This is an unlicensed demake of Donkey Kong Country (1994), ported to the NES called "Donkey Kong Country 4" by Hummer Team. The game is mainly based off of the base game, with visuals being hand drawn to lower quality and music reproduced by ear. Amusingly, it bears a copyright symbol in the opening credits, irony notwithstanding.
The movie consists of optimizing platforming, keeping in mind that all enemies with the exception of bosses, are all on local cycles from when they are spawned.
Game objectives
- Emulator used: FCEUX 2.2.3
- From wikipedia: The game follows the gorilla Donkey Kong and his nephew Diddy Kong as they set out to recover their stolen banana hoard from King K. Rool and the Kremlings
Some emulators set the dipswitches of this game to "either", on FCEUX 2.2.3 it defaults to DKC4, so make sure that you set the dip switch correctly if using a different ROM or emulator. Otherwise the file will instantly desync.
Running is the fastest method of movement, and all optimization is to ensure that the kongs are running most of the time. Jumps are cut to the shortest amount they can be without missing, jumping on enemies is put to a minimum and barrels are thrown away immediately.
Slopes misbehave in the game, sending the kong a pixel forward or back when jumping, and can easily be a source of wasted frames. Hopefully in this submission that is reduced to nil.
Barrels shoot you out according to where you entered. If you entered slightly to the left, the barrel will shoot you out slightly to the left, this is used in some of the barrel sections.
Diddy appears to have faster base speed on land and Donkey in water. Different Kong water speeds is used to optimize pathing in water levels.
There is little lag in the game, but where it does exist, such snakes in industry levels, it has been attempted to be minimized by killing snakes as fast as possible and picking up bananas to reduce the sprite count.
At the end of the movie I spell out crazyjesse using morse code in my downtime.
The best comment I received about this game was doing a speedrun event when someone in the chat said something along the lines of "Why does Jesse know how to speedrun this game and not the regular game? Does he hate 16 bit colours?".
Samsara: I'd ask you to pull out your
Samsara judgement bingo cards to mark off the "Samsara calls judgement 'interesting'" square but at this point you've all marked that square so much that I don't know if the square is even readable anymore. The next version of the bingo cards will just have that square free as well. Also, it'll ensure Samsara finishes her jokes in her submission texts, instead of leaving them completely open and not noticing for several hours.
The issue here lies in optimization this time. Most of this run looks optimized, however there are two stages that stand out: The swimming stages. There is a zip used in the second swimming stage that, upon my investigation, can be used far more effectively throughout both of these stages to save a solid chunk of time. After a little more investigation, alongside being clued in to a clip discovered after the creation of the run, I discovered that these can be used in tandem to save
nearly 1000 frames on the second swimming section alone.
This makes things... Well, interesting. By itself, the clip may not even save any time, but combined with the ceiling zip that was known at the time of this run's creation it turns into a single massive improvement, so it ends up being kind of a weird choice to make regarding whether or not the run as a whole should be rejected for suboptimal play. Without the clip (which, again, was discovered after the creation of this run), the improvements from zipping would be much less substantial, likely only a handful of seconds per stage as opposed to that nearly 17 second improvement. Given the rest of the run's optimization, I wouldn't consider that reject-worthy, but there was something else brought up that's leading me to the decision I'm making here. According to the
linked VOD with the floor clip in the swimming stage, this run is a couple years old (clip is dated from early 2019 and this TAS predates it), and it was apparently not originally intended to be submitted. That leads me to believe that there's likely more improvements hiding around the run that just aren't obvious on plain viewing, something that a future improvement can take more advantage of. Given that, and given that there's now an estimated 20 seconds of known improvement just in the swimming stages, I'm rejecting this for now. It's INCREDIBLY close to acceptable, to the point where I'm not even convinced I'm making the right decision here. It could possibly be given a second look in the future, but right now it's hard to overlook the fact that known tricks weren't used to their fullest potential.
In theory it should be possible in every location like the above (this is in the first swimming section, around frame 3400 in the submission), though for an optimized TAS I would check every single ceiling along the route even if it doesn't look like you'd be able to get in. It might not work immediately: You'll probably have to adjust when you press A to swim upward alongside holding Up or Down to further fine tune the position. Once you're in the ceiling proper, you'll want to hold Up the entire time, and find the frame where pressing A zips you forward. If you can't find that frame, go back to adjusting your position: There may be a frame where you get zipped downward out of the ceiling entirely, repositioning a bit can fix that. Once you can zip, A can be pressed every 4 frames from there to continue zipping at top speed, it should look like this in TAS Editor when done optimally (hyperlinked to image because it's a bit tall).
