Submission Text Full Submission Page
  • Emulator: FCEUX 2.1.2, New PPU
  • Aims to complete each act as fast as possible
  • Abuses programming errors in the game
  • Disregards floors and ceilings
  • Shows that Mario can actually stomp enemies underwater
 
 

About This Run

Somari is a Hong Kong pirate (unlicensed, not a hack) game that takes Mario and puts him in Sonic's world on the NES. It is an extremely difficult, glitchy, hard to control and colorful game. The levels and badniks are modeled after Sonic 1, with the addition of the spindash from Sonic 2. This game is not only difficult, it takes a special person to get through this game unassisted and not have ones ears and eyes bleeding. Try it yourself if you think you have what it takes.
The main goal of the run is to get through levels as fast as possible. I disregard the score tally screens, which unlike most Sonic runs, count rings separately from the time bonus. Some total frames can be saved by sacrificing level time to decrease score tally screen time.

Some technical info and quirks about this game

Speed in this game is basic integer speed. There are two speed values for Somari, which are an internal speed value and the screen speed value. Max normal speed is 5px/frame, although zipping is possible at 7px/frame. After passing certain internal speed threshholds, Somari's screen speed instantly increases or decreases to the next value, which is why at times Somari tends to have a large visible jump in speed. These threshhold points are abused heavily in the run.
All jumps are of a fixed height (unless a ceiling is hit). Speed can't normally be gained in jumps. However, in some cases, touching ceilings, corners or collecting rings can cause the speed value to change (such as increase if you are holding a direction). This is abused heavily throughout the run.
Spindashing and rolling sucks in Somari, because it can't be interrupted. You can't jump or cancel out of it - you have to wait until you slow down to 0 speed (either naturally or hitting a wall) before you stop rolling. However, when coming out of a spindash or a roll, you can start walking and accelerating, even in air. Because of this, spindashes and rolls aren't used much in this run, but it is sometimes used when speed is very low and an enemy needs to be killed, and where taking damage takes more time than the acceleration gained by being able to walk in air.
Somari is limited to 1px/frame max speed in the water. However, acceleration and the internal speed works the same in the water as well as out of it. As soon as Somari leaves the water, he will transition to above water speeds depending on the internal speed (so, if internal speed is 246 and Somari is going 1px/frame underwater, if he jumps out he will go 5px/frame, since a speed of 246 is within the 5px/frame thresshold).
When running off a ledge, Somari will continue to accelerate.
Many things in the game are based on a 60 frame timer (and likely some other timers I haven't found), such as spikes, Orbinauts, and the delays in the arrival and leaving of the score tally and title cards. Enemies are usually based on Somari's position, but can also be affected by if they spawn on an even or odd frame, or position within this timer. Because of this, I delay about 120 frames at the end of Spring Yard Zone 3 to set up spikes and other objects to be optimal and end up with a net 230 savings in Labyrinth Zone 1.
The score tally screen gives points for rings and a time bonus. However, the time bonus isn't actually implemented so it always gives 5000 points. The ring bonus takes 4 frames per ring to count. Again, the goal of this run is fastest level time, not fastest overall movie input time, so I list a SCORE TALLY after each act and possibility of saving frames specifically on this screen.

Flygon started this run but I found a lot of optimizations (it wasn't unusual for me to shave off 15%+ of the frames from some of his levels) so I basically took it over. He ran a lot of the levels first and then I'd come through and basically clean up. I don't think he really cared since the game got done and we both were able to see and find some neat glitches.
Act by act follows.
 

GHZ1

Typical fast level. I beat Randil's run from power on by about 16 frames.
SCORE TALLY: Lots of frames can be saved at the score tally if smarter ring collection is used and if Somari is hit near the end.
 

GHZ2

Speed doesn't really pick up until the second half of the level.
SCORE TALLY: Lots of frames can be saved at the score tally if smarter ring collection is used and if Somari is hit near the end.
 

GHZ3

Typical level. Going the top route is quicker mostly due to badnik placement and the ability to gain speed. My delays around the moving platform don't waste time since I have to wait for it anyway - if I don't wait then I have to make two jumps which is slower than waiting. It might be able to be jumped on a couple frames earlier, but I couldn't get everything to match time-wise.
SCORE TALLY: The rings could probably be lost at the boss without losing much, if any, time. This would save, at most, 244 frames.
 

