Introduction

One day I thought about how different a Sonic any% run would be without Tails’ help. Long story short I decided it would be worth making this tas. I made the whole thing from scratch and made absolute sure I saved every single frame I can. Overall I had fun making this run despite the pain and suffering it made me endure and am really pleased with how it came out.

Emulator and Tools used

  • Gens Re-Recording 11b + Camera hack + Solidity and Terrain dysplay
  • Sonic TAS tools

Movie objectives

  • Aims for fastest in-game time
  • Takes damage to save time
  • Abuses deaths to save time
  • Abuses programming errors and glitches
  • Manipulates luck and enemy positions

Timetable

ActIn-game Time
Angel Island 10:40:55
Angel Island 21:14:51
Hydrocity 10:39:16
Hydrocity 20:22:20
Marble Garden 10:34:37
Marble Garden 20:03:24
Carnival Night 10:29:19
Carnival Night 20:39:17
Icecap 10:56:45
Icecap 20:00:00
Launch Base 10:33:46
Launch Base 20:26:37
Mushroom Hill 10:33:37
Mushroom Hill 20:40:34
Flying Battery 10:45:26
Flying Battery 21:05:43
Sandopolis 11:04:51
Sandopolis 20:48:47
Lava Reef 10:47:43
Lava Reef 20:20:52
Hidden Palace0:26:40
Sky Sanctuary0:52:17
Death Egg 10:58:56
Death Egg 21:36:23
Total16:52:46

Time lost to deaths

ActIn-game Time
Hydrocity 20:05:21
Launch Base 20:00:00
Death Egg 20:04:29
For those of you not familiar with the formatting of the in game timer it goes Minutes:Seconds:Frames with 60 frames being the equivalent of 1 second. I did add the time lost to deaths to the total time.

Level comments

Jumping from a roll

Before I get into the level comments there’s a couple bugs I want to point with how the game handles jumping from a roll. The first one being the game uses Sonic’s taller standing hitbox instead of his rolling hitbox. The other one is you have no aerial control. While these aren’t exclusive to Sonic it does make up an important part of his movement.
It’s possible to regain air control by doing any double jump action, even if it does nothing in the case of regular invincibility and super sonic. This is why you’ll see me using insta-shield a lot without actually hitting anything.

Angel Island 1

This level was actually done by Aglar who saved 13 frames over my initial attempt and made it like a good first level for the run.
0:11 : Saved a few seconds by using this loop zip strat WST found.
0:31 : To give you an idea of how serious I was about the optimization of this tas. I used a kind of movement tech usually only seen in Sonic 1 tases where I do a 1 pixel zip off the wall by rolling over my horizontal subpixels. I did this to get a small speed boost off the wall before jumping on the rock.

Angel Island 2

0:18 : By taking advantage of a few mechanics I was able to do a diagonal clip through the flipping bridge.
0:23 : With careful movement I was able to save 5 frames with this loop zip.
0:27 : Hitting a badnik while going up (or the bottom half of their hitbox if falling) eliminates some of your vertical speed, so I aimed for the Rhinobot to get a faster landing.
0:30 : Unfortunately because there’s a small bump right at the entrance of the tunnel Sonic can’t maintain spindash speed here like can with the other 2 tunnels. Only with the lightning shield he can do it but I wasn’t able to get a tunnel jump out it.
I had to sacrifice 7 frames in order to do the bridge cutscene skip. If it weren't for this the time for the level would've been 1:14'44 instead I could say this was a speed/entertainment trade off because of the ingame time goal but to tell you the truth it would’ve felt wrong not to do this glitch in a tas.

Hydrocity 1

Without Tails Sonic is unable to do a level wrap. Oh well, that just means Sonic will have to go fast instead. A better optimized (and obviously altered to not include big rings) version of the route used in the Sonic 100% run was the way to go here.
0:18 : So here’s something bizarre you probably didn’t know about the speed cap. It’s only a horizontal cap of 16 pixels per frame (the exact speed a red spring gives you) and it doesn’t keep you from building up more speed past this. This is what allowed me to jump so high off the wall.
0:24 : I was able to greatly improve this ramp clip by getting sonic to land instantly at the zipping height.
0:29 : For some reason jumping the frame Sonic rolls into a badnik boosts his jump. This allowed me to make it onto the ledge right behind it a few frames sooner.
0:31 : This seemingly pointless braking was actually done to trigger the quick death during the act transition. This wasted 11 frames which was no big deal considering this death saves in game time unlike the Angel Island 2 bridge skip.
0:33 : Since you don’t need exactly 0 speed to spindash on normal tiles in S3K I’m able to spindash on steep slopes.

Hydrocity 2

Since Tails isn’t here Sonic has to do the offscreen boss kill himself. I used the same technique from the act 1 boss to get a fast kill.

Marble Garden 1

The level wrap was already a huge improvement for this level. Thanks to the first clip right after the level wrap I was able to get the boss to spawn at the first possible frame. Then I saved a few more seconds by hitting robotnik the frame he disappeared after “breaking” the ground like in the ring attack.

Marble Garden 2

No >:C

Carnival Night 1

By replacing both carry clips with spindash clips not only am I able to actually do the level wrap from Knuckles’ area, but also save several seconds. Also found a faster 1P strat for the boss.

