Posts for Randil

Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
I've tried left+right and up+down in this game, and right+left seems to have the same effect as just pressing right, and up+down as just pressing up, so I don't think this can be abused.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
Fabian wrote:
Got my first video with proper audio and stuff up on youtube today, pretty happy with it. Another Elton John song, check it out imo: Elton John - Can You Feel The Love Tonight (Piano Cover)
Very good sound and video! I like your flow through the whole song, and you've obsiously very talented. Great job!
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
This run ought to be good. *watches the run*... wow. This run is awesome, I think I'll even go as far to say that I enjoyed it more than the any% run, and that means a lot. This game is really well suited for TASing. This run will get very high ratings from me. Very big yes from me, and I certainly wouldn't be against either this or the any% run getting a star.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
That strategy on stage 2 is faster, a whole 66 frames faster! Nice!
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
DarkKobold wrote:
There is no mention of the use of MATLAB in the submission! That is a travesty of epic proportions!
Haha, you're right, I will edit it into the author's comments! :D
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
The beautiful thing about lua script is that you don't need to install anything, they automatically work if the emulator supports lua, which FCEUX does. Today I made a lua script for this game that checks if you lose a frame anywhere. Basically, what the script does is compare your X and Y position between two frames, and if they're the same, the program will tell you that you lost a frame there. This works well for this game, because you should be constantly moving in a TAS. I have tested the script, and it works very well. Here's the script, uploaded to mediafire. The only thing you need to do to get it to work is to go to File->Lua->New lua script window. In the new window that pops up, select browse and choose this script, then click on "Run" to start the script. Unpause the emulator if it is paused. The top left corner of the emulator should now say "Lost frames: " and then a number. Now, play one of your movie files with this script running. Whenever "Lost frames" increases, the script alerts that you have probably lost a frame here, so you should investigate it further. Pretty handy, right? No more checking for lost frames manually!
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
This was awesome, one of the best runs I've seen in a long time. Here's a very strong yes vote. Great job, you two!
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
mklip2001 wrote:
Wow, I wasn't expecting your last run to be improved this soon. Nice job with it! The zippings are especially cool. A couple questions, though. In 3-2, is it worth getting into the ceiling so early? (It seems like you spend a lot of time setting up the glitch, when you get onto the ceiling pretty soon later in the stage anyway.) Second, for the world 3 boss, why don't you use the same strategy as in worlds 4 through 6 of moving the pump left and keeping the boss still with Wicky?
Thanks! :) About level 3-3 boss, this is answered in the author's comments: "On level 3-3, the boss can't be held in the right corner because he would cover the pump. " Basically, the boss' hit box is too big, so if I held him in the corner with Wicky, he would cover up the pump so I won't be able to jump on it. The pump here is stationary, so that's why the strategy from world 4 to 6 don't work. About 3-2, I was just about to say that it's not possible to zip through any other place in this level when I suddenly managed to zip through the roof at the beginning of the level, saving 146 frames. I edited this in, and the movie synched. Here's the fixed movie. I will ask adelikat to replace the movie file for this submission with this fixed movie. The only change in this new movie file is that the wall zip in level 3-2 is done at a better place. Now the movie finishes at exactly 10:00 in game time! :)
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
Derakon wrote:
You should specify that n<k; otherwise you could have 100 d4s and a probability of zero that X=k.
I'm terribly sorry, this was my mistake.. the probability I was looking for wasn't P(X=k), but instead P(X=m), for any integer n-1<=m<=k*(n-1). Basically, I want to know the distribution of X. Sorry for the confusion.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
I TASed the first four levels of this some time ago, just for fun. I have no intention of continuing it. We used the the exact same strategy on the first level, but my version is 33 frames faster. I also started the first level 1 frame faster, but that's a small detail. Here's the movie file. Perhaps you can shave off even more frames from it! And also, welcome to the forum! :)
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
Here's one I've been thinking about for quite some time now, but haven't been able to solve: Suppose you have n dies, with k sides each. You roll all of them, and remove the dice with the lowest result. You add all the other results, and call this sum X. What is P(X=k)?
