Moderator, Senior Ambassador, Experienced player
(907)
Joined: 9/14/2008
Posts: 1014
Edit: The competition is over! The game was Streemerz in Streeeeemerz mode used with the permission of the creator Arthur 'Mr. Podunkian' Lee and thefox who ported it to the NES. There were over 1,000 downloads of Streemerz from our mirror and over 150 legit submissions completed the game. While there were a number of problems with the TASJudy script we were able to determine in the end that Mothrayas was the winner (despite preliminary results that kaizoman had won). A proper real submission of the category was made with the help of multiple authors and overall feedback has been overwhelmingly positive. Thanks to everyone who helped make this possible and to everyone who participated!
Mothrayas later made a ghost encode showing the top 10 finishers that's highly worth watching:
Link to video
And Bisqwit made an encode showing different strategies in each room (for the runs done in FCEUX):
Link to video
The original post is below:
I'm happy to announce that there will be an officially sanctioned speed TASing race starting at the end of the TASBot run at SGDQ 2015 sometime on August 1st. From the trial speed TAS race description: In chess, a grand master in an untimed match may take an extremely long time to think out what will happen next and rewind in their head to plan out a game. In a a speed chess scenario, the same grand master has only a limited amount of time to consider each move. Speed TAS'ing is like speed chess - the TAS'er has a limited amount of time to TAS a game, so efficiency with researching the target game and using TAS tools is required.
Here's how this competition will work:
The target game will be announced at the end of the TASBot block (consult the schedule as the time will shift)
The (legally redistributable) NES ROM file will be posted on TASVideos.org along with instructions
Participants will have a minimum of 4 hours (likely closer to 5) to submit a BizHawk (default QuickNES core) or FCEUX movie file that completes the game
The soft cutoff will be an announcement I'll make between Mario 64 and Super Metroid
The hard cutoff will be shortly after said announcement, i.e. there will be a minimal grace period likely ending right at the actual start of Super Metroid
Submissions will be judged and organized by time to last input using an automated bot that will ensure the ending screen has been reached
The single fastest submission will be verified and subsequently replayed at SGDQ directly following Super Metroid (during the setup time)
There will be many more details in the instructions that will be posted when the competition starts, but here are some pointers on the game in question:
The game can be completed in under 6 minutes - how much under 6 minutes is the interesting question
There is an existing TAS movie file that is related to the target game that can be referenced for research, but any research you do will have to be done in the same timelimit for creating the TAS so time management will be critical
The game is challenging but fun to play and should be enjoyable to go through with TAS tools
Perfection is not possible with a short timeframe; consider taking shortcuts and using all of the tools at your disposal, including turbo button presses if they help
It may be very wise to complete a very sloppy playthrough using savestates to get a good feel for the game length to help in knowing how much time you have to optimize the run
This is the first attempt at doing this in front of a live audience with a set timer so things may go wrong; I ask for grace in advance in case something goes catastrophically sideways, such as 30,000 people slashdotting the site or delays in announcing the winning submission due to technical difficulties. With that said I'm expecting this to be a lot of fun and I'm looking forward to the competition. If you have any questions, please feel free to post below. Thanks, and enjoy!
Moderator, Senior Ambassador, Experienced player
(907)
Joined: 9/14/2008
Posts: 1014
Zurreco wrote:
Are you going to select a game that has simple movement mechanics?
Finding a game with good attributes has thus far been the hardest part of organizing this event. I spent an extensive amount of time researching games for this event to find one that could be legally distributed, relatively short, simple to understand, and fun to play. I feel that I've made a good choice based on the feedback I've received from individuals who have volunteered to be beta runners for me. I know from my own experience testing out the target game that the movement mechanics are certainly simple enough to understand in the allotted time, although I also anticipate that given more time there will be tricks I haven't discovered yet that should make for a good TAS.
Thanks for the interest - I hope you're able to participate!
