Sonic's Ability Discussion

I've played through GF1-RV1 with the air dash, no momentum mod.

Overall, it feels quite fun to use. The 'let go of movement to get the higher jump' feature did feel more intuitive than I expected it to be. More importantly, though...I think it's actually quite good to emphasize 'not moving' to new players! It's a great way to introduce players to the concept of 'maintain your current airspeed by letting go of the buttons'.

As Seaball said, it feels a bit too spammable (much like the Thok) on straightaways. This definitely feels like the sort of thing that would be fixed in an acceleration rework, though? It's hard to say whether the spam is better or worse than the existing thokspam; the regular Thok sends the player a greater distance, so they don't need to Thok as many times before they have to turn, whereas the jump thrust doesn't give the player as much of a speed advantage, but is still faster than running, so the player spams it even more times than they spam the Thok.

I played through GFZ-CEZ with controller + simplecam + no autobreak. I did catch this particularly cheeky jump with the double jump. It's not quite a timesave, and you can probably use the spring this way without the move, but it's among the pettiest things I've ever done with a double jump.



I wanted to try it with autobreak on, too, so I turned it on in ACZ. I have not tried autobrake before. I hate it. It's not a bad feature, I just personally really, really hate it. It threw off all my jumps and made the game feel even more unnatural than switching from m&k to controller.

The move does feel a bit worse with autobreak on, though I suppose there's some merit to slowing oneself down when using the platforming-focused higher jump. It does have me wondering if this version of Sonic, which encourages letting go of movement, would also encourage making autobrake off the default option? It's weird.

The move feels slightly too mobile, and there were quite a few situations where I could jump over the level. Still, most of them felt like they took some skill to pull off, so it's not that bad? I guess it's not terrible if Sonic can do some skips, and it's virtually impossible to make a jump correction move that doesn't give Sonic some skips.

I like to use RV1's mandatory monkey-ball sections as a barometer for how mobile a character is. There's two mandatory sections, and I generally don't want to use the boulder on replays, so I try to skip them on every character. The question, then, is whether they can skip them without taking damage.



As far as I can tell, Thrust Jump Sonic can't skip the first one! So his mobility isn't that overpowered. He can't skip the second one, either.

On a tangential note, I used to consider "whether a character can skip the second monkey-ball section while taking damage" to be a metric, too. Only just now discovered that Thok Sonic can do it!



So Amy's the only one of the base game's characters that has to take the ball. I think. Most characters with some kind of speed boost or high jump or limited flight can make it through this section alive. Not as useful of a metric anymore.

I should probably clarify that the dash has a maximum speed cap of 60 fracunits. So it is possible to reach thok levels of speed, but you can't exceed that amount.

Unknownlight said:

I think the main problem with the homing attack isn't even what Mystic said, it's that the homing attack is focused downward. It's meant to hit enemies below you. It doesn't help with vertical movement, and I think any thok replacement should be useful enough that new players are encouraged not to abuse Flight Call as much.

Click to expand...

For sure. And that's why I favor homing's implementation as "Auto Aim" that can be disabled, like "Autobrake," and is otherwise the standard Double Jump. It doesn't help with upward vertical movement, but it does help with downward precision in targeting springs, enemies, item boxes.

As much as I agree with the idea of NuMetal being Thok-only, and while I advocate for Homing/Autoaim and Double Jump for reasons already stated, I would be a bit disappointed if this came at the expense of Thok on Vanilla Sonic.

Again, almost all of my suggestions for revising Sonic have included finding a way to keep his Thok in tact in some form. In case it wasn't obvious! I'm very much for reforming Sonic's abilities within the two button scheme, by simply supplementing the Thok with handicap options that individually have their merits and drawbacks but together form a more cohesive whole, an accidental tapestry that becomes more than the sum of its parts.

I like the Thok conceptually very much because, in my mind, it's the Spin Dash but in mid-air. In-universe, I picture it as that Sonic does the same gymnastics either way, but one is on the ground and one is in the air.

I think it's much more fun to hop n bop Crawlas than roll/spindash through them. It's pointless to thok through crawlas and a lot of enemies, though it feels good on the enemies you can thok through. For the player to succeed, it's best to quickly develop an approximate measurement of the thok distance in your minds eye from memory asap, plus gain understanding of Sonic's hopping arcs to pop crawlas and land on item boxes and springs.

