Introduction
Art Style: PiCTOBiTS, known as Art Style: PiCTOPiCT outside the US, is part of the DSiWare and WiiWare line of Art Style games. This is a puzzle game developed by Skip Ltd., the Chibi-Robo! devs, where the goal is to match colored bits in rectangles of at least 4 bits in size. The main goal of this TAS is to clear all 15 normal stages with in-game time. This game is still available in the 3DS eShop until the online closure that's coming March 2023.
Run notes
- Emulator used: BizHawk 2.8
- melonDS core
- DSi mode
- Aims for in-game time
- Does not color a dinosaur
Important technical info
Much like the other DSi submission to this site this one uses a PWC 2010 firmware that was formatted before the start of the movie. Not sure how important this is but I used melonDS version 0.9.4 to perform the system formatting. Once that's done I then moved the formatted firmware to BizHawk.
So why the long pauses in the middle?
When I discovered better times out there, I mainly improved the individual stages that saved in-game time. Since RNG is real time clock dependent and I don't have enough energy to work on better times for stages that weren't beaten by RTA play, there are gaps where nothing happens until the next stages start. Most notably, due to big time saves in stages 1 to 4, stage 5 has over a minute of waiting before starting. I do apologize but that's what going for in-game time get for me.
Mechanics and techniques
Chains and colors
Combos and color combo orders are the bread and butter for fast completion. Bigger chains means the number of bits added to the picture is multiplied by the chain number, up to 9. How this can be used is dependent on the color situation, though in general the colors that need more bits are used near the end of combo setups.
Some combos are built from using a falling bit shape in a way more complex than the simplest approach. The tradeoff is that it means spending more of the initial bits which does lead to problems down the road. However more bits can be made available by having the falling bits stop and split into individual, usable bits as long as they aren't marked. Marked bits cannot be moved by the player unless they spend one of the held bit slots to make them drop down, which takes a fair amount of time to play out as well as reduce the max number of held bits at once.
Falling bit formation drops
I don't drop some falling bit formations immediately since doing so would cause it to separate into split parts. It's better to wait if the intention is to make bigger combos. Sometimes I do stop bit formations to make more usable individual bits for later since that allows for clearing the next formation with larger combos.
Building off bit formations
Placing bits onto falling bit formations that haven't been fast dropped is a useful way to make a combo with a color that otherwise isn't available yet. While this resource intensive, it's very useful for stages with fewer colors. More colors means it takes more balance to get this to work.
Random number generation
RNG seems to be determined by the real time clock. This determines initial bit layout and some of the upcoming falling bit sets.
Earlier stages could be improved but that might break a lot of the later stage RNG seeds. Possible to set a better initial time for overall ideal RNG. The date I used was just the US release date of the game in the DSiWare store.
Delaying drops for manipulation
Sometimes getting bits delays the next falling bit pattern. This can be used to delay enough such that the RNG pulls from a later bit of time. Usually bit patterns are long enough that manipulation has to be done earlier i.e. before starting a stage but it can be helpful.
Stage in-game times
I sure hope I got that total time right.
| Stage number | In-game time |
|---|---|
| 1 | 0:07.86 |
| 2 | 0:09.50 |
| 3 | 0:21.91 |
| 4 | 0:06.43 |
| 5 | 0:33.56 |
| 6 | 0:12.63 |
| 7 | 0:06.56 |
| 8 | 0:24.05 |
| 9 | 0:13.26 |
| 10 | 0:12.88 |
| 11 | 0:26.26 |
| 12 | 0:11.55 |
| 13 | 0:14.50 |
| 14 | 0:13.25 |
| 15 | 0:03.56 |
| Total time | 3:37.76 |
Possible improvements
There is definitely some RNG seed settable by the real time clock that might make all the times outside of stage 15 faster. I only discovered the really fast in-game times from TokusiN much later in doing this TAS. I could have worked with earlier seeds but it would require having to replan pretty much every stage to meet or beat the first pass, so I went with waiting out for the right seen which I don't like but works okay enough. Some RNG manipulating loremaster would be better off just finding a more ideal RNG seed that would just cut times across the board, as well as cut down on the waiting between stages.
That being said, stage 5 I feel there's a better overall RNG seed for both initial available colors and falling bit formation patterns. I really hate that dang condor so much.
Special thanks
- Walgrey for having a full normal stages run to work as a starting reference point of what can be done
- TokusiN for their really fast RTA strats for stages 1 (though modified) , 4, and 15, along with having a lot of useful in-game times to reference
- CasualPokePlayer for the DSi BizHawk implementation, making this reasonable to produce
- Arisotura, for developing melonDS and working implementation of DSi functions and DSiWare
- Fortranm, for the first published DSi movie to TASVideos which helped on setup
Suggested screenshots
- Frame 13592, with OKAY!!! at the bottom of stage 5
- Frame 21064, full screen combo with no remaining bits in stage 9
- Frame 24549, big smiley face at the bottom of stage 11
CasualPokePlayer: Claiming for judging.
CasualPokePlayer: The TAS seems very well optimized, for in game time. This seems fine in this case, as that optimizes for the main gameplay time. The TAS syncs fine with author provided sync instructions.
Accepting.
fsvgm777: Processing.