Carnival Night 2

Was able to take Knuckles' any% tas route by substituting his glide with a bubble shield.
Because Sonic can’t move the signpost while embedded in the ground I had to choose between getting the signpost bubble shield monitor or using the zip while getting a bubble shield from the tunnel up ahead instead. Getting the signpost bubble shield turned out to be faster in both ingame and real time.
0:08, 0:16, 0:19, and 0:28: You’ve probably seen this in other Sonic tases but you can do stair clips with the bubble shield.
0:14 : So these moving barrels are on proximity based cycles, meaning they start moving from a set position as soon as they load. This particular barrel spawned at exactly 0:10:13 ingame time. I had to delay getting in the vertical tube by a few frames so that I could make it onto the barrel. Vertical subpixels were manipulated before the start of the level so that I could go over the barrel a frame sooner without losing time anywhere else.
0:21-0:26 : I can’t believe that after all these years the barrel of doom gave me trouble yet again. To make the clip work you have to use up the bubble shield bounce before being able to land in the wall. The problem is this greatly limited the positions you could start falling from to 16 pixel increments. After quite a bit of experimenting I found out that there was literally only one position that yielded favorable results.

Icecap 1

Found a way to do the boss clip without having to roll during the act 1 portion of the stage. This did make it a bit harder but I still managed to trigger the transition snow block the frame it loaded.

Icecap 2

Existence is but an illusion.

Launch Base 1

With the help of ram watch I saved a frame by getting the crusher to load a frame sooner (the first jump is what delayed it). Having the crusher clip me in the ground to the left instead of the right allowed me level wrap right away saving 9 frames.

Launch Base 2

Saved 1 frame with better camera manipulation before level wrap. Then proceeded to annihilate both bosses.

Mushroom Hill 1

So I accidentally discovered something kinda bizarre about the boss. Rather than having set spawn coordinates it spawns based on the camera position. This meant that it's possible to get it spawn lower than intended to get a faster kill. Another strange thing about this boss is that it checks for player position for the vertical lock, meaning I have to bring the lower the camera before doing the level wrap. Afterwards I have to avoid scrolling the camera down or else it’ll correct its height (yes this is why there’s no playaround here).
Conveniently this also allows me to use the trick where I get the first hill from act 2 to load on top of Sonic.

Mushroom Hill 2

So I’m not sure exactly who came up with the concept for this route, but I was the one to make it work. Basically you have to hit a pathswapper along the way otherwise the exit to the big ring room will be solid. My solution to this was to hit the pathswapper at the end of Knuckles’ area (shown in the screenshot below) and then clip into the walls again by doing a ramp clip like the one in Hydrocity 1. I was able to do this without slowing down the camera. Afterwards I had to wait a few seconds in the big ring room for the camera to catch up because I needed a red spring to load.

Flying Battery 1

Saved 1 frame by optimizing the first jump. Saved another from triggering the platform a frame sooner.

Flying Battery 2

After I jumped onto the ramp I turned around from inside the wall and did a vertical zip jump. This jump allowed me to trigger the pathswapper (which made the right half of the walls fully solid again) at the top of the rotating mesh and then land on the upside/down slope which allowed me to reach the top layer walls in the map which extends all the way to the end of the level. I’m happy that this new route also turned out to be faster than the other one.

Sandopolis 1

Originally I thought you could only clip into these slopes using a fire shield (which is still mostly true), but an rta runner named BenInSweden showed me a clip of him accidentally clipping into this slope without one. I was able to get a level wrap out of it.
In case anyone is interested here's the non level wrap version I made before it was discovered. Yes I am kinda sad this didn’t end up in the final run but at least it’s here for the public. Consider this a little bonus tas :)

Sandopolis 2

This level has been greatly improved with some new tricks and new routing saving 9 seconds in total.
0:00 : I will admit I kinda lucked out with the platform cycles. I had to start the level at a specific frame so that I could use the first moving column to stop myself by fire dashing into it which saved 4-5 frames over braking. Conveniently none of the other columns in the level got in the way either. This meant that I only had to tas the level twice (the first iteration was a testrun) to get the fastest ingame time.
0:08 : With precise horizontal and vertical positioning and speed I’m able to do a diagonal clip through this slide.
0:22 : It turns out Sonic can clip into the ground using this sand block but only if he’s in the exact center pixel.
0:30 : In order to do this loop zip while saving the most amount of frames I can I had to do the spindash that came before it within a 3 horizontal subpixel range.
0:39 : The way this loop clip works is quite simple actually. There’s a pathswapper that’s in line with the top of the loop. Once you swotch the terrain a part of the wall in the loop becomes non-solid. Normally you aren’t suppose to he able to get back down but you can basically do a stair clip by doing a low enough jump and not moving. I actually would’ve done this with the loop I did I zip off of but the pathswapper on that one is 2 way.

Lava Reef 1

Without Tails here I had to improvise, by Improvising the improvisation? In any case I was able to do the previously Tails exclusive stair clip with Sonic though precise positioning and speed along with the insta-shield to regain air control.
I also managed to clip through the collapsing platforms in the beginning of the level and the spindash elevator by abusing this lip in the terrain.
Just like in Launch Base 1 I used ram watch to make sure I got the crusher to load on the first possible frame.

Lava Reef 2

Saved 1 frame with better camera manipulation before the level wrap. I did try to save another after the second transport tube with some movement optimizations but all I ended up doing was getting a few pixels ahead.

Hidden Palace

Nothing special here. Although there is a 1 frame timesave I will mention because it was overlooked in the 2014 tas and nearly overlooked by me for this one. Apparently the vertical position you activate the teleporter from does affect how high it'll bring you up top.
0:16:14/79377 : oh no
0:17:29/79452 : oh no
0:19:01/79544 : oh no
0:20:33/79636 : oh no
0:22:05/79728 : oh no
0:23:37/79820 : oh no
0:25:09/79912 : oh no
0:26:40/80004 : oh no

Sky Sanctuary

0:23 :Just like the moving barrel in Carnival Night 2 the moving bar up ahead is on a proximity based cycle. I take advantage of this to manipulate its position as Sonic needs it to make it onto the ledge up ahead.
0:38 :I was able to considerably improve the landing on this hidden spring by doing a roll jump to bump into the ceiling. I had to initiate the roll with the minimum 268 speed (12+128+128) to spend the least amount of frames braking on the spring.