But, it's not ideal that a first time Sonic player will mess up the 3 spring jumps in GFZ2, almost invariably. Messing that up doesn't make the player think "My bad. Great vertical level design." It makes them think "This sucks, what the hell." Homing/Auto-Aim helps with that. And once you've got it, you can switch it off. It allows the player to feel good playing Sonic even if they aren't good at the game yet. But the more they enjoy playing it, the more they'll play, and then the game will encourage the player's skill development with bonus characters, bonus levels, and bonus objectives.*

The New Player feels bad for not being able to pull off a basic Sonic maneuver, rather than like a tip of the hat to the level designers. (There are plenty of great "oh wow, you got me" moments later, anyway!) What seems obvious and intuitive to a player/dev who's been with the game for many years and versions may not be to someone who saw a YouTube playthrough who wants to play "classic Sonic but in 3D, kinda like Sonic Utopia but a full game and made in a modified Doom Legacy engine." "What, no way? I wanna give this a try."

I've seen multiple new players almost give up at the 3 springs in Greenflower 2, because before failing the spring jumps, they also had a hard time visually locating the first spring to begin with and was confused that the next move was to ascend over the waterfall vertically using the springs.

I'll tell you about one in particular, my good friend Nic Estrawman. Name changed to protect the innocent. So this person is largely ambivalent about Sonic but charmed by the design, fandom, and DIY attitude of fangames, and saw how much I was enjoying it. He has played most 3D Mario games to completion, but is otherwise not playing many platformers since childhood. The last Sonic game he might've played was probably Sonic 2 or Sonic 3 as a kid in the 90s, and maybe played Sonic Adventure 2 on Gamecube as a teenager. Now they play strategy or simulation games on PC, and some Nintendo games on Switch. They're a grown ass adult with other responsibilities. They're not exactly hardcore, not exactly casual. Somewhere in between.

The question is... how much do you want to encourage that kind of player, as a game developer? How much does the game care about this hypothetical demographic of player? The one who is a little bit of a longshot to hook, but is kind of on the line, and could be persuaded by having a strong favorable first impression. The one who might be a tease, a casual, too fickle and impatient to be worth any consideration. Hey, if they're not willing to meet the game more than that, why bother thinking about them any further?

I like the philosophical part of design discussion.

My view on Sonic as the core idea of what the character/game/franchise is about is that it is works best with cheese. The player who can cheese feels good; remembers the feeling, and the adrenaline thrill of running and spinning and succeeding, the player returns. And as they improve, they can experiment with different characters and different options in the menu to customize their difficulty experience to their liking.

Which is more important? That SRB2 is accessible to new players, or that SRB2 includes nothing from the official 3D games that weren't in the Genesis games to make a somewhat vague and very conjecturable point about how the Sonic games have lost their way. Not to say I'm not sympathetic to the point, I am and I agree with the approach, but as I mentioned before, the level design in SRB2 has already proven its point. Play the SRB2 levels back to back with SA1 or SA2 or any 3D Sonic and find yourself not preferring SRB2, because of the level design.

It's cutting off your nose to spite your face. Sonic Team went the wrong way from Adventure forward, but doesn't mean that homing isn't a good idea for Sonic in any 3D game, whether intended or not.

Homing is like the Spin Dash or Drop Dash... it's nice to have both in the Sonic 1 and Sonic 2 re-releases, and if they ever released S3K, I'd like the DD in there too; any 2D Sonic feels incomplete without it now.

The "auto aim" homing could work exactly as it does with the current Magnet Shield, just as part of Vanilla Sonic's default with Double Jump.

I've said it before but SRB2 feels like it could've been the Super Mario 64 of the Sonic franchise, 25 years later, and it's why it's great; SM64 has a far worse camera and is still an all time 3D platformer that's just fun to run around and play.

Mario in SM64 doesn't have homing, but he has a new arsenal for 3D in that game. It would be a very difficult game for children to play without the punch and the butt stomp and the double and triple jump. Obviously its quite different from the NES/SNES but also it isn't at all. Later games like NSMB3Dworld on WiiU and Mario Odyssey on Switch found other ways to replicate the design and gameplay of the 2D games into modern 3D.

Are the Mario games too complicated with his default moveset in any 3D game? Definitely not, right? So why couldn't 3D Sonic have more than just Spin Dash and one other special ability?

*

The mods with double jump or homing that I've played — Neo, Advance — wouldn't be suitable replacements, because those lack the horizontal mid-air burst. It replaces horizontal axis burst (Thok) with vertical axis burst (Homing). It trades one problem for another, and becomes a coin flip in my preference. I usually stick with Vanilla/Thok Sonic over Neo or Advance, and usually over Modern. Neo Sonic's homing range seems to extend too far, but the double jump is nice. Advance's homing also feels too loose, and the Boost mode feels like it should be with Metal, along with NuMetal Thok.

The Modern Sonic Mod, ironically, has it closer than anyone, but it's dismissed because of the branding, and the problem of having too many extra button abilities — although I do think the full Mod abilities would make sense for Super Sonic, and would provide a tangible advantage for achieving Super form, because now Sonic with stomp and wall jump can go on Knuckles/Tails paths that he couldn't reach before, making him the Greatest Of All Time. As it is now, I like going Super for the novelty but it's not actually helpful and is more of a hindrance, often resulting in unintentionally shooting off course.