Moderator, Senior Ambassador, Experienced player
(907)
Joined: 9/14/2008
Posts: 1014
(This is a test post of what will go on the front page in the TASVideos News column, feel free to ignore or suggest changes.)
A normal TAS is like a grand master in an untimed chess match playing perfectly. A speed TAS is like the same grand master playing speed chess slightly less perfectly. At the end of the SGDQ 2015 TASBot block on Saturday Aug. 1 an NES game and a ~4 hour time limit will be announced during which participants will be able to submit a BizHawk or FCEUX movie file, with the winning submission played back directly following Super Metroid. More details can be found in the competition announcement.
Moderator, Senior Ambassador, Experienced player
(907)
Joined: 9/14/2008
Posts: 1014
RobynS wrote:
Shouldn't just one emulator be decided? Are there not timing diffs between BizHawk and FCEUX.
I've done some testing and BizHawk requires exactly 1 extra frame of no input at the start of the movie. In fact, the auto-judging script is a BizHawk Lua script that accepts both formats but adds one frame to any .fm2 files it imports. Testing so far has not shown any other desyncs or time differences throughout the entire run, so we should be good to go. Good question, though!
Shouldn't just one emulator be decided? Are there not timing diffs between BizHawk and FCEUX.
I've done some testing and BizHawk requires exactly 1 extra frame of no input at the start of the movie. In fact, the auto-judging script is a BizHawk Lua script that accepts both formats but adds one frame to any .fm2 files it imports. Testing so far has not shown any other desyncs or time differences throughout the entire run, so we should be good to go. Good question, though!
thanks, does this means that BizHawk's timing of the game will be the judged one?
Moderator, Senior Ambassador, Experienced player
(907)
Joined: 9/14/2008
Posts: 1014
NitroGenesis wrote:
I can't participate due to a lack of laptop, but the game's homebrew, right?
Sorry to hear you don't have a laptop - maybe you can bum one? :) All I can say about the game is that I have contacted the authors and the ROM file can be legally redistributed. More details on the exact arrangement will be provided when the name of the game has been announced.
RobynS wrote:
thanks, does this means that BizHawk's timing of the game will be the judged one?
Because the script is always going to add one frame to any .fm2's it judges to bring the FCEUX movie to be the same length as BizHawk with the QuickNES core, yes - the actual movie length for the purposes of the competition will be based on BizHawk. I'll add links to each sanctioned emulator to my initial post.
Sounds fun! I might participate just to see how I compare against experienced TASers.
Questions:
- By "submission", are we talking about the standard submission process, or is there going to be a special submission procedure?
- Anything planned if TASVideos goes down during the race?
Moderator, Senior Ambassador, Experienced player
(907)
Joined: 9/14/2008
Posts: 1014
SmashManiac wrote:
- By "submission", are we talking about the standard submission process, or is there going to be a special submission procedure?
I'm working with RGamma to create a static page that does a minimum of three things - hosts the ROM file, displays explanatory text I'm drafting, and allows people to upload BizHawk and FCEUX movies *only* (preferably also limiting size / length) and places them in a folder that can be downloaded from to allow the submissions to be automatically judged. The main submission form should not be used - while I originally envisioned this as the only page accessible for the duration of the contest there's some talk in IRC about leaving the links to the rest of the site active, in which case it would be prudent of us to add a bolded disclaimer on the normal submission page for those four hours or so.
SmashManiac wrote:
- Anything planned if TASVideos goes down during the race?
Um, panic? :) The best we can do is turn things off and ensure the site is as static as possible. We've already experienced high load this week - Nach posted an image that shows the load on Sunday was pretty high. I'm sure we'll survive one way or another.
Moderator, Senior Ambassador, Experienced player
(907)
Joined: 9/14/2008
Posts: 1014
TehSeven wrote:
Is it only one person per movie?