MZ1

I skip about a third of this level by jumping through the ground and zipping.
SCORE TALLY: Some frames can be saved at the score tally if I get hit near the end (maybe the spikes?).
 

MZ2

I realize after the run is done that dropping down at the beginning isn't optimal, but due to later spikes it's doubtful I could get more than a couple frames from fixing this. Somari can't accelerate in lava but he maintains his speed, so when I get damaged I get up to 3px/frame speed before running in the lava.
SCORE TALLY: If frames can be saved here, it wouldn't be many.
 

MZ3

Pass through the floor at the beginning. Getting the third hit on the boss is very difficult - there is a one frame window to get the third hit in.
SCORE TALLY: 104 frames can be saved at the score tally if rings are lost at the boss.
 

SYZ1

There's really no way through the underground part without taking damage, so I optimize it.
SCORE TALLY: Some frames can be saved at the score tally if I get hit near the end.

SYZ2

I go backwards to gain speed to set the timings for the buzzbomber and crabmeats. I take damage at the crabmeat near the end because I can't really find a way past it with the timings that I was given on this level. The only way around it is possibly to delay some frames at the beginning or go significantly faster on previous levels.
SCORE TALLY: If frames can be saved here, it wouldn't be many.
 

SYZ3

I delay about 120 frames here to optimize timings on LZ1, which ends up saving 235 frames.
SCORE TALLY: Maybe hitting the boss is possible? If so, 88 frames can be saved.
 

LZ1

Most annoying level... second longest time-wise. Somari headstomps a Grounder :) There is a ~13 frame wait near the end of the level to wait for the Orbinauts to change their spikeball timings so I can get the most speed out of some areas that act as if you are not underwater.
SCORE TALLY: Some frames can be saved at the score tally if I get hit near the end, or if better ring management is done.
 

LZ2

Yay glitches! I guess they forgot to program crushing... If Somari doesn't move, nothing can hurt him :)
SCORE TALLY: It is doubtful many frames could be saved here.
 

LZ3

I have to get the air bubble... I can almost make it to the top without it, but I need to save another 60 plus frames and don't really have a way to do it. If I can save these frames in the last underwater area, then I can save over 600 frames total. Oh, yeah - I skip Robotnik :)
SCORE TALLY: Lots of frames can be saved at the score tally if I get hit near the end.
 

SLZ1

FAST! YaY SPEED
SCORE TALLY: Some frames can be saved at the score tally if I get hit near the end.
 

SLZ2

I slow down on the downhill parts because traction is lost otherwise and you run in air, where you can't jump. I use this traction to jump over the loops.
SCORE TALLY: 8 frames can be saved at the score tally if the rings are jumped over at the end, although some frames might be lost due to the avoidance of the rings.
 

SLZ3

Maybe the Bomb + Buzzbomber spring can be done better? Depending on timings, the Orbinaut near the end can be run underneath.
SCORE TALLY: No improvement possible =]
 

Final Zone

Robotnik can be hit twice per round in this version. Internally the variable is set to six hits but he only really requires five.
 

The End.

I WILL BE BACK


Nach: I am rejecting this movie, since the author of it did not really try to optimize it, instead opting to make excuses and lie that it was only stylistic choices to look sloppy, while not actually being sloppy, and blaming the game for any perceived problems.
This run ignored basic TAS principles such as: avoid items you do not need, try to get rid of items you picked up that affect the end of level counter, and take damage to save time.
Here is a movie file that is based on the submitted movie and completes Marble Zone Act 3 on frame 26410 whereas the submitted movie does so on frame 26514 which is 104 frames slower, than the movie I made. My run was made without the slightest of effort, which shows clearly the submitted run is the same.
Note to TASers who do want to give this game a shot: This is probably a bad game choice.


