After I saw people working on a TAS for Kuru Kuru Kururin, I decided to make a run for Kururin Paradise, its sequel. I doubt more than a very few of you have heard of it, since it is (at the moment) Japan-only, and not at all mainstream.
Here's my test movie so far.
It goes up to about halfway through World 4 (so it's about 70%-80% done), but I haven't really spent the time to make it really optimized (some of its records on the small levels are slower by a few frames than the best ones I've done) and that's why it's only a test. When I finish it, I'll either submit it (if nobody responds and I feel lazy) or redo it completely (if people give me ideas or I feel motivated).
This is my first TAS, but I spent a lot of time playing this game for speed before the rerecording VBA was released, so even if it's suboptimal, it's not going to be slow.
Take a look - what do you think?
Here is the version so far. I've completed everything in worlds 1-4, and rescued all of the little Kururin guys.
My question is: where do you guys think I should end the movie? Here are the options:
- End now. I've completed the four worlds, and that's all of the normal levels.
- Finish the first secret level and end just before the credits, which are viewed after the first secret level.
- Finish all of the secret levels.
- Get Master Rank (by getting high scores in every Adventure level [the only ones done so far], Challenge level, and minigame).
Fine then. Here is my final test movie.
I finished that first secret level, and ended just before the credits, because that's what I felt like doing. It feels like the ending of the game anyway, even though beating the game unlocks three new levels and I haven't even touched the Challenge levels or the extra minigames.
If nobody responds to this topic by 6:00 PM UTC (that's 1:00 PM in New York, 10:00 AM in California, etc.) on the day after Christmas, I'm going to have to submit this. That seems to be the only way to get people to actually look at this movie. There's no point in redoing the entire thing if nobody has any suggestions on how to improve, is there?
Yes, but there are a bunch of big differences:
- In this game, you can use the R-trigger to speed up your rotation. This means that the mazes can be completed much quicker and there are many more options on how to complete a given maze.
- The levels in this game look better and are generally less linear than those in Kuru Kuru Kururin. The moving obstacles are more dangerous. There are a few levels that have a number of small obstacles scattered about, sometimes giving the player dangerous shortcuts.
- In this game, there are keys hidden in levels. When you get a key, it will open up a part of a level you've already completed, and going through the new exit in that level will give you a whole new level. I get all of the keys because I want to open up all of the levels.
- In Kuru Kuru Kururin, you just go through mazes, but in this game, to rescue your brothers and sisters you must also beat twelve different minigames.
wasn't bad, looked decent, albeit the sound effect at the start of every level was annoying.
i'd likely vote yes if you submitted it, as the game is a lot harder to play than you made it look :P
I just started version 2. I'm going to try to make the level times even better, as well as optimize the minigames a bit. This time around, I'd really like some comments on which minigames looked bad and what I should do, so anyone who's seen the run, please comment here.
So far, I improved (by at least 1 frame each level) the training levels. I also learned that if you don't pick a name on the name select screen it gives you one randomly (so I just pressed Start + A).
Note that the timer in the levels is accurate to the frame (sometimes it goes up 1 and sometimes 2, but it's always the same sequence - generally x"x0, x"x1, x"x3, x"x5, x"x6, x"x8, but also x"97 and x"99 instead of x"96 and x"98), so I can tell exactly how many frames something took (e.g. a time of 2"21 means 133 frames). That's nice. But the minigames timer isn't so useful, so I'll just have to guess there.
But the minigames timer isn't so useful, so I'll just have to guess there.
Or you could hit '.' to use the emulator's built-in frame counter which is always accurate to the frame even when the game doesn't show any timers. (To find out how long a minigame took you can just subtract the count at the start from the count at the end, of course.)
I already have the frame counter on while I'm playing. Should I use frame advance and time exactly when the screen goes to black (after the CLEAR!"), then try to improve that? Sounds good.