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
Solution to the alien symbol problem: I assume that the aliens' numerical system also includes negative numbers. This solution is not the fastest one, but it works. For simplicity's sake, let's call the symbols A,B,...,J (10 symbols for the digits 0,1,...,9). To find what symbol stands for 0, calculate A-B, and if the result is A, we know that B is 0. If not, continue with A-C and so on, up to A-J, until you find the symbol so that the result is A. If you don't find it, you know that A is 0. In order find the digits 1 to 9, first remove the zero symbol. Let's assume that J was the zero symbol. To find what A is, calculate the 8 values A-B, A-C,...,A-I. Find how many of these results were negative, and call this number n. Now A=9-n. Do this step for all remaining symbols, and the problem is solved. This algorithm will work in any basis, not only basis 10. Duplicates can be found if the resulting difference between two symbols is 0, which the algorithm will find when calculating A-B, A-C, etc.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
I was wondering, would it be possible to add a feature that puts converted .fcm's into the movie folder of FCEUX? For example: If I convert an .fcm that is in the movie folder of FCEU 0.98.28, the converted .fm2 is created in this folder too. Then I have to manually move the .fm2 from FCEU's movie folder to FCEUX's movie folder, which gets a bit annoying after a while if you do a lot of converting. I personally want all my .fm2's in FCEUX's movie folder. EDIT: This gave me another idea: In Windows, I think it would be very convenient if you could right click on a .fcm-file, and have the options "convert to .fm2" in the menu that pops up. Would this be possible to implement?
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
]0,c[ isn't dense in [0,1] since it doesn't include elements >c, so I don't think that will do. I think the answer is some kind of exotic set, since it seems pretty counter intuitive to me that the set has to be dense in [0,1] and yet have a lebesgue measure <1. Perhaps it involves the cantor set, or some strange set including only irrational numbers? I'll give this some more thought...
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
Sonikkustar wrote:
Very Nice! I loved every moment of it. However, the ending can be improved slightly. Its no big deal, and its an easy fix. :) This looks very submittable to me.
Thanks! If you mean that I could have ended input earlier and still beat the game, I did consider this, but chose not to. I aimed to actually reach the ending sequence as fast as possible, and ending input earlier would make me reach the end of the level a few frames later. I hope this made sense. If this was not what you mean, could be please be more specific?
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
Double posting to inform that I have now finished the run! I've worked really hard these past few days, and I think it really payed off to redo this run. Here's the .fm2. The rerecord count is off due to lots of editing and splicing. The correct rerecord count should be around 30 000. I might edit this before submitting. This new run is 1276 frames, or 21.27 seconds, faster than my previously submitted run. This submission includes zipping through walls on 3-1 and 3-2, and zipping through flag poles on 5-3. The levels on world 7 have been greatly improved, thanks to strategies by narimasa. Almost all levels have been improved, so if you liked the previous run, I suggest you take a look. I will leave this movie in this thread for a few days, for people to take a look. If you think something looks weird, or think you have an idea for improvement, please let me know. I'd like to avoid having to cancel my run a third time. Trivia: interestingly, this movie synchs on the (J) version of the game, so this run and narimasa's run are comparable.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
I've been redoing my submitted run from scratch. I've just finished world 1, and have saved 20 frames over my last submitted run so far: *4 frames saved on 1-1 by picking up Wicky without jumping, and optimizing the section around frame 820. *4 frames saved on 1-2 by not slowing down at the beginning of the level. *12 frames saved on 1-3 by faster boss fight. Here's the .fcm. Please post any ideas/comments you might have! I have analyzed in detail how the herb moves and where I should kill the boss in order for the herb to appear at a convenient place. I can post details if anyone's interested. I have also found a way to zip through flag poles, but this resets your speed to 0, so I'm not sure it will save time anywhere. EDIT: Just finished 2-1, and boy did I save time here! I used narimasa's strategy at the beginning of the level, and managed to get up the height around frame 6450 much faster. All in all, I saved 71 frames on this level alone! I'm quite amazed, actually. Here's the .fcm. EDIT2: I don't know if anyone reads this, but anyhow: I've just started world 4 now, so I'm about half-way through the run. I'm saving a lot more time than I thought - I'm already 9.4 seconds ahead of my submitted run. Both 3-1 and 3-2 use wall zipping. World 5 will include zipping through flag poles, there are two places in this world where this will be included. I think his new run will look great! I've switched emulator to FCEUX now, so from now on all WIPs will be in .fm2 format. I hope no one minds. Here's the new WIP.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
I made a script here that doesn't brute force anything, but instead uses RAM addresses to know when to press a button. This means that no savestating is required. You might want to take a look at it.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
Thanks, and I feel silly for not reading up on lua commands in FCEUX help, I'll make sure to read up on that before asking more questions.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
Sonikkustar wrote:
Nice one Randil! Is it possible to do the zip in 3-2 too? I look foward to the final product. :)
I haven't had any luck zipping in 3-2 yet, until now - ironically, while answering this post, I decided to give it another try just to make sure, and actually managed to pull it off! I was just about to write that I didn't think it was possible, so thanks for reminding me :) Here's the .fcm showing this. The wall zip probably isn't optimal, but it still saved 45 frames, so you will see it in the next run.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
I'm wondering, is it possible write a lua script that only performs some commands when a movie is playing? I mean something like this:
while (movie is playing)
(commands)
end
And while we're at it, is there a way to get the length of a movie? I know movie.framecount() gives the frame that is currently playing, but I don't know how to fetch the actual length of the movie that is playing.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
Regarding the matlab program. Your Y speed when you make a jump depends on your X speed when you press A. It also depends on if you hold down right or left when pressing A. And in turn, your initial Y speed will have an impact on how many frames your sprite is inside the block when you hit it with your head. The matlab program I wrote wasn't very complicated. I simply told it how the jumping physics worked in this game, and it told me what X speeds that resulted in a jump that gave me maximum time inside the block. That's mostly what I use matlab in TASing for. If the game's physics are easily predicatable, you can just write a matlab program that simulates these physics and can brute force test all (or at least extremely many) possible ways of solving something.
