This is a Melee mod. If anyone knows how many mods are in line in front of me, please, don't post about it, because this mod is special: it's special because I’m making as much of the design process as possible public. Specifically, I want to do this in a semi-democratic fashion, where by "this" I mean creating the list of character changes.
But in order to help the process reach a reasonable conclusion, I want to set the foundational philosophy, and the goal is to -- only being a little arrogant here -- create more actual high-level gameplay than anyone before us. See? Only a little. The name is “Sanded Melee”, from “sanding”, using sandpaper to remove imperfections in a surface. The fundamental question that I start with is: who in hell is going to play this mod, and why? "Who" is Melee players, and "why" is... well, to beat each other up. But more importantly, "why" is: occasionally. Not always. Most of the time, they are still Melee players. Most people who have played SD Remix are probably people who play Melee most of the time. Only a handful become devotees.
In order to do better than all of the previous mods, I don't need to be smarter than the people who made them, and I'm probably not. The advantage that I have over them is that I'm doing this in the future, the same way that every modern scientist knows more about the world than Einstein did. I want to draw on the experience we had with the previous mods, and the most common complaint from Melee players about SD Remix and Project M sounds something like this:
This is commonly said in so many words by luminaries like Mew2King and other people who know way more about Smash than I probably ever will. Unfortunately, it's absolutely impossible to test this theory. Nobody can magically train people to understand the ins and outs of a character in a mod the way that Lucky understands Fox without a lot of very talented people having to spend thousands of hours in front of a screen, grinding reads and reactions, and that takes a looooong time, even longer than writing assembly.
The closest that we can come is taking the skill that people already have with Melee characters, and using as much of it as possible to play our new modded characters. Only then can we begin to approach true balance. One of the first goals of Sanded Melee is going to be skill transferability.
Skill transferability will in many cases consist of the muscle memory that people have with their characters' moves and movement, and so in order to achieve that transferability, we want to keep as many moves the same as we reasonably can. Noted exceptions are moves which are easy to use the same way after being changed: for instance, removing self-damage from Pichu's moves allows them to be used in much the same way as before, although now Pichu can throw out more down smashes. Another exception is when you make something easier which people already did when it was difficult: below, we'll take a look at making it easier for Link to ledge-cancel his aerials. This is already an important part of the Link meta, and all of the old tech will still work.
The next thing we want to focus on is a complaint that many people outside of Melee have with Melee itself:
The secondary focus of our efforts should be to soften the worst matchups. In general, I've mostly considered this in the context of the majority of other Melee mods, which consist primarily of buffs to low tier characters. Most of the discussion and disagreement, to the extent that there is any, concerns a one-dimensional concept of "how good" the low tier characters should become. Should every character be as good as Captain Falcon? How about Mario? Why not Pikachu or Ganondorf or, **** it, Fox?
Instead, I'd like to take a two point approach:
* reduce aspects considered "depressing". Here "depressing" means things like "getting back aired over and over":
https://www.youtube.com/watch?v=-zCX6NP8eOI#t=0m27s
Depressing is bad because, unsurprisingly, it causes people to quit. Such people can include me, or Sethlon. Sethlon is probably more important than me.
* focus particularly on giving mid-tiers the tools to handle their worst matchups.
Mid-tiers are overall characters who we know a decent bit about because some people have mained them and done okay at tournaments, or even gotten onto a regional PR or two. Almost every mid-tier is understood kinesthetically by a top player as a secondary or former main: Luigi (Plup), Yoshi (Leffen), Dr. Mario (Shroomed), Mario (Mango, PPMD), Ganon (n0ne), Young Link (Armada, Axe), DK (Westballz), Mewtwo (Leffen).
A major dilemma is whether to buff the B tier (Peach, ICs, Falcon) and C tier (Pikachu, Samus). These characters could be solo-mained at a high level but for one (B) or a few (C) matchups, respectively. It's tempting to try to fix Peach-Puff, ICs-Peach, Falcon-Falco, but players who already play those characters may not want to use a modded version for fear of messing up their Melee skill.
In my initial list of proposed character modifications, I'm going to buff everyone below Puff. In the final list, they might not get changed. In fact, all of my actual suggestions could be thrown out. Who knows?