Death Egg 1

0:06 : Through a precise landing on the slope I was able to instantly zip in the wall.
0:18 : I used the spikes to clip me into the conveyor belt which in turn clipped me into the wall. I avoided getting crushed by panning the camera down to temporarily deload the spikes.
0:28-0:42 : I saved several frames here over the 2014 tas through some small optimizations.
0:46-0:50 : Freaking Redeye. The 1st phase nearly drove me insane. The final version you see in the run didn't take me that long to execute. It was trying to save more frames beyond that which made me want to use a toothbrush made out of Sonic's spines. I even asked Aglar if he could save any frames and he had no luck either.
0:54-0:58 : The 2nd phase was pretty fun though. It wasn't too hard to get a frame perfect kill here.

Death Egg 2

With a bit of experimentation I found a way Sonic can do the level wrap without having to pan the camera.
One thing I wanna point out about the Death Ball real quick. The speed of your spindash does effect how fast the spike balls fly up (down?), so I had to make sure the last one was hit at point blank range with a fully charged spindash to save the most amount of frames.
Right before transitioning to the final bosses I did a second spindash to trigger a glitch where the solidity there gets shifted to the left. I did the smallest possible shift of 32 pixels which is indeed the length of one block. This did waste 2 frames but I more than made up for that by getting a frame perfect kill on the boss.
There were some timesaves with the death egg robot as well. With the 1st phase I found a way to destroy every finger on the second pass. With the 2nd phase I used the same trick from the Tails tas (just executed a bit differently) to get a frame perfect kill.

Potential Improvements

  • Angel Island 2: As I mentioned in the level comments it is possible for Sonic to carry full spindash speed into the last tunnel with the lightning shield, but despite my best efforts I was not able to get a tunnel jump out of it.
  • Sandopolis 2: Theoretically you can save 1 frame after the second fire dash slope clip with better vertical subpixel management. Unfortunately in this run I wasn't able to get the desired results to save said frame.
  • Sky Sanctuary: This actually wouldn't have saved any time in this run because of the 4 frame rule the teleporter has but it could help for future runs. With the spindash-jump from the vertical screen wrap at 0:32 if you have low enough subpixels it's possible to extend the jump by a frame. Normally I would've done this anyways but the managing my vertical subpixels in this level was already pretty difficult.
  • Death Egg 1: There’s potential to save some frames in phase 1. The spikes made this practically impossible to achieve please, no more

Special Thanks to

  • Aglar: For his input on Angel Island 1 and his previous tases.
  • BenInSweden: For that slope clip in Sandopolis 1 that led up to the level wrap.
  • Chrezm: For the fast Sandopolis 1 boss strat. (yes it is the thing that saved the 108 frames feos mentioned with the file replacement)
  • Evil_3D: For finding the stair clips in Carnival Night 2 (originally used for our ring attack tas) and his other tases.
  • DMTM: For the first terrain clip in Marble Garden 1 used immediately after the level wrap.
  • Marzojr: For his guides on the Angel Island 2 bridge cutscene skip and quick death at the start of Hydrocity 2 and his previous tases.
  • WST: For the second loop zip strat in Angel Island 1.
  • Everyone else for their contributions and generally being awesome.

Suggested screenshots


feos: Judging...
feos: Replacing the submission with a file that saves 108 frames in Sandopolis 1. Overall time is the same though.
feos: We've researched the hell out of this movie and the situation it is in. This branch and [2741] Genesis Sonic 3 & Knuckles by Aglar & marzojr in 26:53.06 are the most similar branches of Sonic 3 & Knuckles that exist today. And this game has 8 branches already, if we don't count the S3K hacks. Due to massive similarity I was ready to reject this run, but arkiandruski suggested a brilliant solution: since this movie is faster than Sonic+Tails, in both in-game timer and actual gameplay duration, it counts as an improvement! I verified the times, and everyone agrees with this idea, so I'm finally accepting this run to obsolete [2741] Genesis Sonic 3 & Knuckles by Aglar & marzojr in 26:53.06. Stars and flags will be decided after it gets some ratings.
Stovent: Processing...


