can save minutes.
9) Begin mentally planning next run, wait 6 months
10) GoTo 2.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
Derakon wrote:
The fact that it's so hard to figure out what the designers did and didn't plan for is part of what makes it fun to break Super Metroid.
Exactly my point. Without the subtlety, the fun which comes from circumventing the creator's will is not there. When you get to the end of MZM with 14% items, and you see your special ending, you feel gypped, because you know it was planned all along, and somehow that's not as fun. You feel less like a clever maverick and more like a tool who has wasted time not doing something truly profound and new.
Personally, I'm just downright glad they acknowledged the misaimed fandom and nodded to them, whether or not they completely recreated what made SM so fun. We've all seen what happens when they take ALL the nonlinearity out of Metroid (Fusion and Prime 3), so in that sense, they gave the fans exactly what they wanted. You shouldn't hate them for that, like so many have. Besides, they couldn't have made it too much like SM, otherwise it would have been just another version of SM. Go play SM instead. They gave it its own unique and good personality as a game.
Besides, while I've never broken SM myself, I found it involved plenty of cleverness trying to beat all of MZM's self-imposed challenges, as the whole path through the game completely changed depending on what you were aiming for, and it wasn't always totally obvious what you had to do.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
I think the writer makes a very valid point, though he has a very tiny reference pool. Also included should be every major "sequence breaking" community, Metroid especially.
Though it raises some interesting questions about the theory of game design. Do players have fun because they are defying the will of the designers? It appears so. My favorite example is Metroid Zero Mission. Great game, great redesign of the old and outdated Metroid 1. But, in a kind and noble nod to the fans that made Metroid and Super Metroid as great classics as they are, the designers of Zero Mission included many sequence breaks on purpose, most of which were masterfully implemented and added layers upon layers of depth and replay value to the game.
...However, simply because the sequence breaks were intentionally placed in the game, that somehow sucked all the fun out of it, so the game was mostly hated by the Metroid fandom. Shame.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
Some better ones for DKC2:
Diddy Fan: Beat every level with Diddy Kong (so only Diddy's face displays on the world map for every level)
Dixie Fan: Beat every level with Dixie Kong.
Prize Winner: Pound the target at the end of every level to obtain a prize.
Grand Prize Winner: Obtain a 1-up from every end-of-level target.
KONG Completionist: Spell KONG on every level.
Trivia Whiz: Answer all of Swanky's questions in the game on the first try.
Family Patron: Complete every trivia game, purchase every advice, save the game at every college, and use every Funky barrel at least once.
Treasure Hoarder: Obtain 99 Banana Coins.
Animal Fanatic: Guide every animal buddy in the game to the end sign.
Clean Sweeper: Obtain every Hero Coin and beat every bonus round on the first trip through every level. (Hero coins from the end-of-level target might be optional.)
Bludgeoner: Kill 15 enemies with one cannonball.
Ariel Ace 1: Kill every enemy without missing a shot on the Parrot segment of Squawk's Shaft.
Ariel Ace 2: Kill every enemy without missing a shot on both the Parrot segments of Bramble Scramble.
Animal Master: Complete Animal Antics without getting hit.
Substance Abuser: Get hit by each of the 3 types of clouds on K. Rool Duel in the same match, and still beat him.
Earthquake Survivor: Defeat Kudgel, letting him stun you every time he hits the ground.
Last Stand Fail: Get KO'd by K. Rool's final cannonball.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
I've had one problem with WMM that I couldn't figure out how to fix. Whenever I overlay text across a moving video or do a fade effect, the whole screen expands by about 2 pixels, then shrinks back down once the text disappears. I figured it was a problem with the compression options and codec I chose, but I couldn't find a combination that would fix it. And it looks really bad.
Demonstration: http://www.youtube.com/watch?v=7Iffe_1Qch4
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
Publication already? Wow, that was quick!
Possible publication text:
Tired of being mocked by the viruses, Dr. Mario switched to his "fast-acting" and "extra strength" medication. The viruses never knew what hit them.
This video beats Level 20, Speed Hi of Dr. Mario in record time. Don't blink, or you might miss the twist ending.
--
MUGG wrote:
After further thinking about this submission, I came to the conclusion I'd better have voted meh on it. While the glitch is cool, there isn't much comboing going on. Your first combo challenge video on Youtube was really impressive. And this submission is merely a glitch demonstration in my eyes... IMO, a TAS of this game should also be longer than just one minute.
I do think you were spoiled by the previous movies.
I submitted this movie because its short length gives it wide appeal (who can grow bored in 1 minute?) and the WTF-factor is priceless. It is not advertised as a combo video, but just as a straight completion video like Baxter and ziplock attempted years ago, which I think it beats with flying colors.