But the Modern Sonic Mod is close, because it has the Colors Double Jump and the Homing on the "A" button, and if you only use "B" button Boost in mid-air in taps instead of holds, it's almost like the Thok for the horizontal forward momentum. Sonic can go up and down (double jump or homing) the vertical axis with "A" button, and forward horizontal burst with "B" button, either on the ground or in the air (spin dash or thok).

Alternatively, Thok could remain as Jump, Jump, and then Double Jump/Homing is Jump, Spin, so the shield power replaces Double Jump/Homing instead of thok. I feel like this one doesn't make as much intuitive sense to a newcomer but I think it'd be an easier adjustment for veteran players.

I just honestly believe the compromise solution is the strongest! <Sonic>Friendship!! Yeah!! Way past juicin'!!</Sonic>

Triple Dash (c'mon it's got a... ring to it)
A button - Jump
B button - Spin

hold B, release B - Spin Dash

A, A - Double* Dash (double jump with homing/autoaim enabled by default, works like the shield so momentum issue is as intended by devs; double jump bounces upward like Colors unless targeted on nearby object, then goes downward to meet it)

*
A, B - Thok** Dash or Shield Ability (when equipped)

**

Spin Dash, Thok, Boost... these are variations on a common theme. Immediate forward momentum. Ground or air. Spam/mash the button (thok/homing) or hold (boost/spin dash). Both have their place.

And that place is on each and every one of your controllers! Two buttons! Three dashes! Endless potential!

Imagine Sonic is like the game mascot version of one of Yuji Naka's sports cars. Gotta go fast. Spin Dash and Thok are like driving stick; Boost is automatic. Now, I like stick. I can see the value of automatic, but make no mistake. By adding Auto-Aim/Homing, this isn't switching everything to "hold button and forward and win" automation. It is a bit more "mash button and forward to win" but it always was, spamming the thok is what speedrun players do. Auto-Aim is a safety feature, like adding sensors that can automatically guide the steering away from nearby objects for protection. You exchange a bit of control in exchange for a smoother experience. It's not throwing away the whole design and starting over. You're still driving stick but the handling is more precise, a safer experience while you're racing through wacky courses. There will be fewer deaths.

Fewer deaths for Sonic. Can we do it? Yes we can! <Obama 2007 campaign voice>For you see, there is no modern Sonic, there is no classic Sonic... there is only one united state of Sonic the Hedgehog. Simple! Easy! Greatly Accessible! </O07> SEGAAAA

Putting the homing attack in the game and turning it on by default doesn't make new players think "I'm not supposed to be using the homing attack, but it's there until I get better". It makes them think "this is a game about using the homing attack".

I've been thinking about it since my post earlier, and I may have come up with a better solution than I had then. The moveset I have in mind for Sonic at the moment is as follows:

Spin on ground while not moving= Spindash
Spin on ground while moving= Spin Attack
Jump = Spin jump
Jump + Jump = Air Thrust
Jump + Spin = Thok (Without shield) or Shield Ability (With shield)

The Thok could be used in any direction (Simple mode) or only forwards and backwards (Standard/Legacy). I also still feel like having it so that it allows you to keep your speed if you are holding a movement direction when you land (or even just forward) would work really well for preventing thok spam.

This isn't entirely ideal in that shield abilities would end up replacing the thok, making it so that the flame shield just feels like a worse thok replacer and the wind shield just feels like a different air thrust (Basically it would occasionally feel kinda redundant) but it's probably preferable to having the Air Thrust on Jump + Spin as what the Air Thrust lacks in instant speed, it makes up for in mobility that the shields can't always account for.

As for Metal Sonic, while I am of the mind that his hover could use some tweaks, I would sooner keep the same hover we have now for him than replace it with the thok. For reasons I have already mentioned previously about the thok being unlockable, and also because I don't really see a lot of hype for a player to unlock a character that is basically just the same thing that used to be available from the start, except maybe with a situationally useful dash mode.

Magnemania said:

A novice wouldn't be able to do this? Turning backwards to save oneself is an extreme use case, but there's plenty of situations where using a tool that quickly sets Sonic's horizontal and vertical momentum in any direction to a consistent amount helps with platforming.

Click to expand...

Maybe if they planned to do so, then yes, with a good chance of failure. Unplanned, though? Absolutely not. There is no way that a new player can possibly jump back to the platform there, for a variety of reasons.

I think a lot of people here are severely underestimating the problem here, and also misunderstanding what are and aren't helpful for those players. New players try out our game with a variety of prior knowledge and experience in platforming games, but some things apply and some things do not. New players all need to learn some things in order to be effective at the game, and those things vary by the player. I'll go over the two most important things here.