WST
She/Her
Active player (489)
Joined: 10/6/2011
Posts: 1697
Location: RU · ID · AM
McBobX wrote:
Imagine another TAS improving this one by a lot of time, haha!!! Good job though!!
Actually, it’s not hard to imagine for any Sonic TASer. This game proved it’s complexity many times, and there are really tens of examples when a “good” run got updated by many seconds, or even minutes. This game is full of inobvious strategies and things like «going slower to go faster», a good example is MGZ2 boss, where delaying the last hit by a few frames makes the whole boss battle many seconds faster. For years this was not known because every TASer just tried to beat that boss in a frame-perfect way.
S3&A [Amy amy%] improvement (with Evil_3D & kaan55) — currently in SPZ2 my TAS channel · If I ever come into your dream, I’ll be riding an eggship :)
Skilled player (1458)
Joined: 11/26/2011
Posts: 656
Location: RU
WST wrote:
McBobX wrote:
Imagine another TAS improving this one by a lot of time, haha!!! Good job though!!
Actually, it’s not hard to imagine for any Sonic TASer.
I believe that it is true for many other games too, if to dig deep most of published movies can be improved. But most of other games not have such great TASers community as Sonic games to regulary find new and new improvements.
I show you how deep the rabbit hole goes. Current projects: NES: Tetris "fastest 999999" (improvement, with r57shell) Genesis: Adventures of Batman & Robin (with Truncated); Pocahontas; Comix Zone (improvement); Mickey Mania (improvement); RoboCop versus The Terminator (improvement); Gargoyles (with feos)
Former player
Joined: 4/17/2009
Posts: 88
This cannot be rejected because "too many branches", there's already 3 Sonic ring attacks and an Amy Rose hack submissions, these are just made up challenges & an unofficial hack, unlike this Sonic solo. To delay or reject for this reason would be unfair to the author. No more sonic ring attacks or hacks should be accepted anymore imo.
Samsara
She/They
Senior Judge, Site Admin, Expert player (2238)
Joined: 11/13/2006
Posts: 2822
Location: Northern California
Pikachu wrote:
This cannot be rejected because "too many branches", there's already 3 Sonic ring attacks and an Amy Rose hack submissions, these are just made up challenges & an unofficial hack, unlike this Sonic solo. To delay or reject for this reason would be unfair to the author. No more sonic ring attacks or hacks should be accepted anymore imo.
There are a number of problems with this reasoning. 1. One of the ring attack TASes is for Sonic 2. It's unfair to say there are three of them without specifying that one of them isn't even for this specific game. 2. Hacks are published based on audience reception, if one makes it through then it deserves a spot on the site based on that alone. Amy is sitting very comfortably in Moons right now, meaning that even after all this time, people are generally still positive about it, and with an improvement run in progress I can't see that changing at all. 3. Consider that there's already a solo Sonic run that shows off more of the game, and consider that the any% branch is already technically playing as Sonic. Optimization level/new tech aside, what does this run have that those runs don't? My take on separate branches is that they need variety to warrant publication. Sonic characters all tend to play fairly similarly, just with each of them having one or two unique movement abilities that, in my opinion, still don't make them varied enough to warrant every single one having distinct any% and 100% categories published. I can understand only having 100% for every character, and I even support that, as more run logically translates to more variety, giving each individual character more of a chance to show off how different they are. This brings me back to the ring attack runs. Ring attack runs, to me, fill that gap of being varied really nicely (though I do think having 2 runs with the same character for 1 game is a bit much) (EDIT: disregard this, misread categories while tired, sorry!). They're basically full-game exhibitions, the strategies are super unique, they're constantly moving around and doing interesting things, and it never gets boring because each stage is laid out completely differently so there's always innovation in how they're handled. I feel like I just reworded the same thing four times. No wonder I was always great at essays! It's not a matter of being unfair to the author, either. It's just a natural process of how the site handles things in its current state. Even if the run gets rejected, there's always a chance that a rule/site structure change down the line could see it come back and be published. It's happened before. It just happened very recently, in fact. It could easily happen again with a high quality run like this. Honest question: If ring attack runs are arbitrary to you, why is this run not arbitrary? Is there significant enough of a difference for it? Is a general viewer unfamiliar with Sonic TASing going to get something noticeable out of it that the 100% run or the normal Sonic+Tails runs wouldn't otherwise give them? I'd like to see someone with more intimate Sonic knowledge (NOT THAT KIND) chime in on things like that. NON-EDIT EDIT: Everything above this was written yesterday while lacking sleep, so sorry if it sounds weird. The following should technically be a separate post, but double posting is bad. I got curious about similarities and just watched this run and the Sonic+Tails run simultaneously, level by level, comparing them. My verdict is that an optimized any% run would be virtually identical to this one, save for the following stages: ANGEL ISLAND 1: Tails is able to move Sonic during the opening cutscene. This is a minor difference at best, the levels otherwise looked 100% similar except for the new tech in this run. HYDROCITY 1: The published any% run is able to use Tails to perform a full level zip. Solo Sonic can't do this. This is the biggest difference between the two runs. LAUNCH BASE 2: Solo Sonic seems to be unable to do a boss/cutscene skip that any% is able to perform. MUSHROOM HILL 1/2: Tails can break the victory cutscene in Act 1, though the new tech in solo Sonic to go through Act 2 might warrant that obsolete anyway: Rough time estimates on the encodes: This run spends about 30 seconds in MH2 (starting from the end of the MH1 victory cutscene), while the published S+T run spends about 40 seconds in it (starting from using Tails to break the MH1 victory cutscene), both ending at the very first hit on Eggman. I'm assuming that an optimized S+T run would break the cutscene and activate that new zip immediately, following the same general route, though I'm not familiar enough with Sonic TASing to know whether or not that's actually the case. HIDDEN PALACE: Tails is able to quick kill Knuckles and skip the rest of the level afterward. DEATH EGG 1: That weird hit-the-targets thing in the