But we're going to need a different philosophy when looking at the low tiers, the truly bad characters, who cannot be saved by merely helping them monkeywrench their way through a wall of Foxy back-airs. In fact, my version of the changelist will be characterized by me forfeiting on fixing Ness, Zelda, Bowser and Kirby. These poor broken toys cannot be shaped by my creativity into reasonable Smash characters. For Link, Roy, Mewtwo and Pichu, there are some flaws that I consider obvious enough to fix directly, while reasonably hoping not to create a "monster" like Pichu from early versions of SD Remix. Or Game and Watch. You know.
The last directive I'm working with is a slightly more specific attitude towards buffing moves. I apologize to whoever said this for the disdain I'm about to show for the phrase "no bad moves". There will always be bad moves. Marth's fsmash is bad if you use it wrong. Furthermore, if you try to make sure a character has no bad moves, you're vastly more likely to give them too many good moves, resulting in a character who is not only overpowered but very hard to master, meaning it'll take a long time in order for people to realize how overpowered they really are. Mew2King has said as much in regard to PM Link, as I recall. That's a metagame disaster waiting to happen, not to mention eternally frustrating for anyone who's trying to stick it out with Fox amidst the jank stampede.
Actually I want to leave most of the bad moves alone. What I want to do is to give characters some new good moves and mechanics, with a focus on use-cases, e.g. how is this (modified) move going to help Young Link's godawful out-of-shield game or his inability to edgeguard. We want the new moves to be easy to learn, so we want to have as few new moves and mechanics as possible. I'd also prefer to depend on fundamentals, like walljumps and ledge-cancels, rather than adding new superpowered recoveries and f-air finishers.
But wait, there's more. There are also some things that I do not think need to be goals of a mod:
* Canonicality. The characters' new mechanics do not need to be canonical. It's cool if they are, but gameplay should be central, and poetic license is a good thing.
* Preserving the "play style" of low tiers. The play style of low tiers is to lose in predictable and not-fun ways. We want to destroy the playstyle of low tiers completely, while drawing on the micro-situational knowledge of the people who play them. Link, for example, will probably get a significantly buffed defense. That's okay!
* Pointless changes to moves that don’t actually make them any better. I’m not going to increase the damage on Pikachu’s Thunder from 10 to 12. I mean, like, why? Just why?
* I’m not making a goddamn YouTube video. It’s insanely annoying to quote and analyze and critique a video. A video implies a finished product, which is absurd because I’m a drunk in his apartment, not a game design team with degrees and ****. I expect you to tear me apart, and in order for that to be possible you’re going to need the copy and paste functions on your keyboard.
So our final set of goals is:
* skill transferability, i.e. Melee ability leading to Sanded Melee ability
* focus on bad matchups/dynamics rather than how "good" a character is
* a short list of use-targeted changes to moves and mechanics
Never mind the ********, here's the sex pistols
I wouldn't be posting this thread if I didn't have my own changelist burning a hole in my metaphorical pocket. There's probably one or two people in the room at every tournament who have tried to come up with ideas for a mod; it's not a new concept, after all. If I'm lucky and not just deluding myself, this list will also be a tutorial example of how to implement the above goals in terms of actual changes to Melee characters.
I'm probably deluding myself.
6. Peach
The only thing I care about is Jigglypuff. Peach can fight Jiggs, but when percents get high she just can't pop this pink balloon. The hardest thing to do is change something about Peach in a way that makes her better against Jigglypuff, without making her even more absurdly good against everyone below the C tier.
The best thing I can do is give the wink-eye turnip (6.9% odds) upwards knockback with high knockback growth. Other characters face being popped up and combo'd, but the same turnip hit would in vanilla Melee lead to an edgeguard situation, which for every character except Jigglypuff is worse. Giving Peach the opportunity for a kill off of turnip farming evens up the odds considerably.
The idea here is: it's okay that she still loses to Puff as long as it's winnable. The rest of the list will follow this general idea.
7. Ice Climbers
As we all know, Peach kills the crap out of Nana, but as Chudat has shown the world, Sopo can survive for a surprisingly long time even against a determined edgeguarding Peach. I think the trick is to give Nana a significantly improved AI only when she's offstage, helping her take full advantage of ICs side-B and airdodge recovery options. Other characters, who rely on getting Nana to higher % in order to kill her, will be less affected by this change.