First is understanding the rules of the game. Most players will enter our game with some basic stuff already covered. For instance, I doubt many players would pick up a Sonic fan game without already knowing that as long as you hold some rings, you won't die to damage. Some things are a bit more obscure and many players will learn them first here, having not played the original games as much as others. Other things are exclusive to our game and will need to be learned from scratch.

The most important thing that players will need to learn from scratch is our game's physics and jump arcs. For instance, if you're a new player, and you see this monitor, can you jump up there as Sonic?



I'd wager most players here can just glance at that and know that they won't quite make it, but that knowledge isn't obtained instantly from booting up the game. That knowledge is learned from experience by playing, which for most new players means they will attempt to jump up and realize through testing it that they cannot make it unassisted. Here's another situation in GFZ1:



The trail of rings here implies that you can, indeed, make that jump. How do you do so? There are a couple ways, but learning the rules to the game is something that new players will need to do before they're remotely able to play effectively. This area is designed to let players test out various ways of getting up there in a safe, penalty-free environment until they find a method that works. Both of these situations in GFZ1 exist on purpose to teach new players the game's rules through trial and error. We want new players to try various things to learn what they can and cannot jump to. Getting players to try out their abilities and figure out how to use them well is super important to the learning process. That information becomes super important later on when there are actual penalties for trying to make a jump and failing.

The reason this is important is because new players are extremely unlikely to have memorized the rules of the game well enough to even make the connection that they can turn around and thok back to solid ground. In fact, I've watched several new players in this exact scenario; the predominant decision new players make in such a situation is to instead thok forwards, because while they know that action is probably futile considering they can see in front of themselves, they still try anything they have access to because obviously doing nothing isn't going to help. The realization that turning around and thokking backwards is a reasonable thing to try just doesn't even occur to players until they've started to achieve mastery of the game's rules.

Second, and more relevant for twitch reflex, is something called "muscle memory". Over time while playing games, your brain starts to remember the repetitive patterns we use to accomplish tasks. In this case, the hand motions you use to control the game are being ingrained while playing, making it such that expert players never have to question what input does what action. A simple way to know when you've put a game input into basic muscle memory is when you stop thinking about what key to press to do something entirely and only think about what action you want the character on-screen to take instead. Your brain has already memorized what your hands need to do in order to take that action, and the details of how to execute that action aren't thought about anymore.

Quite simply, SRB2 has a complex control scheme, even in simple mode with a controller, that takes longer for most new players to get into muscle memory than many other games. A new player having gotten to ACZ1 for the first time is incredibly unlikely to have gotten the game into muscle memory, and therefore is not going to be able to make the decision mid-jump that this jump was a bad idea and turn around to land back on the platform. This is because of that intermediate step in there where the brain is still translating from the idea they had to the inputs they need to execute it. In that time, the player has long-since fallen into the pit.

This is why I say that your metric is a bad one and that none of the Sonic abilities demonstrated would save a new player from dying. They would need both good understanding of the game's rules and muscle memory to pull such a thing off, and new players generally have neither.

time gear said:

Crash Bandicoot for example. In Crash 3, you have a double jump, a body slam, a tornado spin, a bazooka that shoots wumpa fruit, a crouch, a slide, a slide jump, sprinting, etc. You are given access to a ton of different abilities, but I have never heard anyone say "Crash uses too many buttons and button combinations!"

Click to expand...

This is because Crash 3, like many games with large movesets, staggers granting abilities to the player over time. You don't start out with all of your abilities from the start, but instead the game teaches you only the basics from the start and gradually introduces new abilities over the course of the game. Another common example of having a huge ability set, Mario 64, doesn't prevent you from using the abilities from the start like Crash does, but intentionally teaches you over time about the various abilities through signs in front of areas that require them. For instance, the designers just assume the player doesn't even know about the long jump until Hazy Maze Cave where it's explicitly explained and required to progress.

This is a great solution for games that intend to have a massive set of abilities with the time to explain them all one at a time, but SRB2 does not fit that description. Not only do we have a ton of different characters with variable abilities, we do not have the type of game design that allows for stopping after beating a level and telling the player about some cool powerup ability they just got, and then teaching the player through a mandatory section that requires its use. Hence, we need to keep our ability setup to much simpler things that players can understand from a small blurb of text. It's not just to fit it into the character select screen, but it's more that anything that wouldn't fit onto that screen is way too much for a player to internalize at once and remember for later on when it matters.