TASVideoAgent
They/Them
Moderator
Joined: 8/3/2004
Posts: 15630
Location: 127.0.0.1
This topic is for the purpose of discussing #2544: True's NES Somari in 24:14.09
Post subject: My avatar is pornographic according to Wikipedia
Joined: 11/4/2007
Posts: 1772
Location: Australia, Victoria
I vote Yes. I will be back... with an encode, and a suggested screenshot. true said the grounder stomping, I say the Labyrinth glitch... decisions, decisions. Glitch: Grounder: Coming soon! Already optimized. Archive.org (Still encoding) DailyMotion (Still uploading) For those wondering, I'm chucking a lossless version of the encode to DailyMotion, which encodes much faster.
sgrunt
He/Him
Emulator Coder, Former player
Joined: 10/28/2007
Posts: 1360
Location: The dark horror in the back of your mind
Oh yeah! I'm glad to see this in the queue finally. I'll summarise the comments I made to true on IRC earlier: This is a relatively slow-paced game. However, it's also insanely difficult, and the sheer glitch factor makes this run accessible to people who aren't familiar with the game. Even though the music makes my ears bleed, I think this deserves a Yes vote.
Editor, Emulator Coder, Expert player (2159)
Joined: 5/22/2007
Posts: 1134
Location: Glitchvania
So you two publishers again. EDIT: I just now found something.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days <adelikat> no doubt <adelikat> klmz, they still do
NitroGenesis
He/Him
Editor, Experienced player (556)
Joined: 12/24/2009
Posts: 1873
I did not know we accepted pirates. BUT JEEZ! How glitchy. AND entertaining. Nice job making a slow paced game fun to watch. I am going to vote yes.
YoungJ1997lol wrote:
Normally i would say Yes, but thennI thought "its not the same hack" so ill stick with meh.
NitroGenesis
He/Him
Editor, Experienced player (556)
Joined: 12/24/2009
Posts: 1873
Oh no! you posted a rom link in the description!!! EEEK
YoungJ1997lol wrote:
Normally i would say Yes, but thennI thought "its not the same hack" so ill stick with meh.
Joined: 11/4/2007
Posts: 1772
Location: Australia, Victoria
The game isn't copyrighted, so, we can't get banned for spreading a 512kb file around now, can we. <3
Joined: 7/24/2008
Posts: 9
Haha, I was just messing around with this game yesterday.
MarbleousDave
He/Him
Player (13)
Joined: 9/12/2009
Posts: 1560
Labyrinth Zone 2 Glitch City FTW
Skilled player (1417)
Joined: 10/27/2004
Posts: 1978
Location: Making an escape
NitroGenesis wrote:
Oh no! you posted a rom link in the description!!! EEEK
I've done it before.
A hundred years from now, they will gaze upon my work and marvel at my skills but never know my name. And that will be good enough for me.
Experienced player (642)
Joined: 11/30/2008
Posts: 650
Location: a little city in the middle of nowhere
Hate the game. Love the run. Yes.
Measure once. Cut twice.
Joined: 11/4/2007
Posts: 1772
Location: Australia, Victoria
DailyMotion (Excuse the quality, DailyMotion... is incapable of processing Somari with any decent quality) Archive.org I'm reencoding it into even higher quality settings, but unless the improvements aren't significant (For the file size actually becomes smaller), consider this the final encode.
Expert player (2574)
Joined: 12/23/2007
Posts: 831
Sonic+Mario=Somari? Voted Meh.
Recent projects: SMB warpless TAS (2018), SMB warpless walkathon (2019), SMB something never done before (2019), Extra Mario Bros. (best ending) (2020).
Joined: 5/29/2004
Posts: 757
I honestly found myself very unentertained and this game an absolute pile of crap. The ONLY redeeming factor were the odd level where it was minimallly entertaining.. but for the most part I had to keep ff'ing through from boredom. It is slow paced.. it is difficult.. just not my cup of tea. I voted No on this. Sadly I see a lot of people voting yes because it's essentially a Sonic Game on the NES, so that alone will make them overlook the crappiness. Mind you, I can see some people voting yes because they genuinely enjoyed it.. but not many for that reason. I will say this. WAYYYY more entertaining then the Columns runs though. Mr. Kelly R. Flewin
Mr. Kelly R. Flewin Just another random gamer ---- <OmnipotentEntity> How do you people get bored in the span of 10 seconds? Worst ADD ever.
Joined: 7/2/2007
Posts: 3960
I'm inclined to agree with Mr. Flewin here. The game looks sloppy, no matter how tightly-optimized it is. Some games just don't make good TASes. Voting meh; seemed decently well-run but just not entertaining.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
Joined: 10/3/2005
Posts: 1332
Voted no, same reason. It may be a homebrew, but it looks like a bad hack. It's also a game/series for which we already have many polished TASes of superior, official titles.
Editor, Player (69)
Joined: 1/18/2008
Posts: 663
