Tool-assisted game movies
When human skills are just not enough

Submission #4877: Samsara's PCE Splatterhouse in 12:15.72

Console: TurboGrafx 16
Game name: Splatterhouse
Game version: any
ROM filename: Splatterhouse (U).pce
Emulator: BizHawk 1.11.2
Movie length: 12:15.72
FrameCount: 44015
Re-record count: 9795
Author's real name: SJC
Author's nickname: Samsara
Submitter: Samsara
Submitted at: 2015-10-13 10:55:19
Text last edited at: 2020-07-12 00:26:59
Text last edited by: Samsara
Download: Download (11656 bytes)
Status: published
Click to view the actual publication
Submission instructions
Discuss this submission (also rating / voting)
List all submissions by this submitter
List pages on this site that refer to this submission
View submission text history
Back to the submission list
Author's comments and explanations:
Scrollerhouse is the scrolliest game since The Elder Scrolls. Somehow, "scroll" stopped looking like a word, then looked like a word again, and now doesn't look like a word anymore. And now it does again. Huh. How about that.

Game objectives

  • Emulator used: BizHawk 1.11.2 (started in BizHawk 1.8.1)
  • Aims for fastest time
  • Takes damage to save time
  • Splatters houses
  • Oh my (RELIGIOUS MENTION REMOVED IN AMERICA), it syncs on both (U) and (J)!

Special Dual-Version Encode!

It's almost like you're watching an AngerFist or agwawaf run, except my name doesn't start with A and this is way less interesting!

(Link to video)


Hey look, all the trademarks of a recent Samsara run are here! It was started a year ago, it's another Splatterhouse game, and the submission comments are incredibly cheap and lazy because I ran out of jokes around the time I was writing for my first Run Saber submission. Oh well.

Splatterhouse is the first game in the Splatterhouse series. It's not to be confused with Splatterhouse, which is the latest game in the Splatterhouse series. This is the TG-16 home port of the superior (but way less TASable) Arcade version, with very tinkly music and a lot of removed religious references. It's pretty great, if only because it's so self-aware about what it is as a series: Pixel gore and horror movie references!

So yes, this was technically started September of last year, when I created a quick test run to see how TASable the game was. After that, I started working on an optimized version, got through Stage 5 and stalled out for a good year right about there, but we'll get to that when we... well... get to that.

I don't want to say it was a rough road, because (another trademark of my recent runs) I spent most of the time not working on it and doing other things, but I'm still glad I worked through the one major roadblock I had and improved my test run by a little over 6 seconds.

I may do the other two Splatterhouse games I haven't done yet. Splatterhouse 2 is pretty much exactly the same as this game, except different, while Splatterhouse 3 will be a challenge should I ever decide to take it on. But hey, it's an option!

Oh Boy, Difficulty

So this game does have a Hard difficulty that I chose not to use, because all it changes is enemy/boss health and maybe enemy speed. The boss fights are very much pattern-based and aren't all that interesting to watch, so I decided against prolonging them. As for enemy speed, it's basically the same situation as Run Saber: It's not noticeable at all.

A Couple Things to Note

Movement Speeds: Walking is 1.375 pixels per frame (11 pixels per 8 frames), jumping is 1.5 ppf (12 pixels per 8 frames), and sliding is 1.875 ppf (15 pixels per 8 frames).

Attacks: Rick's normal punches and kicks do 1 damage, while the slide kick does 2 per hit. The slide kick can hit multiple times, and thus is always used when possible to cause the most damage to bosses.

Stage by stage comments

Stage 1

The first stage is fairly straightforward, though it showcases the most oft-used and important trick in the run. Jumping on the first possible frame after a slide will keep your slide speed throughout the jump, which can be chained into another slide in turn. This keeps Rick's movement at a breezy 15 pixels per 8 frames throughout most of the run, saving quite a lot of time in the process.

Jumping routes have to be planned out in advance to ensure that Rick is able to avoid obstacles and enemies sufficiently quickly. It was not uncommon over the course of the run that I'd get through most of a stage, only to find out there was no way I could avoid something without losing time, meaning I'd have to go back and find another jump pattern. Enemy placements are fixed, but it's still tricky to plan for them when you're moving faster than intended and constantly jumping.