I'm keeping the order of the levels the same because there's no reason whatsoever to change them. My improvements thus far (everything that isn't a shortcut is from optimization:
(Training)
01"16 -> 01"15 (1 frame)
01"95 -> 01"93 (1 frame)
02"30 -> 02"20 (6 frames)
02"56 -> 02"53 (2 frames)
02"11 -> 02"10 (1 frame)
(World 1)
03"11 -> 03"11 (0)
02"97 -> 02"95 (1 frame)
07"05 -> 06"90 (9 frames)
08"97 -> 07"80 (70 frames, spin shortcut)
16"95 -> 16"68 (16 frames)
(World 2)
04"51 -> 04"05 (28 frames, small shortcut)
(World 1 unlocked)
11"95 -> 11"66 (17 frames)
15"50 -> 13"78 (103 frames, hit two less springs for spin shortcut)
12"71 -> 12"03 (41 frames)
That's it for now. Here's my WIP:
http://www.mzrg.com/filepost/KP/Kururin_Paradise_v2_part3.vbm
Do you see any obvious improvements? I'd redo most of the above if you found something big (but small optimizations aren't worth it).[/url]
Looks good to me. The only things I can think of are:
- Are you sure all the corners are being cut as closely as possible? For instance around frame 4720, it looks like you could stay closer to that edge (one "pixel" away from taking a hit at each frame) to reach the heart zone a couple frames or so faster.
- When you hit the wall 3 times in that heart zone, it looks like each hit slows down your movement briefly, is it possible to move down a bit more first so you hit the wall only 1 or 2 times a little later, or am I wrong that each hit slows you down? (I know it was also done to affect your spinning but it looks like you could still squeak through with less pushes.)
- I think you should do some stuff with the L button, like putting it on autofire when you're taking a shortcut, or matching colors with the level or something, for what little extra fun you can get out of it.
Are you sure all the corners are being cut as closely as possible? For instance around frame 4720, it looks like you could stay closer to that edge (one "pixel" away from taking a hit at each frame) to reach the heart zone a couple frames or so faster.
No, I'm not sure at all. Some of this is out of laziness (I'm savestating only every 2 seconds, should be doing it every half second or closer - I'll put more work into this on version 2.1 - I should probably start again just in case) and some of this is places where cutting corners close is either impossible or irrelevant. For example, there's a section in the first unlocked level (from world 1) where I turn so slowly that cutting corners wouldn't matter at all. I do agree with you, though, and I'll try to put more work into that.
nitsuja wrote:
When you hit the wall 3 times in that heart zone, it looks like each hit slows down your movement briefly, is it possible to move down a bit more first so you hit the wall only 1 or 2 times a little later, or am I wrong that each hit slows you down? (I know it was also done to affect your spinning but it looks like you could still squeak through with less pushes.)
I've done it with one or two hits before (usually two - one requires you to be too far away from the wall to save any time) but I sort of felt that three would make it faster. But I guess it is true that it slows you down. I'll try a few things and see how the timing works out on that.
nitsuja wrote:
I think you should do some stuff with the L button, like putting it on autofire when you're taking a shortcut, or matching colors with the level or something, for what little extra fun you can get out of it.
Oh yeah! I forgot completely about that. But wouldn't matching with the level make it sort of hard to see sometimes? Hmm... autofire L does make it change insanely fast, and I can always use the savestates to make it stop on the right color. Looks good! I think I'll try it on some of the more insane shortcuts. Maybe. I can also use the autofire to change every time I enter and exit a heart spot (if I feel like it).
Hmm... I just noticed something. If you press R or L on the world map, you'll move an entire world at a time. This might save ten or so frames on map movement throughout... but why not. It looks cooler that way.
OK, so I've redone all of version 2.
Here is my WIP (version 3, movie 4).
The changes are:
- I try to cut corners a bit closer than last time (although it's still not perfect, and I can't really ever expect it to be, especially because there are places where I need to just rotate so I need to introduce some kind of delay). So I did improve the times for a bunch of levels.
- I tried to hit the wall in heart zones as few times as possible (once, usually), and I think I'll continue doing this.