Mystic said:

This is because Crash 3, like many games with large movesets, staggers granting abilities to the player over time. You don't start out with all of your abilities from the start, but instead the game teaches you only the basics from the start and gradually introduces new abilities over the course of the game. Another common example of having a huge ability set, Mario 64, doesn't prevent you from using the abilities from the start like Crash does, but intentionally teaches you over time about the various abilities through signs in front of areas that require them. For instance, the designers just assume the player doesn't even know about the long jump until Hazy Maze Cave where it's explicitly explained and required to progress.

This is a great solution for games that intend to have a massive set of abilities with the time to explain them all one at a time, but SRB2 does not fit that description. Not only do we have a ton of different characters with variable abilities, we do not have the type of game design that allows for stopping after beating a level and telling the player about some cool powerup ability they just got, and then teaching the player through a mandatory section that requires its use. Hence, we need to keep our ability setup to much simpler things that players can understand from a small blurb of text. It's not just to fit it into the character select screen, but it's more that anything that wouldn't fit onto that screen is way too much for a player to internalize at once and remember for later on when it matters.

Click to expand...

I did actually take this into consideration, which is why I followed up with the Spyro 3 example. However, I can also simply give the example of Crash 2, which has a moveset almost as expansive as Crash 3 and you start out with all of it. Right from the start of Crash 2 you have access to run, jump, crawl, slide, spin, slide jump, belly flop, and crouch jump. In N.Sane Trilogy's version you eventually unlock the sprinting ability, but let's ignore that for the sake of the argument.

If we were to take SRB2's inputs for Sonic into account, the following inputs would likely be what would be used if Crash 2's Crash were to be ported into the game:

Jump - Jump
Spin - Spin
Jump + Jump = Belly Flop
Jump + Spin = Nothing

The vast majority of Crash's moveset is left unaccounted for under these limitations. That is of course to be expected, Crash wasn't designed to be functional within a Sonic game (Though Sonic was one of the inspirations, as evidenced through the well known "Sonic's Ass Game" title the original Crash game had during development). The point I am going for is that both Spyro 3 and Crash 2 exist as examples of 3D platformers that use way more button inputs than SRB2 from the start without complaint. Spyro 3's moveset is even the exact same as Spyro 2's final moveset once you have unlocked everything. They didn't take away any of the unlocked moves going into the third game. I have never actually played any of the Banjo games, but from what I have heard Banjo-Tooie does something similar as well.

While it's true that a player isn't going to have a proper grasp on everything from the start, having access to a few more moves from the start isn't something that a player is incapable of getting used to. There's even a number of character mods released that make use of the additional inputs provided by custom 1, 2 and sometimes even 3 such as Modern Sonic towards which I haven't seen any complaints about too many inputs.

The distinction here is that the structure of Super Mario 64 and its clones actively helps with the increased ability pool. Those games all introduce things one at a time, either through unlocking them or by teaching them in the levels. Sequels back in those days were generally made with the actual assumption that most players would have already completed the previous game already, so the increased moveset existing from the start in various sequels wasn't contrary to the design. In fact in most of those cases that assumption was even correct, so it wasn't even unreasonable to do so.

Basically, we don't have the equivalent of the signs in Super Mario 64 to stop gameplay when you encounter a new hazard that needs you to use a new ability to handle it, and personally I think that type of structure wouldn't suit SRB2 to begin with. There's a reason we went with an isolated tutorial stage instead of building signs and tutorial box style teaching into the early stages instead. We totally could have gone with the equivalent of Omochao to teach mechanics, but we chose against it for a number of reasons.

I just tried the file uploaded by Cobalt, I pretty much felt betrayed. Like if I just throw more than 15 years to the garbage, learning the thok and all that since I was a kid, zooming through zones with it and mastering it until I got to complete SRB2 1.0 - 2.0 - 2.1 in so little time ... These are just feelings of course, no arguments, yet:
Sonic started feeling clumsier when I stopped spamming the thok (Spamming thok is not just spamming the jump button guys, you actually gotta learn a timing and to use spindash to gain more speed, learnt it in 15 years of thok "spamming"), was not as precise as the good old thok. the double jump seems a bit not accessible, I have to stop all my inputs in order to it appear and that just doesn't help when platforming, just makes things more complicated and I avoided using it because of it. When using the Dash, it just didn't help going fast and stopped using it too.
I just think FSonic moveset has a better option for the game, the dash adds height as a reward for keeping the momentum, the thok is not that necessary for beginners but if you want to you could master it, the instashield seems hard to use and all, but that's left there for more experienced players.

Mystic said:

Basically, we don't have the equivalent of the signs in Super Mario 64 to stop gameplay when you encounter a new hazard that needs you to use a new ability to handle it, and personally I think that type of structure wouldn't suit SRB2 to begin with. There's a reason we went with an isolated tutorial stage instead of building signs and tutorial box style teaching into the early stages instead. We totally could have gone with the equivalent of Omochao to teach mechanics, but we chose against it for a number of reasons.