The ladder at the end of the stage is just a timer. It doesn't matter how high or low I grab it, the fadeout always happens at the same time.


Aptly named. The only one that needs to be killed is the final one: This is basically an autoscroller that doesn't scroll up until the last one bursts out of the hanging corpse. Much like the facehuggers in Wanpaku Graffiti, I elected to not kill any of the boreworms to make it more interesting to watch. Did it work? You be the judge of that! Because I can't. This isn't Combatribes.

Stage 2

I'm linking the Arcade version's song here, because of how drastically different it sounds when adapted to TG-16. It's probably my favorite song discrepancy between the two versions. The TG-16 version is so tinkly and cute!

Songs aside, this is another level that required some serious jump route planning to get through it quickly. The end of the first part is a bit awkward, but I couldn't pull off anything faster. The last hanging corpse is just too in the way to be able to jump into the hole. It takes two hits and regains its hitbox pretty quickly. Fun fact: Delaying starting the game for an odd number of frames causes Rick to get hit by this guy. It's so in the way that it desyncs in an alternate universe.

The sewer autoscroller, which is a combination of words that probably just made at least one of you reflexively gag, is ripped directly from my testrun. I tried to sync every single sound to the music, mostly succeeded, and just copied it over because it's an autoscroller. I did optimize the end a bit over the testrun, though, so you can't accuse me being TOO lazy.


It's possible to break every phase of this fight, which you know is the hallmark of a great boss. I abuse the game engine a bit on the chair phase in an almost unexplainable way: When you hit the chair, it normally flies to the opposite wall, then starts bouncing toward you. In this case... well, that doesn't happen. It hits the wall (since I'm technically kicking it to the right) and bounces right over and despawns on the other side of the wall, saving time. The hit I take is necessary, as I need to hit the chair at the peak of its jump and I can only do that by being inside of it.

There's a way to despawn the knives by making them fall through the floor, but the only evidence of this is a single video and I just couldn't get it to work. It doesn't look like it'd save any time, unfortunately, but it'd be cool to try to figure out. In this case, I just hit them once and slide into them to take out their 3 HP each. The two-knife part is pretty precise. Watch out for that. The painting, for whatever reason, dies instantly to slide damage even though it has more than 2 HP, so I simply slide into it to end the fight.

Stage 3

The beginning section not only required serious planning, but it required the discovery of a new trick compared to the testrun to be able to work. Rick doesn't fully keep his momentum with frame-perfect jumping, which is why I need to slide after every jump to maintain top speed. You normally lose 2 pixels per 8 frames on every frame perfect jump: Jumping twice will actually put you at walking speed, which is slower than jumping speed. However, if you crouch three frames before you're normally first able to jump, you can jump a frame earlier, and you only lose 1 pixel per 8 frames this way. This is used to keep Rick's speed up over the bridge in order to avoid the holes (which lead to a long, unnecessary autoscroller) while not having to slow down or make awkward jumps. Finding that was pretty big, even if it doesn't save too much time. The layout of the level also requires the crouch-jump trick in order to save as much time as possible.

Biggy Man

Straightforward fight. He has 20 health, and the shotgun/slide do the same amount of damage. 7 shotgun blasts followed by a 3 hit slide is the fastest way to take him down, and ending at the far right of the screen is, of course, faster, since Rick is scripted to walk off. Unlike Wanpaku Graffiti, you can't trick the shotgun into recoiling you forward while still firing forward, which is a shame. But there's that.

Stage 4

Rick has an obsession with harpoons in this first autoscroller. Keeping the harpoons with me for entertainment required me to kill most of the enemies getting in the way to avoid slowdown, otherwise I would've gone full pacifist here as well.

The mirror section has great music. Rick acts like a scaredy-cat. Unfortunately, the game doesn't handle as well as Wanpaku Graffiti in terms of how Rick moves, so it's harder to give him "character" in these autoscroller sequences. But I did my best... a year ago... Look out for the fistbump that I'm definitely not using as a screenshot suggestion. How dare you accuse me of that.