Oddly enough, I was planning on submitting a "Combo Challenge" video as a concept demo, but I could never get past level 05. Actually, after that first video on YouTube, I could never beat level 03 again. So I just gave up.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
Dr. Mario has always had a lot of the same appeal to me that Tetris has: it's a mad struggle to survive, but the game mechanics are so simple (only 3 colors) that you can manage if you focus hard enough, and you have to really think fast and get creative if the game gives you pieces you don't want (just like the S-blocks in Tetris).
Unlike Tetris, you can actually win levels, so it adds more intensity. You can't just bury the last virus underneath a pile of junk pills and survive forever, you have to dig back down to get it if you want to win, while the pills start dropping faster every 10 turns. In fact, this often makes the end of the level just as difficult as the beginning, since your rejected pills will have all fallen down to the bottom of the bottle and buried the last viruses. Also unlike Tetris, if you're clever and think fast, you can make great combos for high score, or to drop junk on your opponent in 2-player mode.
When playing tool-assisted, the focus shifts to setting up huge, mind-boggling combos, which sort of has a Chess quality to it. You have to really imagine and test just how the huge towers will fall and line up to make horizontal combos and such, often without wasting a single pill given to you. Unfortunately, tool-assisted movies really make this game look too easy and don't really give an impression of what the real gameplay looks like.
Obviously I can't make you like the game just by explaining it, but hopefully that explains why I do.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
Sorry, but I'm not going to submit the glitchfest run. To do so would be to undermine the most basic principle of this site, that the game itself remains unmodified. YouTube is a good enough home for it.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
adelikat wrote:
Why does the level end before you clear all the viruses? Does the overloaded combo glitch cause that?
Yeah, it's a glitch, all right. Something in the processing of the complicated combo makes it set the rest of the bottle with the "cleared" flag, ending the level immediately.
I needed to make a sufficiently large combo to trigger the glitch (after clearing enough of the top of the bottle to make enough room necessary to set up for it) involving lots of different colors and (most importantly) a horizontal clear.
I used the pill-changing lua script to test which pills were needed to trigger the glitch and which were not. It didn't glitch as intended unless the bottle contained certain pills in certain positions, most notably in the top left position of the bottle.
The biggest miracle was that I got such a favorable pill sequence to set it all up in the first place. I only got about 4 completely unneeded pills total.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
It works now, but there appears to be a bug. Something disables the menu options for reset and power-on (and the hotkeys as well) and forgets to re-enable them again. After restarting, it works as normal.
Experienced Forum User, Published Author, Active player
(325)
Joined: 2/23/2005
Posts: 786
Here's some more brainstorming.
For the achievements, the modified emulator would look for a lua script that automatically gets run when the game starts based on the rom's checksum. It would be stored somewhere like /gameprofiles/(game). If you're online, the emulator checks the lua script against the one on the server, and if it's not the same, downloads the server's copy and updates the local one. If you're playing offline, it would still record your achievements, but it would invalidate them if it senses that the lua script has changed without being updated from the server.
Lua would need to implement new functions, or maybe a new API set.
arcade.giveachievement(id) - awards achievement, arranges to confirm achievement to the server, and pops up a text box telling the player they've gotten the award. (Text strings are defined elsewhere, possibly in the lua script itself, possibly in an external text file or a database on the server. The achievement might also have a Steam-style icon, which could be stored in the /gameprofiles/(game)/ directory.)
arcade.registercounter(id) - Defines a value that is to be saved and loaded to a data file every time the game is run or closed. Basically, it behaves like SRAM, and will keep a running count of certain values needed by scripts, like a running count of all the goombas you've stomped in every game session.
arcade.addcounter(id, amount) - Adds to the specified counter variable. Accepts positive and negative amounts. If the counter overflows or underflows, the error should be silently discarded and the counter should be set to its maximum or minimum amount.
Likewise, arcade.setcounter and arcade.getcounter.
arcade.toggleachievement(bool on/off, id) - enables or disables a certain achievement. If an achievement is disabled, it cannot be awarded. If the ID parameter is not specified, enables or disables all achievements. This is useful for coding special challenges, specifically those which involve loading savestates or memory.writebyte, when you don't want other achievements to be activated during the "special session".
arcade.menu(string, string, string, string, string) - Displays a hard-coded onscreen menu with the specified options and returns the index of the chosen option, 0 if the menu was canceled. Useful for giving the player an option of activating a "special session". If # options is not enough, the menu can simply open another menu.
I can't see anyone actually following through with this idea in the near future, but it's fun to talk about.