middle of the level is about 3 IGT seconds faster with Tails, otherwise it seems like both runs would be the same. Here's the brief notes I took while watching. They're not great, but I'm including them for posterity. Keep in mind I'm not a Sonic TAS expert so my untrained eye might be wrong about things. I'd gladly appreciate corrections to anything I'm wrong about! As it stands, I don't really see these as significant enough differences to warrant a solo Sonic any% publication. Hydrocity 1 is the only major gameplay difference, as solo Sonic actually has to navigate through the whole level (and it's quite entertaining, in fact). Everything else is a boss/cutscene/uninteresting gameplay skip. Don't get me wrong, the new tech on display in this run is amazing, it's incredibly well made and optimized and super entertaining to watch, but apart from a single stage with different gameplay, an optimized S+T any% run would serve the exact same purpose as this run. This just feels like an S+T run that arbitrarily doesn't use Tails. It's a fantastic run, but I don't think it deserves a new branch. ...But holy hell am I excited to see an S+T any% improvement with all this new tech.
TASvideos Admin and acting Senior Judge 💙 Currently unable to dedicate a lot of time to the site, taking care of family. Now infrequently posting on Bluesky
warmCabin wrote:
You shouldn't need a degree in computer science to get into this hobby.
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
biggest specific changes are in Hydrocity 1 and Hidden Palace
Hidden Palace in "Sonic" is the same as in "Tails" though.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Samsara
She/They
Senior Judge, Site Admin, Expert player (2238)
Joined: 11/13/2006
Posts: 2822
Location: Northern California
There's that too, yeah. I was just strictly comparing this run to the Sonic+Tails any% since they're the easiest to directly compare. I had the idea to compare to some of the other published runs, but wanted to wait for some feedback/explanations/corrections first. With the main difference just being down to Hydrocity 1, I might as well do it now. Solo Tails is able to get the zip in Hydrocity 1 that this run can't do (not technically important compared to this run, but stating it regardless). Knuckles can't get this zip and has to play the stage normally, but Knuckles is able to use a different route through the first half of the level. The Knuckles and solo Sonic routes do eventually converge at that halfway point and look identical up through to the end, though. Sonic 100% also plays Hydrocity 1 with the exact same restriction as this run, and the bonus stage is at the very end of the level, meaning an optimized run in 100% would be the exact same nearly the entire way through, except it spindashes left instead of right at the very end. Painful as it is to say this, without Hydrocity 1 or Hidden Palace, this run officially doesn't have any strictly unique content as far as I'm aware. Note that I mean strictly unique content as routing/strategies that are unique to this specific run. The new tech, awesome as it is, doesn't count as unique as it's applicable to other runs, namely the already published solo Sonic 100% and S+T runs. Different setups to hit the same full-level zips also don't count, as that is technically just suboptimal strategy (assuming the optimal strategy is using Tails to clip, of course). I'd still love to hear feedback from Kaan or anyone else with more technical Sonic knowledge in case my analysis is wrong in any way.
TASvideos Admin and acting Senior Judge 💙 Currently unable to dedicate a lot of time to the site, taking care of family. Now infrequently posting on Bluesky
warmCabin wrote:
You shouldn't need a degree in computer science to get into this hobby.
Memory
She/Her
Site Admin, Skilled player (1556)
Joined: 3/20/2014
Posts: 1765
Location: Dumpster
So I finally watched the TAS for myself. Now I haven't done an in-depth comparison like Samsara, but at no point did I feel that it was that different from what I had seen before in S3K. If the differences exist, they're subtle, which is very bad for a game with nearly 10 published branches.
[16:36:31] <Mothrayas> I have to say this argument about robot drug usage is a lot more fun than whatever else we have been doing in the past two+ hours
[16:08:10] <BenLubar> a TAS is just the limit of a segmented speedrun as the segment length approaches zero
Banned User
Joined: 6/7/2017
Posts: 420
Location: Somewhere
Sorry I took a while to respond. I can tell you that the differences in an updated Sonic and Tails run would not be as subtle. To help prove this I'm going to list out the updated strats. This may take some time.
Banned User
Joined: 6/7/2017
Posts: 420
Location: Somewhere
Angel Island 1- First half would more or less be the same as the published run with the second loop zip. Second half will use the same zip from the NG+ tas thanks to a new 2P stair clip technique. Also has frame perfect boss fight Angel Island 2- This does need testing but Sonic and Tails may end up taking the top route. There is another lightning shield monitor I can grab and with Tails' help Sonic can take the top route without losing his shield. Otherwise it will more or less be the same as this run. Just with small timesaves courtesy of Tails. Hycrocity 1- As mentioned by other this level has a level wrap. Updated version will likely just being a slightly faster version of published run. Also has frame perfect boss fight Hycrocity 2- Will probably be the same as this run with a frame perfect boss fight. Marble Garden- Will probably be the same as this run with 1 possible timesave with Tails. Carnival Night- Possible return of wheel glitch thanks to new strats found by the rta community. Also faster Act 1 boss fight. If wheel glitch ends up being slower then the level will just be a faster version of this run. Icecap-Sonic can get slope glitch much sooner with the first ice block. Will also use act 2 boss skip like in this run. Launch Base 1- Will probably be the same as this run with faster boss fight. Launch Base 2- Sonic can't skip the 1st boss and eggmobile ride by himself. He can't do the act 1 in act 2 glitch and even if he could he still wouldn't be able to reach Knuckles' boss area. Mushroom Hill 1- Sonic will do the level wrap without dropping down inside the walls since Sonic and Tails can beat the boss fast enough together to leave the screen unlocked. Mushroom Hill 2- Sonic and Tails won't be taking the same route as this run. The reason for this is only the horizontal boundary remains unlocked during the score tally while the vertical ones stay where they are which leaves the death plane at the bottom of the screen. As a result they have to go up top like in the published run to be able to progress during the score tally. As for updated S+T strats. Sonic may end up using the lightning shield from the act 1 boss arena. Will also have a faster boss fight