Evil Head/Cross

Okay. Bear with me, here.

This is the biggest disappointment of the run for me, and one of the oddest decisions I've ever had to make regarding one of my TASes. You'll notice that shortly after collecting the golden cleaver, I stop, pause for a moment, and start walking normally. There's a reason for this that requires a bit of a lengthy explanation, so here we go.

The Evil Head has a quick kill that's utilized in this run. Getting this quick kill relies entirely on Evil Head's position when you start to fight it. Normally, when you reach the end of the scrolling section, Evil Head will begin to arc over Rick going left to right. This is a horribly slow pattern, of course. However, it doesn't begin this pattern immediately: It actually starts the arcing pattern once it reaches a very specific point in its initial up/down cycle, around 1/5 of the way up from the bottom of the screen.

For whatever reason, hitting Evil Head while it's in the up/down cycle will cause it to pause in place while it's hurt, then immediately snap to the point it would have been in its cycle had you not hit it. Putting two and two together, this means that you can hit Evil Head before it reaches the pattern-switching point, and it'll jump past it without changing patterns. This is how the quick kill works: Evil Head isn't allowed to change patterns, so it remains on the right side of the screen within easy reach.

There's just one small problem: Evil Head has to be travelling upward when you reach the end of the scrolling. You need enough time to actually get over to it and hit it before it changes patterns, and you can't get that if he's travelling downward. Had I kept top speed throughout the entire scrolling section, I would not be able to perform the quick kill, which wastes far more time than what I lost from the normal walking. This is the earliest cycle that can be reached, and had I reached the end of the scrolling section a frame earlier, it would have immediately launched into the arcing pattern.

Strangely, slides mixed with normal jumps allow you to reach Evil Head at the perfect time. I could have done that (i.e, copied it from my test run), but something told me that it would seem lazy if I mentioned I'd done that, so I did the entire boss over and ended up saving a little bit of time. Glad I made that choice, even if it does look weird to suddenly stop and not move as fast.

Stage 5


The second room of Stage 5 is probably my favorite part of the run. It's normally a very difficult and awkward room with all the slopes and such, but I found ways to bypass nearly all of the awkwardness. I slide on the backward slopes so I don't get pushed back, and I jump normally on the forward slopes since they actually let me travel at 2 pixels per frame, the highest speed Rick can reach. The room went very well, especially when compared to my test run.

Get ready for some autoscrollers.

Another sewer scroller is first, though it's much shorter. After that is a section that has a gimmick I completely skip: The floating pumpkinhead dude revives those zombies if you let him float around for long enough, but this is a TAS so I don't give it any time to float around at all. Rick fits into a tiny hole at the end of this section, a hole that leads to another autoscroller! It's there. I like it I guess.


Reminds me of my ex-girlfriend, really.

This is an incredibly technical boss fight, not that you'd notice if you've never played the game before. Jennifer has a gimmick where she transforms back into a human at certain damage values. These values are very specific, and the transformation sequence is twice as long as the one before the fight. This can happen multiple times in the fight and obviously, it's a massive waste of time to let one happen.

Thankfully, since she's coded to transform at certain, specific HP values, we can actually skip all the usual transformations with a very specific pattern of normal hits and slide attacks. A well-timed slide attack will pass over the values, preventing her from transforming and saving huge amounts of time over a "normal" fight.

Damage is taken during the fight to allow this specific pattern to continue, and to make sure she dies on the right side of the screen. Like Stage 3, Rick is scripted to walk off at the end, so ending on the right is the best course of option.

Shoutouts to zallard for the knee drop at the end of the fight, because it actually saves 2 frames while simultaneously ruining the closest thing this game has to a sentimental scene. Truly, this is art.

Stage 6 (+ Mother)

Everything from here on out was done very recently, so I know what I'm talking about now!

This was my biggest roadblock preventing me from completing the run back in 2014. It's the only RNG in the entire game and to this day, I have no idea how to manipulate it. My initial pass through the stage ended in bitter failure: The boss pattern in particular was simply not conducive to speed. The spawns were ridiculous and always got in the way, and no matter how much I tried I always lost more time than I had saved over my test run. I eventually dropped the project, vowing to come back at a later date when TAStudio was functional enough to allow me to test things quickly and easily.