A more targeted change is giving Nana a certain chance of catching an item thrown at her in certain situations. Because the Links are getting buffed, and they're trouble for ICs as well, this might actually be a good idea.
8. Captain Falcon
His recovery sucks, but he gets a lot more out of it by DIing up, which works against every character that isn't a blue bird with a laser that can easily zap all of that height away. There are a couple of options: we can let him recover below the stage by canceling side-B into walljump, or we can add a little projectile armor to down-B or side-B to allow him to recover through some lasers. I'm actually not sure what exactly to go with, but unfortunately you probably can't help the Captain respond to Falco's up-tilt without also making him way better against the characters he already wrecks.
The other option would be, instead of messing with his recovery, end his down-throw animation one or two frames earlier to make it easier to reaction tech-chase spacies. I don't think you want to do both. But do people really enjoy long RTC strings?
9. Pikachu
"Can't kill floaties" is the simplest description of Pikachu's trouble with high level matchups. He does well against floaty characters at mid-level, but once they get used to his very few up-smash setups, you basically can't hit them with it, and there are no real alternatives. Puff lives to 160% until you can kill her with nair and for Peach it's even longer. There were two changes I thought of that might give him a better endgame: either allowing him to jump out of dash attack after about frame 35 or so (so a lot earlier than the animation ends), leading into possibly jump-cancel upsmashes or a well-placed nair, or reducing the land-lag of fair so that Pikachu can follow up with upsmash or uptilt.
I think I prefer the second option, reducing fair's land-lag to around 9 frames with an L-cancel. It also makes his neutral less one-dimensional without making it too good, and Marth and Falcon in particular can easily beat his fair, so it won't affect them too much.
10. Samus
The most depressing thing for a Samus player is when you shoot a charge shot at Sheik, she replies with uncharged needles, and her needles not only block the charge shot but go through it and hit you. Charge shot should beat needles, but only when it’s fully charged.
She’s also very easy to shield poke, because even though she has a large shield she also has a large… suit of armor, which is a hurtbox because Melee is like that. So… we give her an even larger shield in order to deal with that. She already does okay against all of the characters with ultra-safe shield pressure (Fox/Falco/Peach) so this way we shouldn’t affect those matchups too much.
11. Yoshi
His worst matchup is Peach, followed by all three fast-fallers. They are all particularly vulnerable to one of his weaker moves, though: Egg Lay. Peach’s float naturally has her moving horizontally in the air, which is the best-case scenario for Egg Lay, and the faster a character falls, the bigger threat Egg Lay is when Yoshi’s near the ledge. Also, the high damage on spacie back-air and Falcon everything makes it easier to send Yoshi into a lethal tumble animation, and Egg Lay is the fastest way to stop tumble (seriously).
So we just want to make Egg Lay better. I’d shorten the animation by a third, from 39 frames to 26, and move the hitbox up to frame 14 or 15, ending on 17 or 18. It’s still a very slow move, but it does a much better job of stabilizing Yoshi in the air and letting him air-dodge. It might also make offstage edgeguards possible with Egg Lay, which is pretty cool IMO. And last, we let Yoshi swallow projectiles (if he catches them with his tongue/mouth), because a tumbling Yoshi has generally been a sitting duck for turnips or charge shots (and probably Link’s bombs as well), and killing him clearly shouldn’t be that easy.
A different possibility I had considered was giving Yoshi a zair, so he could lick the ledge. This works well for Samus against spacie edgeguarding, for one thing. Modifying parry to work better against multi-hit moves (maybe extending parry by 1-2 frames each time the shield gets hit?) is another way of targeting the Fox and Peach matchups. I would only do one of these, or maybe zair + one of the other two.
12. Dr. Mario
In order to deal with a wall of back-airs, we want to give Dr. Mario a move with better range; in order to not make him too good, it shouldn’t be grab which is what some people have suggested in the past. The other thing that sucks about being the Doctor against a spacie is platform camping, and we can rectify all of that by extending his up-air hitbox in the top-front corner. We want it to either beat, trade with, or lose to bair depending on spacing and timing; a move that always beats Fox’s bair is stupid strong, but because the hitbox lasts much shorter than bair we can safely make it a lot bigger.