For what it's worth, as part of my defiance regarding the MM3 MM4 submission debacle, I purposely made parts look sloppy if it lost no time. I understand some won't like this but this aesthetic decision was only made later into the run and is, again, a show of defiance. Some parts do look pretty bad though due to the really limited and simple engine in this game, which can't really be helped, so it may be hard to tell what is due to me or what is due to the game itself. I also understand this is a niche title. If you aren't interested in fapping to Mario and don't enjoy Sonic runs, I don't expect you to enjoy something ear grating, an eyesore, and a skin deep ripoff of Sonic 1. (But yes, no matter how you put it, it is better than Columns.) However, if you do fit in at least one of those above groups, or really like seeing glitches due to bad programming, or had played this back when the emulation scene was starting to hit it big, then you'll likely enjoy this run in some degree. And those who did play it back in the day also have the understanding of the sheer difficulty of running it unassisted... I ran this for them :)
Dromiceius wrote:
Voted no, same reason. It may be a homebrew, but it looks like a bad hack.
It is also a pirate that was actually put on cartridges and sold for Famicoms. Not quite homebrew. But a hack of what? (Or do you mean hack in the traditional sense?)
Dromiceius wrote:
It's also a game/series for which we already have many polished TASes of superior, official titles.
I guess nobody should run any platformers anymore. There's plenty of TASes published for them. Darn. I'd hope this third party game would be judged by its own merits (of which you may not find any, that's fine) rather than that of other similar in appearance but otherwise very different first party games. Also, for what it's worth, I don't recall any published or submitted runs where Mario goes through loop-de-loops or kills Crabmeats and Buzzbombers. If I am missing one, though, please point me to it.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
mklip2001
He/Him
Editor
Joined: 6/23/2009
Posts: 2227
Location: Georgia, USA
You mean the MM4 debacle? You wrote MM3. For what it's worth, this looks like it's a pretty good run of the game. There seem to be a lot more enemies in this game than in Sonic 1, which makes it harder to maintain consistent speeds. The glitching in Labyrinth Zone 2 was pretty awesome, as well as a few other wall skips in other parts of the run. That being said, the game is not good. It's trying to be Sonic 1, and I have to give the developers some credit for trying to make a Genesis game for NES hardware. Unfortunately, this means that the game turns out obnoxiously slow-paced for a Sonic game (three and a half minutes to complete Labyrinth 3?! Come on!). The fact that they nerfed the spindash, making it unusable for a speedrun, really kills the entertainment too. There are two ways I could judge this. One way is by saying that because the game is essentially trying to be Sonic 1 with a Mario sprite, this game falls very short of Sonic 1 and hence should get a No vote. The other way is by saying that this is a good run of a hard bad game with some redeeming moments. I will use this second way, and you get a Meh vote.
Used to be a frequent submissions commenter. My new computer has had some issues running emulators, so I've been here more sporadically. Still haven't gotten around to actually TASing yet... I was going to improve Kid Dracula for GB. It seems I was beaten to it, though, with a recent awesome run by Hetfield90 and StarvinStruthers. (http://tasvideos.org/2928M.html.) Thanks to goofydylan8 for running Gargoyle's Quest 2 because I mentioned the game! (http://tasvideos.org/2001M.html) Thanks to feos and MESHUGGAH for taking up runs of Duck Tales 2 because of my old signature! Thanks also to Samsara for finishing a Treasure Master run. From the submission comments:
Shoutouts and thanks to mklip2001 for arguably being the nicest and most supportive person on the forums.
Skilled player (1652)
Joined: 11/15/2004
Posts: 2202
Location: Killjoy
True wrote:
For what it's worth, as part of my defiance regarding the <strike>MM3</strike> MM4 submission debacle, I purposely made parts look sloppy if it lost no time. I understand some won't like this...
As in, it won't get published? I don't think you get how things work here. If you submit subpar movies, the site doesn't suffer. courtesy of Failblog.org... Protest FAIL.
Sage advice from a friend of Jim: So put your tinfoil hat back in the closet, open your eyes to the truth, and realize that the government is in fact causing austismal cancer with it's 9/11 fluoride vaccinations of your water supply.
Joined: 6/4/2009
Posts: 893
my... i was surprised as the hack isn't "so bad" some levels uterly suffer from the downsample ( like MZ) but levels like LZ are actualy not that visualy bad, the game isn't that laggy and have a realy decent speed for a nes hack... i was even surprised that they bothered to program the underwater air management.... the run is good and the glitches made me smiles, yes vote but this run should be published as a concept demo/hack.