Cut to recently, with BizHawk 1.11.2's release. TAStudio hadn't played nicely with this game until 1.11.2 came along. In 1.11.1, loading TAStudio would slow the game to 1-2 frames per second around Stage 4, rendering it impossible to use. 1.11.2 fixed that and I was able to get back to testing patterns. Nothing worked! So, on a whim, I decided to just go ahead and see how much of the stage I could salvage.

To my surprise, I ended up saving over a second and a half compared to my test run. In my test run, I had apparently missed the first opportunity to attack the boss by over a second. This was repeated in my initial attempts in the optimized run, and that was what lead to me thinking that the pattern was bad. It's actually really nice, though I still had to take a forced hit (but that doesn't lose me too much time).

I ended up running through the level a few times recently, the first few times to get a feel for the spawn pattern and to test out what I could do on the boss, and then the last time around was to add the entertainment and fully optimize the boss fight. The optimized boss fight here was 30 frames faster than my last test. I feel pretty good about it, to say the least!

One note about the stage: Those spawn will keep... spawning. There's no sprite limit or anything. You can have 20 of the little buggers jumping around, slowing the game to a halt. It's pretty great, but it does mean I have to kill as many off as possible, especially for the boss fight, since they'll get in the way if I don't.

One note about Mother: It has 32 HP and can only be hit twice by the slide due to a longer invulnerability time. The pattern employed is to slide into it, then jump left with a kick, and repeat. Usually does 5 damage a cycle, but often the spawn will get in the way during the jump kick. It ended up looking smooth in the end.

Less smooth is the graphical glitch I found by accident during my boss testing. It's frame precise and kinda funny I guess. Yep.

Stage 7

Damage boost city, dogg. Took me quite a few tries to come up with the boost strategy seen here, and this one is only (technically) 1 frame better than the previous best I found! Lots of weird positioning is needed in order to get damage boosted forward, and there was nothing I could do about the positioning in a lot of cases so it looks a little bit awkward at times, but... yeah, nothing I could really do about that.

Hell Chaos

Also known as Captain Mozzarella, apparently. But Hell Chaos. The baddest-ass of names.

Hell Chaos, apart from being the title of my next album, is a pretty standard fight. On its first phase, you can take off 9 HP before it retracts back into the ground. Every phase after that, it only takes 8 HP. This can be circumvented with carefully managing its health (like I did with Jennifer), but it leads to a slower fight in the end.

There were a couple alternate strategies discussed in the game thread that ended up not working: I couldn't, for example, slide into Hell Chaos, then jump kick and slide again. Its invulnerability time is too short to be able to pull that off without getting hit, and obviously I can't get hit in this fight. So I'm forced to stick with the same strategy I used in my test run: Slide for 3 hits, then two normal attacks (three on the first phase).

I end input early on the last phase by tagging it with a jump kick first, then inputting a slide to take off the last 6 HP. And that ends the run.

Special Thanks

  • First and foremost, thanks to zallard for his useful posts in the thread and his great RTA runs of this game
  • Usual shoutouts to IRC for being even chiller to me now for some reason. I wonder why.
  • Word to your mothers

Screenshot Suggestions

Frame 18832:

Frame 40534:

Samsara: Quick note to publishers - Despite the run technically being done on (U), I would prefer the primary encodes to use the (J) version if possible, purely because of the lack of censorship. Once again, no input changes are required, as the file syncs perfectly on both versions since the only changes between the two are cosmetic. Apologies for the weird nonsense.

ars4326: Claimed for judging!

ars4326: Hello again, Samsara. Overall, this run kept up a good, smoothly executed pace (enjoyed the dual encode, too!); most especially due to maintaining Rick's slide momentum speed, throughout. And as usual, your "trademark" style of incorporating bits of creative movement choreography added considerably to the run's entertainment value (e.g. Stage 4 and the hall of mirrors). Well done work, once again!

Accepting for publication to Moons!

fsvgm777: Processing.

Similar submissions (by title and categories where applicable):