The other thing we want to do is make his recovery a little less predictable, and the easiest way to do that is to make cape a little faster. Contrary to popular belief, Doc’s edgeguard on spacies is pretty tough on reaction because cape is a frame 14 move and Firefox starts moving at frame 45 so assuming 15 frames of reaction time you have a quarter of a second to get to where you need to be and often you won’t. We can safely move the hitbox up to frame 10-ish and shorten the animation to 28 frames or so. And of course we’ll let him use it multiple times to recover.
13. Luigi
Whenever I start to think of adding buffs to a character, one of the first things that always comes to mind is “give them a walljump”. But Luigi really *does* deserve a walljump. Sheik, Falco and Marth can easily wall him off the stage and exploit his inability to get back on. In order to make below-stage recoveries a thing, we also want him to be able to cancel side-b into a walljump, probably whenever the horizontal momentum ends. Walljumps are less important when recovering against characters who edgeguard offstage (Fox + floaties) which is fine because Luigi does okay against them already. We’ll also give him Mario’s up-B-cancel-walljump, which actually might be an automatic consequence of him having a walljump, and which seems like a good compromise between buffing his up-B and not.
Or, more simply, because walljumps are good when you can’t grab ledge, and because being prevented from grabbing the ledge is Luigi’s biggest weakness, give him a walljump.
The other thing about Luigi is the “sitting duck” effect where Luigi in the air has actually no movement options that aren’t side-b. Since increasing his airspeed would be too much, instead I want to let him move downwards, increasing his fall speed to 1.83 (even with Ness). He still has terrible airspeed, but being able to fastfall lets him avoid Falco lasers a little better.
14. Ganondorf
Bad matchups here are Sheik, Fox, and Falcon; uniquely, Ganon actually does very well against almost the whole rest of the cast, and PPMD has gone on record saying that Ganon-Falco is even although I think 45-55 is probably more accurate. So here we really don’t want to create a monster. Sheik and Falcon are really good at tech-chasing, so the first thing to do is give Ganon longer tech-rolls, which will affect those matchups the most.
The next thing is… well, it shouldn’t be too easy for Ganon to get a grab, but Sheik shouldn’t become immune to grabs by merely crouching. So first we extend his grab hitbox downwards a little so that he can grab a crouching Sheik with the right spacing. Then we’ll also increase the base knockback on dthrow a little so that the Fox chaingrab starts earlier (aiming for about 30%), which has a nice side effect where Ganon’s chaingrabs on low tiers become a little easier to escape at high %s.
15. Mario
Of course I’m giving him the faster, multi-use cape. Did you even need to ask? It’s pretty much the same rationale as for Dr. Mario, too.
But he’s not getting the up-air. Mario has to worry about Puff and Marth even more than Doc, and is less concerned about spacies. Instead, noticing that his worst matchups are Jigglypuff, Marth and Falcon, the logical thing to do is… give him a better projectile. In order to keep things interesting Fireball will only get better in the air, so we’re splitting it into an aerial animation and grounded, similar to spacie lasers; similar to spacie lasers, aerial Fireball will autocancel on landing. In fact, that’s enough of a buff right there.
The other thing every smash mod does when they buff Mario is making fair stronger. But partially thanks to his fist of mass destruction, Mario is one of the most hated characters in PM, generally referred to as “cancer”.
SD Remix considered making it a spike, but ended up just making it faster, because a spike was too strong. I don’t agree with this. Let’s consider where fair happens: usually at the end of a combo. It’s relatively safe to throw out offstage, and particularly it’s much less of a risk than Falco or Marth’s dair. The meteor smash will kill anyone who isn’t used to it, but if you know the matchup, fair won’t kill you even at 150%. But this is exactly why people complain about Melee low tiers: they get kills because of jank, and now it’s even easier for Mario to jank low-level players right into the blast zone. Meanwhile, once you learn to abuse Mario, it’s useless, so ultimately the better fair is just more annoying.
In order to analyze fair I want to split the possible uses of fair in half. In one case, Mario has stage control; in the other case, his opponent has stage control. As it stands, when Mario has stage control, he can combo into a fair relatively easy, and it’s a likely kill on an unpracticed opponent. When he doesn’t have stage control, the opponent gets hit towards center stage and can easily tech away. This sucks particularly because Mario has a very hard time being cornered.