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
TaoTao wrote:
Randil wrote:
It would be awesome if it could be incorporated in the run.
I agree. I hope this glitch is useful :D And, I found narimasa had uploaded his improved run (J version). He says it is in 49474 frames: http://dic.nicovideo.jp/v/sm8222456 (apparently, fcm/fm2 is not available)
Interesting! It would have been a lot easier for me if there was a movie file available. Although my movie is generally faster, he pulls off a jump in level 2-1 that I think will save time, but I'm having problems pulling it off. Do you know this narimasa, TaoTao? Perhaps you can ask him to upload his movie file somewhere? I'm also wondering, is there an easy (and legal, of course) way to download videos from nicovideo? If the .fm2 or .fcm isn't available, it would be easier for me if I had the nicovideo movie on my hard drive. I have also managed to perform the wall zip in level 3-1! Here's the .fm2 of this submissions with this edited in. The zipping is optimized, I wrote a matlab program that optimized it for me. This movie is 264 frames faster than this submission so it saved quite a lot of time. In the light of both this new zip trick, and narimasas new movie, I will cancel this submission. I'm very sorry for the inconvience, in particular since the reason for cancelling this submission is very similiar to why I cancelled my old version. I will need some time to fully investigate this new wall zip trick, as well as analyze narimasas movie. This game is extremely edit friendly, so hopefully it won't be too long before I can submit the next version. And I'll make sure I don't abandon this project for a year again! At least we know for sure that the next submission will be at least 264 frames faster than this one! Any future discussion about this submission and Kid Kool in general should probably be carried out in the NES thread..
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
TaoTao wrote:
I'm not so familiar with this game, but today I found a person mentioning about a glitch to pass through ceilings in a Japanese BBS. http://schiphol.2ch.net/test/read.cgi/retro2/1175717336/192 (in Japanese) I tested the glitch and made a movie: http://dehacked.2y.net/microstorage.php/info/1083112279/KidKool-CeilingGlitch.fm2 Does this glitch help to improve your run?
Wow! That's crazy! Thanks for telling me. Hmm, I'm not certain it will save time anywhere, I'll have to take investigate this further. It would be awesome if it could be incorporated in the run. Any judge should probably wait with accepting this until I have found out more about this. EDIT: Ok, I think I know what is causing the zipping. It is the roof that is pushing me forward, and not the wall that's dragging me in. The trick is to movie inside the block above me for 3 frames. This is done by setting up your Y subpixel position in a certain way. As soon as you move inside the block, the game will try and push you out. If you've moved deep enough into the block, it will take 3 frames for it to push you out. If you change horizontal speed within this 3 frames span, the block will push you out horizontally the opposite direction you're moving in. And since you turned around inside the block, it will push you in the right direction. The game doesn't care about if it's a wall in the way, it will still push you out like this, and this is how you can move inside walls. I'll edit this post again once I've tried out if this saves time! EDIT2: Hmm, scratch that, it seems to be some more factors behind this...
Experienced Forum User, Published Author, Skilled player (1890)
Joined: 4/20/2005
Posts: 2160
Location: Norrköping, Sweden
I liked how you had 4 minutes to beat each stage but usually managed to beat them in under 10, sometimes 5, seconds. Obvious yes vote!