Click to expand...

SRB2 definitely doesn't need more explanations or text, and definitely not Omochao. More than any of the proposed movesets, additional tutorializing would really drag down the game's feel as a 3D successor to the Genesis spirit.

But SRB2's levels are chock full of ability lessons and tests. Your post above showed two of them in Greenflower, where Sonic's natural jump isn't the solution.

In that example of GF, the goal is "hit the box that is unreachable by normal jump." Even with Vanilla's current moveset, there is more than one possible solution. So, the lesson is "some level challenges will require thoughtful consideration of how you achieve vertical movement, where the ordinary jump will be insufficient."

So, if the data says — most (new) players are struggling with this lesson — then how do you help the player so it is more likely they will achieve the goal, and learn the right lesson in doing so?

1. More text/Omochao is very direct, but inelegant and unartistic.
—Nobody wants this solution, I'm sure we all agree.

2. Change the level design so each specific Sonic lesson is more forgiving to the existing movesets.
—Possible solution, but more work than is necessary for uncertain results.

3. Change the existing moveset to provide the player with more abilities so there are more possible solutions to the existing challenges.
—The goal is more easily achieved, and lesson more likely learned.

But wouldn't there invariably be players who use the added abilities to achieve the goal, but learn the wrong lesson? Yes. This is the same problem of Tails' Airlift, but there's a key difference.

Airlift teaches the Sonic player to rely on Tails, skipping any lesson about Sonic's abilities in favor of learning how to use Tails' flight. In addition to being a bad solution from the Dev point of view for how it disengages with the level design, it's also a bad solution from the Player perspective, because calling Tails is a small admission of defeat. Everyone knows it, even as they're using it. Any player would of course prefer to kick ass as Sonic, rather than be carried around by Tails.

Doublejump/Homing empowers the player while keeping them as Sonic. Which means the next time the player encounters a similar dilemma in a later level, they will be more likely to engage & experiment with Sonic's abilities (having previously experienced success with that general method), rather than giving up and calling Tails.

Answering Magnemania's questions

Magnemania said:

Is there an existing double jump in the game that you're describing? The only double jump I'm aware of that's currently in the game is the Whirlwind Shield, so I'm basing my impresssions of the move on that.

Click to expand...

Check out this SRB2 wiki page to learn about all abilities that currently exist in vanilla SRB2. Or use in-game cheats. Or both.

Magnemania said:

You can't isolate elements of a character's toolkit by the inputs used for them. Are Amy and Fang's higher jumps their "ground jump" ability? Does Knuckles have a strictly worse "ground jump" ability than Sonic/Tails/Metal Sonic? There's hardly even any commonality between the inputs between characters; every character except Metal Sonic can use their second jump as an attack, Sonic/Tails/Knuckles/Metal can use their 'spin' as a means to gain speed whereas Amy's 'spin' slows her down and Fang's 'spin' can't be used while moving.

Click to expand...

Isolate? Well, maybe. I just want to say that Spin ability doesn't necessarily have to be helpful for platforming. Fang's moveset includes a "strictly platforming-based" ability (Spring Tail) and a "strictly combat-based" ability (Popgun). If you want me to tell you how useful it is - Popgun is comparable to Tails' flight: while Tails can skip most levels by flying over difficult obstacles, Fang can skip most bosses by shooting them to death without getting dangerously close to them. It slows you down severely because you have to completely stop? Yes. The same goes for Spindash, to a lesser degree. But if you play the game you probably don't use Spindash all the time as soon as you touch the floor. Same with Popgun. You don't have to go on a robot killing spree just to extract some sort of value from it.
And when it comes to Amy... She probably needs a thread of her own for that discussion (and it's probably not worth starting one just for that).

Tried Cobalt's Air Dash mod. I didn't play much with it so I don't really have the right to judge but it certainly is an interesting ability.

XleederTH said:

I pretty much felt betrayed. Like if I just throw more than 15 years to the garbage, learning the thok and all that since I was a kid, zooming through zones with it and mastering it until I got to complete SRB2 1.0 - 2.0 - 2.1 in so little time

Click to expand...

That's exactly the reason some veterans defend the option of letting it stay as it is. "It sucks because it's new/different".

Thoughts on Cobalt's ability:

I really appreciate the variety of movement this allows for. The extra height from the neutral jump came in much handier than I expected it to, and it works very well with SRB2's lack of air friction. The additive momentum from directional jumps made building up speed feel much more rewarding when I was able to manage it. ACZ1 was particularly fun with so many downwards to slopes to get an initial jump off of and then boost it even more with the directional jump.