Banned User
Joined: 12/23/2004
Posts: 1850
I don't see why. The game isn't a hack, it's just a shitty unlicensed game.
Perma-banned
Editor, Player (69)
Joined: 1/18/2008
Posts: 663
mklip2001 wrote:
You mean the MM4 debacle? You wrote MM3.
Yeah. Fixed. The user previously went by the username 'mega_man_3' and I still refer to him as that... all this confusion got me confused :)
mklip2001 wrote:
The fact that they nerfed the spindash, making it unusable for a speedrun, really kills the entertainment too.
It isn't unusable, but it isn't usable in the traditional way. I still used it in the run.
mklip2001 wrote:
reasonings behind vote
I clarify above (a little bit) why I made the run and what audience it is supposed to reach. It's fairly fun to watch just for the glitches (if I didn't know the game, I'd think that it is interesting and notable being a pirate ripoff of Sonic 1 and that there is enough random stuff to keep me entertained, plus I like Sonic-type runs) but those who played this when it hit the exploding scene years ago will find the most enjoyment.
DarkKobold wrote:
As in, it won't get published? I don't think you get how things work here. If you submit subpar movies, the site doesn't suffer.
No, the way things work here is that rules are the extreme law of the land, even if the rule is completely wrong. (And the visible suggestion to fix these rules has been to add several more...) No, the way things work here is that a judges biased opinion supersedes everything so if the game can't naturally be rejected, a rule is used as a scapegoat. No, if it doesn't get published simply because someone doesn't like MY aesthetic choices, which I find more entertaining than a bunch of follow-the-song "dancing" or other crap, and ESPECIALLY without actually trying things themselves, then I think that goes against the nature of a speedrun and I don't want it, or anything else I am working on, submitted or published here. No time is lost where I make it look sloppy; it is still a "superplay" and it took just as much effort to make it look aesthetically sloppy without technically being so as it would to make it look somewhat polished (good luck making this game look really shiny). This is a minor protest and guess what? Sometimes, some battles don't work out in my favor. I hope this one does. But at least I've got the balls to do it instead of just getting in line and taking it. And while may be ultimately minor and insignificant, in front of my eyes it is weeding out the crybabies, the people who don't care, and the people who understand. Obviously the majority are enjoying this run well enough.
Nicos wrote:
...the game isn't that laggy...
If you're talking about frame lag, there is none except during score tally and title card screens.
Nicos wrote:
...but this run should be published as a concept demo/hack.
Why? To extend Xkeeper, this game was published on real cartridges and sold to run on real hardware.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Joined: 6/4/2009
Posts: 893
True wrote:
Nicos wrote:
...but this run should be published as a concept demo/hack.
Why? To extend Xkeeper, this game was published on real cartridges and sold to run on real hardware.
because it's missing something important : :D more seriously, i'm for the concept demo / hack tag because it's not a "official" game aproved by nintendo on the console, it's would be like tassing homebrews as games because they can run on a real playstation... (btw is anybody's bad enough for a TAS of rocks n gem on the net yaroze ?)
upthorn
He/Him
Emulator Coder, Active player (392)
Joined: 3/24/2006
Posts: 1802
I could only stand to watch about 3 minutes of this run. Without an in-depth understanding of the physics, I can't tell if I should wholly blame the game, or if there are obscenely poor stylistic choices involved here as well, but either way, this particular TAS does not seem acceptable for publication. Especially given the author's comment that he intentionally made parts look sloppy, which goes against the principal goal of TASing.
How fleeting are all human passions compared with the massive continuity of ducks.
Post subject: Added comment
Joined: 11/4/2007
Posts: 1772
Location: Australia, Victoria
On a less negative note: Here, have a higher quality encode. Having personally helped work on the game, I can say this... the physics are almost impossible to work with to create any style out of, as are the timers. And yes, I agree, Labyrinth Zone 1 and 3 are terrible (Sheerly because of how impossible it is to go above 2px speed), but I don't see them as sheer reasoning to not publish the run, I thought the other Zones were fairy good and hasty.