followed by the early capsule button glitch like in the ring attack. Flying Battery 1- Much faster level wrap as shown in the video below. Also with Tails' help Sonic can reach the boss without having to wait for objects to load. Link to video Flying Battery 2- Will likely use the same route as the one in this run. Sandopolis 1- Other than a slightly faster boss this will be more or less the same. Sandopolis 2- Good possibility of level wrap via light switch glitch. Will also have frame perfect boss fight. Lava Reef 1- Faster level wrap (shown in the video below) that skips basically all the content in the level. With Tails' help Sonic will be able to kill the 2 circular turrets faster Link to video Lava Reef 2- Possibility of slightly faster level wrap with Tails' help. Otherwise the level will pretty much be the same. Hidden Palace- Minor timesave with the first spikes. Considerably faster Knuckles fight with 2P strat. Also skips all the cutscenes. Sky Sanctuary- Will more or less be the same as the published run. Death Egg 1- Will likely use a different route from this run. Also gravity room skip is a thing with Tails' help. Frame perfect phase 1 redeye fight. Link to video Death Egg 2- Fast death during act transition as shown in the published run. Will likely have faster level wrap with Tails' help. Timewise the only boss Tails will save time with are the fingers. Also no invisible floor glitch in the final boss arena as it's not necessary and wastes 2 frames. As I was typing this the problem became more and more apperent to me. Compared to the 2014 Sonic and Tails run an updated run would indeed be more different from this one. In fact a good chunk of the game is heavily impacted by not having Tails. However there are indeed some levels that would only have slight differences which I feel is okay to some extent. I made this run with the intention to show off how Sonic would handle each level by himself. I thought that by itself would be interesting. If the game wasn't as heavily impacted as it is I wouldn't have even gone through with making this run. With the possible exception of Lava Reef 2 no level would end up being exactly the same.
Samsara
She/They
Senior Judge, Site Admin, Expert player (2238)
Joined: 11/13/2006
Posts: 2822
Location: Northern California
Going through that post, I'm seeing a lot of either "this will more or less be the same" or just talking about level wraps and zips in general. Here's the big problem that I see: S3K is zipping. It's a lot of zipping, level wraps, watching a screen fly by in fast motion until it finally lands at the boss of the area. The only way these runs differ is in the decreasingly often sections of actual gameplay, which most of the time is just to set up the next zip, and most of the time the setups are near identical except one method is slightly slower than another. This would be a MUCH different story without the zipping, if every character actually had to play the levels and were able to get through using their unique abilities. Going off of one of your earlier posts in the thread, I think the branch thing isn't mentioned in Movie Rules because it's almost never a problem, it's only in very few cases where a game could actually warrant the publication of several branches. I admit that's a flaw with the rules. It's mentioned in the Judge Guidelines, which is why the current Judges (and myself as a former Judge) are even talking about it in the first place, but I do agree that it should be more clearly communicated to TASers and audience as well. I understand that a game's community would see no reason not to have every possible branch published, especially in S3K's case where there are actually different playable characters that would normally lend itself perfectly to having a run for each character, but the community has to try to see things from the general audience's perspective. When it comes down to it, is there a reason to watch all 9 published S3K runs (competition mode is excluded here, Amy hack is included)? Is it worth it for a member of the general audience to specifically sit down and watch every single branch of this game? I'd argue it isn't, and part of my argument is that I already am a member of the general audience, and recently I've watched four of those runs, plus this one multiple times, and I can honestly say that there wasn't much difference to me, and I'm clearly not the only one who thinks that. Keep in mind that even if this run gets rejected, it doesn't mean it can't be looked at again and published in the future. The rules can change, and they HAVE changed, and runs have gotten second chances due to these rule changes and a lot of them have ended up published after all. Timeframe isn't an issue either: Look at when this run was submitted VS when it was finally published. 11 years. Rejection is only really the end of a run if it's noticeably suboptimal (which this run absolutely isn't) or if there are other glaring issues (emulation, flagrant rule-breaking, things like that, which again this run isn't any of those things). The branch thing isn't even so much a hard rule as it is a philosophy, even if there are times (like now) where we do enforce it as a rule. It's a really difficult case. I just hope that if it does end up getting rejected, it doesn't kill the community's motivation to improve the other runs, as there seems to be a LOT that the published runs are missing. Maybe we can take a look at this run again when the other categories are finally improved and it's a lot easier to compare? Otherwise, we can take a look at it again in the future when the rules change, and yes I'm fully expecting that to happen.
TASvideos Admin and acting Senior Judge 💙 Currently unable to dedicate a lot of time to the site, taking care of family. Now infrequently posting on Bluesky
warmCabin wrote:
You shouldn't need a degree in computer science to get into this hobby.
Sanqui
Any
Player (26)
Joined: 4/25/2011
Posts: 33
From an entertainment perspective, I do gotta say I would like to watch Zipless more than a new character branch like this
ovo
Memory
She/Her
Site Admin, Skilled player (1556)
Joined: 3/20/2014
Posts: 1765
Location: Dumpster
Samsara wrote:
Words. Good ones.
I agree with basically everything said here.
[16:36:31] <Mothrayas> I have to say this argument about robot drug usage is a lot more fun than whatever else we have been doing in the past two+ hours
[16:08:10] <BenLubar> a TAS is just the limit of a segmented speedrun as the segment length approaches zero
BigBoct
He/Him
Editor, Former player
Joined: 8/9/2007
Posts: 1692
Location: Tiffin/Republic, OH
I took some time and watched the movie in emulator, fast forwarding though things that were familiar from the S+T any%...and I skipped the vast majority of the movie. I don't think we need another any% run like this. However, I DO think we need more any% runs that don't use zips or terrain-ejection glitches.
Previous Name: boct1584