- I have introduced nitsuja's idea of changing the color on entering a level (tell me what you think about this! please!) and I've also set L to autofire on a certain level where I actually do take shortcuts. Note that since most of my shortcuts are going the wrong way through corners and stuff like that, so the real "shortcut"-y shortcuts (going through close-together objects that I'm not supposed to go through). Most of the first few levels are quite linear. Of course, this is quite difficult to do correctly, because I have to turn on the autofire for just the right amount of time while carefully navigating the (usually difficult) shortcut. If you want me to remove the autofire-on-shortcut feature, tell me and I will, because I could probably save an extra two or three frames on that level if I didn't have to worry about that (not to mention the future ones). But I might just keep it in for entertainment value, who knows.
Here's the time breakdown from v2 to v3:
(Training)
01"15 -> 01"15 (0 frames)
01"93 -> 01"93 (0 frames)
02"20 -> 02"20 (0 frames)
02"53 -> 02"51 (1 frame)
02"10 -> 02"10 (0 frames)
(World 1)
03"11 -> 03"10 (1 frame)
02"95 -> 02"95 (0 frames)
06"90 -> 06"86 (2 frames)
07"80 -> 07"76 (2 frames)
16"68 -> 16"53 (9 frames)
(World 2)
04"05 -> 04"03 (1 frame)
(World 1 unlocked)
11"66 -> 11"58 (5 frames)
13"78 -> 13"46 (19 frames)
12"03 -> 11"95 (5 frames)
Here is my WIP version 3 movie 5.
I hexedited out the flashing colors on the shortcut (removed the L presses from the movie file) because I decided I didn't like it. The absence of randomness in this game meant no desyncs. I think L presses (except on the world map) are irrelevant to your player's and the enemies' movement.
Take a look around frame 20350 - this shortcut was really hard to get and took about 25 minutes (even with frame advance), so I really don't want to reoptimize the beginning of the level. If anyone ever wants to try to redo this shortcut, kudos to whoever can accomplish it, but I don't think I ever want to have to do it again. Grr. On the plus side, it saves 60 frames. Maybe there are similar shortcuts in other places? (I hope not.)
Tell me what you think! All of the levels are being improved, anyway. Note that I'm going for in-level time first and game time second. I'm still trying to take no damage in the run.
A few more levels done, a few more improvements.
Here is WIP version 3 movie 9.
Almost completed up to the third world.
I think I'm going to make two files:
- One file ends at the credits, because people on this site seem to like credits.
- One file ends after completing the last secret level (demonstrates record times for the secret levels too, you see).
I don't know which one you all will want to watch - I prefer the second but I'll probably submit the first.
Anyway, I've been trying to optimize minigames a bit, but since the minigames are so long and complicated it's more difficult to optimize them than I feel is necessary. I simply don't have the time for utter perfection. Also, note that my game time is 1st priority and real time is 2nd. I'm trying to stick to this goal as much as I can.
Emulator Coder, Site Developer, Site Owner, Expert player
(3599)
Joined: 11/3/2004
Posts: 4739
Location: Tennessee
Are you planning to finish this run, and do you plan to cancel the current submitted run?
I've watched the submitted one, and I highly question the decision to do minigames/secret levels or whatever. This game is a bit too repetitive. Keeping it short is the key.
Are you planning to finish this run, and do you plan to cancel the current submitted run?
It's on indefinite hold for now. I have a lot of stuff to do, and I'm still not sure if I should end at the credits or go to the secret levels.
adelikat wrote:
I've watched the submitted one, and I highly question the decision to do minigames/secret levels or whatever. This game is a bit too repetitive. Keeping it short is the key.
Yes, but here's the thing: you only see the credits screen when you save all of your little friends (siblings?) and then beat the first secret level, and the only way to save them is to complete the minigames. Unless there's a glitch that will bypass this, the only way to achieve any kind of signal that you've finished the game (other than simply doing all of the normal levels, but that would be stupid) is to save them all, and that means completing every minigame. It may be unfortunate, but unless there's a faster way of getting to the credits (where we like to end movies, right?) I have to go through the repetition.