I was secretly hoping no one would ask me this. :) The answer is that TASVideos has team-based contests as part of the Dream Team Contest competitions. I specifically designed this contest in a way that it can be completed by an individual, both in the length of the competition and the difficulty of the target game. While my intent in creating this contest is seeing a swarm of individuals each having fun playing through the game on their own, I don't want to create excessive rules against people collaborating that I wouldn't be able to enforce anyway.
So, while the intent is that each individual participant submits their own movie, I fully expect that there will be groups of people who communicate using various open, semi-open, or private methods to collaborate on glitches and helpful techniques. I consider this type of behavior natural and to be expected and as a result there is a very real possibility that a movie file with multiple author names embedded or multiple otherwise identical submissions of the same frame length and inputs from different people may occur, in which case I will simply give credit to all of the winners.
So, go have fun and don't feel encumbered by heavy rules!
Moderator, Senior Ambassador, Experienced player
(907)
Joined: 9/14/2008
Posts: 1014
MUGG wrote:
Is this just for everyone or only for visitors of sgdq?
This is an open free competition that doesn't even require participants to be registered on TASVideos.org (in part because, um, can you imagine the chaos of the lowest denizens of Twitch Chat descending on the registration system in flocks and droves while DeHackEd looks on in horror as they pick apart the serv... I'm getting a little too detailed, I should stop before I frighten off staff... :)
Being a bit more serious, this is the most open event TASVideos has done and I anticipate a large amount of outside publicity and page views. I wouldn't be surprised if a wave of new users appear following the competition. I encourage everyone here to welcome any new members into the fold - while there will always be a few bad eggs I'm not anticipating this to be an Eternal September as those who are genuinely interested in TASing will stick around and those who aren't will fade away like so many disinterested freshmen, but I'm waxing poetic again. I for one, will welcome our new members and encourage everyone here to show the same grace that you were shown when you first joined.
Moderator, Senior Ambassador, Experienced player
(907)
Joined: 9/14/2008
Posts: 1014
MUGG wrote:
What version of FCEUX is good to use?
The one linked above - it shouldn't matter as the game seems stable but I will explicitly note the latest version of each emulator in the instructions.
I think this is a fine idea and it sounds like you've already laid out the rules, but may I make a suggestion? If possible, have an unassisted runner play the game, either live or pre-recorded for comparison. It seems to me that half the purpose of doing such a demonstration is to show off tricks that cannot be feasibly replicated unassisted. If an unassisted human completes the game in 10 minutes and a TAS completes it in 6 minutes while abusing collision detection, zipping, and other glitches, that would be very impressive. Give the audience a benchmark for comparison.
I understand if you don't have the luxury to set aside extra time for a human-TAS race.
Moderator, Senior Ambassador, Experienced player
(907)
Joined: 9/14/2008
Posts: 1014
Bobo the King wrote:
If possible, have an unassisted runner play the game...
I understand if you don't have the luxury to set aside extra time for a human-TAS race.
We pitched exactly that idea for both AGDQ and SGDQ. Both times the submission was rejected due to time constraints. I practically forced this competition into existence by convincing staff that I could do it without adding time to the marathon. If this competition is successful I hope to do the contest again with human runners competing. Thanks for the feedback!
Joined: 6/5/2006
Posts: 188
Location: Malmö, Sweden
Just submitted my movie. This was a really fun idea, a well chosen game and a reasonable time limit. Lots of thumbs up.
If possible you should try to display it on stream as a race between the top 3-4 times I think.
Even the best player is limited by the speed of his fingers, or his mind's ability to control them. But what happens when speed is not a factor, when theory becomes reality?
I agree on the fun, good game choice, and on the reasonable time limit.
Without knowing how long the game is going to be (i.e. how well I'm doing compared to the advertised goal time), the time limit seemed somewhat tight, and I began doubting how much I can spend time optimizing, but in the end it went well.
Can't wait to see how badly I was beaten :-)