But for a new player, I think the most useful thing here would be the neutral jump - that extra height does far more for correcting jumps than the horizontal momentum variations. However, and I might be totally wrong on this, neutral-ing the controller is not something casual gamers are used to in platformers. I'd like to hear other's thoughts on this, but it seems to me that the instinct for a lot of players is to hold a directional input, especially when they panic. Another related problem is this would further reduce parity between Autobrake On and Off. I play with Autobrake Off and I can't imagine the neutral double jump being nearly as useful if it Sonic tried to brake whenever you let go of the directional input.

With regard to the horizontal jumps - like I said, I had a lot of fun with them when I got everything to line up. However, it felt to me like it worked best on straightaways - turning corners I'd do one of two things: undercorrect and fly into a wall or off the stage. Or I would make the turn accurately but lose most of my momentum in the process. This could just be a factor of learning a new ability for the first time, but I do think it could do a little more to redirect momentum in new directions; maybe I'm just used to Pointy?

Finally, holding backwards for the double jump never did the trick for me - it always felt like what we would call a "limp dick jump." Sonic would gain too little extra height to make any difference, and rather than helping Sonic gain speed to get back to the platform behind him, it would just make him kind of stall out in the air and fall to the ground. Maybe it'd be too complicated or even too powerful, but I'd consider holding backwards convert existing forward momentum into vertical momentum, allowing Sonic to reach greater heights from faster speeds, at the cost of that speed. Almost like a glider of sorts.

Rather than "holding backwards", I'd say "inputting in a direction opposed to Sonic's current momentum". Simple controls don't really have a backwards.

Boost said:

solate? Well, maybe. I just want to say that Spin ability doesn't necessarily have to be helpful for platforming. Fang's moveset includes a "strictly platforming-based" ability (Spring Tail) and a "strictly combat-based" ability (Popgun).

Click to expand...

The Spring Tail is not strictly platforming based, it allows Fang to bounce on enemies. It's Fang's only method of attacking that doesn't require him to completely kill his momentum (the act of bouncing puts him at a temporary stop but preserves his momentum). The existence of the two moves in the same moveset is fairly important; Fang is forced to choose between coming to a complete stop to safely kill an enemy, or maintain his spring bounce and either dodge or land on the enemy at a high speed.

If Fang had the spin dash instead of the popgun, his bouncing would be devalued because he could pick up speed and attack enemies without bouncing. Fang also loses most of his momentum when he starts his spring chain, so the spin dash wouldn't synergize with his it. Comparatively, Tails and Metal Sonic preserve their momentum when they use their double jumps, so the spin dash synergizes with their double jump. Knuckles comes to a full stop after using his ability, so spindashing is the natural follow-up to his glide. Sonic, on the other hand, has very little synergy between his Spindash and his Thok because both are attacks that launch him forward; but the Thok doesn't require any setup and launches Sonic farther, so the spindash only tends to be useful in context-sensitive situations with slopes.

The entire kit is connected.

Having now had a chance to try out Cobalt's airdash thing, I really like it, but I agree that neutralling the stick to go upwards is a bit clunky and doesn't fit with how people prefer to play platformers. Holding backwards doesn't really work with simple controls, so what if it goes horizontal or vertical depending on how long you press the button? Hold (more "jump" in the input) to go high, tap (more "speed" in the input) to go long. A panicking new player is more likely to hold the button than to tap it lightly, and a vertical jump is generally less likely to get them into trouble if they weren't already in it.

I've decided to try making the game as unfamiliar as possible to test out the Air Dash as a mistake-fixer. I normally play with M&K + Traditional Camera + No Autobrake, so I've done the following things:

• Playing on controller with simplecam
• Turning on autobrake to make the physics feel off
• Playing through Shrewd Blizzard Zone Act 1 from the CyberDime Realm level pack, which I don't have too much experience playing
• Playing with the controller upside down

The last one, of course, is more dramatic than the others. Both directions on the control stick are inverted, so aiming feels very difficult, the control stick is on the wrong hand (I hate using arrow keys/numpad because they put movement on the right hand), the jump button is a bit harder to reach, the spin dash's relative position to the jump button is inverted, and I can't center the camera (I put it on shoulder) without letting go of everything else.

It's impossible to completely handicap oneself to feel like a new player, since the game knowledge is still there, but this almost completely destroys muscle memory. I've gotten the idea from some developers at Insomniac Games, as apparently a few of them did this to test difficulty after they got too good at their own games.

I have recorded the playthrough here:

https://www.youtube.com/watch?v=-myAQRZGRSI

It is seven minutes long, partially because I can't confidently go fast with these controls, and partially because I have no sense of direction and it's harder to keep track of where I've already been when I'm struggling with the controls. Key moments include:

0:14: Immediately messed up the inverted horizontal controls. With about 2 seconds in the air to figure out what I was doing, managed to save myself from the pit with the Air Dash. This is a 90 degree turn rather than a 180 degree turn, which is much more likely for a new player to manage.