Reviewer, Active player (287)
Joined: 12/14/2006
Posts: 717
Personally, I don't there should be a hard limit on how many branches are published for each game, or even a soft limit along the lines of "when you reach this number of movies, it's time to start seriously question new branches" Instead, I like to focus on significant gameplay differences, or "compelling differences" like the Judge Guidelines say. In my personal experience, route differences are usually less noticeable unless they are drastically different, and I mean drastically. What stands out more and I feel is a better criteria is looking at the techniques used. Does a new branch consistently apply techniques that are not present in the current branches? Let's use the Ring Attack runs as an example, since I saw those contested earlier. One of the byproducts of collecting every ring is that a large portion of the level needs to be visited, so just naturally with that goal, some techniques, like for example glitching to the end of the level become less useful. I'm feeling the effect of that on the Aero the Acrobat 2 100% run I'm currently working on. In the any% run, wall zips were the way to beat most levels. However, in 100%, when I have to visit most of the map anyway, they become much less useful. It doesn't mean that the move is disallowed, just that the goal changes the usefulness of the technique. We've had a few runs rejected because they didn't apply techniques that weren't present in the published movies, even if it makes a different route or is highly optimized. A really good example of this is #5978: The_Tas_Master's N64 The Legend of Zelda: Ocarina of Time "All Medallions" in 53:21.70. While it's reliance on techniques that weren't present in other branches wasn't the only reason that movie was rejected, it's still a significant one. One that's more in line with this submission is #1572: mr_roberts_z's N64 Super Mario 64 "Coinless-Cannonless-Capless" in 1:03:10.80, which people also felt was too similar to published movies. Like a few others, I watched the currently published "Sonic and Tails" TAS and compared it with this TAS and the Tails only TAS level by level, noting where things changed. Keep in mind, that the casual viewer is not going to look at these things in that amount of detail. They are going to watch the entire movie and mostly go by feel, and if it feels the same as a published movie, it might as well be the same. Since Sonic and Tails has access to Tail's moveset, I was expecting Tail's techniques to be similar to the any% movie. I was pleasantly surprised to find that Tail's levels actually do feel quite different for the most part. Enough that I believe a new branch is justified for the category. Sonic alone, however, didn't. I think it was about 2/3s of the way through the Sonic 3 portion where I thought of a big reason for this. Everything that Sonic can do Sonic and Tails can do as well. It's just that Sonic alone doesn't have access to Tail's moveset. With this in mind, the run is less of a run with a different character and more like Sonic and Tails, but Sonic is not allowed to use Tails. Movies where the moveset of a character is limited in some way can, should, and have been accepted for publication if the limit leads to different solutions in almost every circumstance. Movies like [1868] SNES Super Mario World "no powerups, maximum exits" by PangaeaPanga in 1:18:23.22, [2062] N64 Super Mario 64 "70 stars, no Backwards Long Jump" by Jesus, Kaylee, MICKEY_Vis11189, MoltovM, Nahoc, snark, sonicpacker, ToT, CeeSammerZ, coin2884, Eru, Goronem, Mokkori, Nekuran, Nothing693 & pasta in 42:58.52, or to use a non Mario example, [3375] SNES Disney's Aladdin "capeless" by jaysmad in 16:36.78 are accepted because limiting the moveset like this removes a tool that was used constantly in the other runs to the detriment of all other possibilities. So oddly removing options actually leads to increased variety of techniques employed in almost the entire run. I noticed in the any% movie that Sonic relies on Tails a lot less than I initially thought, and unless he's using Tails, the solutions are going to be the same as a Sonic only run. In terms of different characters being justified for branches, let me use another game to help provide a barometer for my limits. In Super Mario Bros 2 (US version), I don't think there's any reason to have a Toad only, Luigi only, or Mario only run. Peach only is much more justified. Even though all four characters play differently, Peach is the only one with a noticeable unique ability. She can float, which leads to significant changes. I see no reason why the others would be really unique in that regard, even though Luigi's extra jump height does give him different route options in some levels. Also, very minor point that shouldn't have bearing, but I feel is worth mentioning: the speedrun.com leaderboard for this game lumps Sonic solo and Sonic and Tails into the same category. One small thing I find disappointing is the currently published any% is five and a half years old, and watching the submission in comparison to that, I can really see the difference that those years make in understanding the game and adding further optimization. The improvements this movie makes really should be represented in a published Sonic 3 and Knuckles TAS. I don't think this movie is different enough to justify a new branch, but if we could find any way to publish this movie without adding new branches, I think we should make that happen. I also would be interested to see a level by level breakdown for how this compares to the currently published any%, especially to see if it's actually faster on any levels.
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
arkiandruski wrote:
I also would be interested to see a level by level breakdown for how this compares to the currently published any%, especially to see if it's actually faster on any levels.
--------------------------------------------
| Level            |   Sonic   |   S+T     |
--------------------------------------------
| Angel Island 1   | 0:40::55  | 0:35::55  |
| Angel Island 2   | 1:14::51  | 1:14::53  |
| Hydrocity 1      | 0:39::16  | 0:18::37  |
| Hydrocity 2      | 0:22::20  | 0:21::58  |
| Marble Garden 1  | 0:34::37  | 0:50::56  |
| Marble Garden 2  | 0:03::24  | 0:03::24  |
| Carnival Night 1 | 0:29::19  | 0:32::28  |
| Carnival Night 2 | 0:39::17  | 0:40::22  |
| Icecap 1         | 0:56::45  | 0:59::06  |
| Icecap 2         | 0:00::00  | 0:00::43  |
| Launch Base 1    | 0:33::46  | 0:27::52  |
| Launch Base 2    | 0:26::37  | 0:25::50  |
| Mushroom Hill 1  | 0:33::37  | 0:32::27  |
| Mushroom Hill 2  | 0:40::34  | 0:32::50  |
| Flying Battery 1 | 0:45::26  | 0:42::49  |
| Flying Battery 2 | 1:05::43  | 1:06::13  |
| Sandopolis 1     | 1:04::51  | 1:10::13  |
| Sandopolis 2     | 0:48::47  | 0:57::19  |
| Lava Reef 1      | 0:47::43  | 0:48::48  |
| Lava Reef 2      | 0:20::52  | 0:20::53  |