1:15: Tried to make it to the next platform with the Air Dash, misdirected it and fell.

1:35: Failed the same jump again, this time succeeding with the Air Dash.

3:05: Safely air dashed into a wall in a ice-platforms-over-a-pit section to come to a safe stop.

4:40: I start using the double jump aspect of the air dash to make some unintended jumps, which causes me to start going backwards. Similar problem to what first-time players tend to experience when playing as Tails and Knuckles.

6:30: Finally figure out which way is forward.

Mystic said:

The distinction here is that the structure of Super Mario 64 and its clones actively helps with the increased ability pool. Those games all introduce things one at a time, either through unlocking them or by teaching them in the levels. Sequels back in those days were generally made with the actual assumption that most players would have already completed the previous game already, so the increased moveset existing from the start in various sequels wasn't contrary to the design. In fact in most of those cases that assumption was even correct, so it wasn't even unreasonable to do so.

Basically, we don't have the equivalent of the signs in Super Mario 64 to stop gameplay when you encounter a new hazard that needs you to use a new ability to handle it, and personally I think that type of structure wouldn't suit SRB2 to begin with. There's a reason we went with an isolated tutorial stage instead of building signs and tutorial box style teaching into the early stages instead. We totally could have gone with the equivalent of Omochao to teach mechanics, but we chose against it for a number of reasons.

Click to expand...

This isn't actually entirely true. To begin with Crash 2 is anything but an SM64 clone. It's far more akin to the classic 2D Mario games than it is to SM64. Furthermore, while sequels back then did often expect that at least a portion of their playerbase beat the previous game already, they were also designed around the fact that this wouldn't be true for every player. Crash 2 in particular is a notable case in that the game is actually much more lax with it's difficulty compared to the previous game. It was treated as a sort of soft reboot to the series, a second chance to draw more new players in with more accessible difficulty and new features.

Crash 2 never needs to stop the gameplay to teach you how to use Crash's moveset. All of the tutorial content in the game is invisible. You have an incredibly brief level to familiarize yourself with how to move Crash and the basic premise of the gameplay, and then you are thrust straight into the first warp room to choose any of five levels to take on next. The game knows that a player is going to feel experimental and figure out what Crash can do on their own. Players learn much akin to the GFZ1 examples you posted.

This is noteworthy because Crash 2 and SRB2 are much more alike to each other than they are to SM64. They each follow a design mentality around getting to the end goal to finish the level, rather than completing given objectives. You do need to collect all the Crystals in Crash 2, but the vast majority of these are placed right along the main path and can't be missed.

Boost said:

...That's exactly the reason some veterans defend the option of letting it stay as it is. "It sucks because it's new/different".

Click to expand...

That's true, but I try to give some suggestions and not be that guy; hard to avoid when all the suggestions lead to Sonic being just underpowered or just plain bad.

---------- Post added at 02:08 PM ---------- Previous post was at 01:56 PM ----------

Kaysakado said:

...But for a new player, I think the most useful thing here would be the neutral jump - that extra height does far more for correcting jumps than the horizontal momentum variations. However, and I might be totally wrong on this, neutral-ing the controller is not something casual gamers are used to in platformers. I'd like to hear other's thoughts on this, but it seems to me that the instinct for a lot of players is to hold a directional input, especially when they panic. Another related problem is this would further reduce parity between Autobrake On and Off. I play with Autobrake Off and I can't imagine the neutral double jump being nearly as useful if it Sonic tried to brake whenever you let go of the directional input.

Click to expand...

I can confirm this, I cannot get myself to naturally stopping all direction inputs because I want to double jump, and if you ask me, I find it actually better to not even use the double jump at all because it made the game more complicated than it has to be, I just flat out ignored it and kept playing without any double jump (managed to complete the game without it too); I think the thok should stay and people have to learn to play the game instead of crying because they can't instantly be good at it, everything needs a little bit of effort and practice.
As I said, I instantly found it better off not using that double jump nor the dash, it just felt clumsy and manoeuvring was little bit hard to understand (Coming from a really old player... Complaining about manoeuvring and doing tricks, isn't it weird?).

Made some more tweaks this morning.

• Changed how vertical and horizontal thrust is applied (should feel more natural while also allowing for more technical play)
• Increased air maneuverability while in dash state
• Improved spindash scaling
• Some sfx and gfx for good measure
• Thok on Metal (for comparison)

I gave the new version a try. I tested it in both simple and standard modes. Overall, I find it to be an improvement over the previous version. I enjoyed using it quite a bit, actually, though I did have a little bit of a problem with aiming it once or twice resulting in me not going where I wanted to go.

I feel like it's definitely getting there though. It's somewhat got the feel of the thok while also having more or less the maneuverability of the double jump.