| Hidden Palace    | 0:26::40  | 0:16::36  |
| Sky Sanctuary    | 0:52::17  | 0:52::13  |
| Death Egg 1      | 0:58::56  | 0:57::28  |
| Death Egg 2      | 1:36::23  | 1:37::15  |
--------------------------------------------
| Total            | 16:52::46 | 17:02::05 |
--------------------------------------------
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Banned User
Joined: 6/7/2017
Posts: 420
Location: Somewhere
Lots of good arguments here. Unfortunately I've kinda just run out of stuff to say. I could try to argue but I'd likely just repeat an argument I've already made. Also I have a compromise I wanna propose. So I still do want a new any% branch, especially since this would be the last one. However I am willing to compromise. Since some people don't feel this Sonic alone run is different enough from a Sonic and Tails run, why not just have this submission obsolete that one? I mean I did beat the total ingame time by a good margin (as you can see from feos' post), and if a new Sonic and Tails were to be made it could very easily beat my Sonic alone run.
arkiandruski wrote:
Also, very minor point that shouldn't have bearing, but I feel is worth mentioning: the speedrun.com leaderboard for this game lumps Sonic solo and Sonic and Tails into the same category.
It shouldn't have any bearing. In the rta Sonic category you aren't allowed to control Tails with a second controller. It's similar to how they do Sonic 2 runs. I opted to not use Tails at all because not having 2nd player control felt like a dumb restriction. Also they have a coop category which is the actual rta variant of a Sonic and Tails run.
Reviewer, Active player (287)
Joined: 12/14/2006
Posts: 717
feos: Based on that chart, this movie actually beats the published any% in terms of level times. Now this might sound like a completely crazy suggestion, but maybe it could obsolete the any%?
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
It doesn't actually sound crazy at all, since those 2 runs are also the most similar in content across all the chars: Post #494162.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Samsara
She/They
Senior Judge, Site Admin, Expert player (2238)
Joined: 11/13/2006
Posts: 2822
Location: Northern California
In-game time is definitely faster in this run. It's actually even faster than the table suggests:
Samsara wrote:
This run spends about 30 seconds in MH2 (starting from the end of the MH1 victory cutscene), while the published S+T run spends about 40 seconds in it (starting from using Tails to break the MH1 victory cutscene), both ending at the very first hit on Eggman.
So S+T essentially gets a 10 second head start before the IGT starts running in that run. S+T also get a 4 IGT second head start in Angel Island 1 from the cutscene breaking, meaning the total time save is actually closer to 20-25 IGT seconds based on pure comparable gameplay. I'd support this run obsoleting any% until a new any% comes around, with the caveat that this run is still branched as Sonic only and the description makes it clear that the improvement is purely IGT, the runs are close enough in content to warrant the obsoletion, and the realtime loss is solely because of a cutscene that solo Sonic can't skip. It wouldn't sit right with me any other way.
TASvideos Admin and acting Senior Judge 💙 Currently unable to dedicate a lot of time to the site, taking care of family. Now infrequently posting on Bluesky
warmCabin wrote:
You shouldn't need a degree in computer science to get into this hobby.
Banned User
Joined: 6/7/2017
Posts: 420
Location: Somewhere
Well I guess it's settled then. Just have this submission obsolete the current Sonic and Tails any% run. This even solves a lot of issues that were brought up throughout this thread. Now that I've thought about it some more, the idea of obsoleting a Sonic and Tails run with only Sonic sounds amazing.
Samsara wrote:
I'd support this run obsoleting any% until a new any% comes around, with the caveat that this run is still branched as Sonic only and the description makes it clear that the improvement is purely IGT, the runs are close enough in content to warrant the obsoletion, and the realtime loss is solely because of a cutscene that solo Sonic can't skip. It wouldn't sit right with me any other way.
I second this. I feel that the published video should still distinguish the fact that this is a Sonic alone run in some way.
Site Admin, Skilled player (1254)
Joined: 4/17/2010
Posts: 11475
Location: Lake Char­gogg­a­gogg­man­chaugg­a­gogg­chau­bun­a­gung­a­maugg
If anyone disagrees with this, please speak up!
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
EZGames69
He/They
Publisher, Reviewer, Expert player (4460)
Joined: 5/29/2017
Posts: 2761
I agree with this solution!
[14:15] <feos> WinDOES what DOSn't 12:33:44 PM <Mothrayas> "I got an oof with my game!" Mothrayas Today at 12:22: <Colin> thank you for supporting noble causes such as my feet MemoryTAS Today at 11:55 AM: you wouldn't know beauty if it slapped you in the face with a giant fish [Today at 4:51 PM] Mothrayas: although if you like your own tweets that's the online equivalent of sniffing your own farts and probably tells a lot about you as a person MemoryTAS Today at 7:01 PM: But I exert big staff energy honestly lol Samsara Today at 1:20 PM: wouldn't ACE in a real life TAS just stand for Actually Cease Existing
EZGames69
He/They
Publisher, Reviewer, Expert player (4460)
Joined: 5/29/2017
Posts: 2761
feos requested that I make a comparison encode between this movie and the published Sonic + Tails TAS. So I did that: Link to video Edit: I’ve changed the encode to remove the real time comparisons, as they’re irrelevant for IGT.
[14:15] <feos> WinDOES what DOSn't 12:33:44 PM <Mothrayas> "I got an oof with my game!" Mothrayas Today at 12:22: <Colin> thank you for supporting noble causes such as my feet MemoryTAS Today at 11:55 AM: you wouldn't know beauty if it slapped you in the face with a giant fish [Today at 4:51 PM] Mothrayas: although if you like your own tweets that's the online equivalent of sniffing your own farts and probably tells a lot about you as a person MemoryTAS Today at 7:01 PM: But I exert big staff energy honestly lol Samsara Today at 1:20 PM: wouldn't ACE in a real life TAS just stand for Actually Cease Existing
Invariel
He/Him
Editor, Site Developer, Player (171)
Joined: 8/11/2011
Posts: 539
Location: Toronto, Ontario
The side by side encode makes it clear that feos is making the right call here. Also, this is a spectacular run, kaan55.
I am still the wizard that did it. "On my business card, I am a corporate president. In my mind, I am a game developer. But in my heart, I am a gamer." -- Satoru Iwata <scrimpy> at least I now know where every map, energy and save room in this game is
Former player
Joined: 4/17/2009
Posts: 88
Can the Sonic Solo & Sonic & Tails runs be merged, so the fastest input is kept? Or is the game not hex edit friendly?