Ideally I want fair to be better when Mario doesn’t have stage control and worse when he does. I also don’t want it to be a meteor kill because that’s hella jank and annoying. So our change to fair is to make it a spike, but the knockback goes behind Mario, which means that if you’re comboing your opponent off of the stage, you have to get them pretty far to kill with fair or they can DI back on. It also makes Mario more dangerous when cornered because a surprise fair can turn the situation completely around. A newb doing survival DI against an aggressive plumber won’t get janked, but a Falco abusing his bair on a cornered Mario risks dying. That’s what I call “justice”.
To accomplish this we change fair’s knockback angle from 280 to 240. So here I have cured cancer.
16. Young Link
If you’ve watched Laijin play against spacies and enjoyed it, maybe Chris Hansen should check if there are any other forms of child abuse you enjoy. YL shields = he loses a stock. He also has the most useless moonwalk in the game, gets chaingrabbed by Sheik forever, and has absolutely no way of hitting Marth (bomb pulls can be punished like turnips, only harder). His matchups against other characters are actually decent, but every high tier has their own special way of wrecking this fool.
In terms of his OoS game his best options are nair and upair, but I don’t want to make his nair even faster because it’s already too good. This applies to his other aerials as well, which are generally his best moves. So the only other option is to buff up-smash or up-B, and for reasons of hype and coolness, plus versatility, we’re just going to make a slight change to his aerial up-B: a hitbox active on frames 1-2 which corresponds with the lens flare that appears next to his sword when he draws it. This still requires some tech to use, of course, but he gets a frame 5 hitbox out of shield if you do it right. This fast hitbox behind YL also gives him a way out of downthrow to downthrow to downthrow to downthrow to downthrow to downthrow to downthrow to fair. The appropriate knockback direction, of course, is straight up, because down would be too good and anything else would let him get punished for hitting, which is never cool. It’s still a super high commitment move, of course.
He also has nearly the worst rolls in the game, as bad as G&W and Bowser. But G&W and Bowser are at least floaty and heavy respectively, whereas YL is really easy to combo, so it might be worse for him. Most other characters have a 31-frame roll; Yoshi has 33, but the Links have 37. YL’s roll is even shorter than that of most other characters. So when I suggest giving him the same frame data as everyone else, it’s not like I’m giving him a great roll. I’m just giving him a fighting chance. YL deserves a 31-frame roll. Everyone is already used to punishing 31-frame rolls anyway; most YL mains in Melee don’t have to deal with people taking advantage of YL’s roll because he’s so unknown.
The other thing is: he can’t close stocks, and some high-priority moves just wreck him. Falcon is a bad matchup, too. But if we fixed every situation like this by giving the character a huge hitbox, we just end up with a new set of too-good top tiers. YL’s moonwalk hints strongly that he deserves a better bair, but how can we fix all of these problems without making something too good?
Instead of a bigger hitbox, the modification to YL bair that I like is giving him invulnerability (but still grabbable) in between the two kicks, so frame 10-17. Now, rather than being Swiss cheese, YL bair is pretty safe, and capable of trading with almost anything, although it beats almost nothing. It’s particularly useful against characters whose hitboxes don’t stay out for very long, i.e. Marth, who otherwise hard-counters YL by out-ranging him in every situation.
17. Donkey Kong
DK’s worst matchups are the characters who can abuse his recovery, particularly by way of projectiles. In PM this was fixed by putting armor on his recovery, which made it too safe. Instead I want to target projectile gimping specifically, by putting a shine-like projectile-reflector hitbox on his up-B. The other modification that seems to fit here is allowing DK to press B to cancel the up-B animation into a special fall (possibly with delay), which makes it harder for floaty characters to abuse its predictable trajectory.
The other particular matchup which causes a particular pain point is Jigglypuff, who has an easy time landing Rest. My favorite approach here is adding a lot of upwards knockback to charged f-smash, which generally should only ever land on Jigglypuff anyway. If Rest punish kills at 30% on Dreamland I think that’s good.
Also, because it’s too easy to get backthrown off of the ledge, it seems natural that the monkey should be able to jump off of the wall. Unless this would make grab-and-walk-off too easy for DK to survive… but I don’t think it will.
————————— low tiers%%%
OK, remember all that stuff about not making disorienting changes? I’m not following that as much anymore. The remaining characters are so bad that nobody can do anything with them, so they have to change significantly at least somehow.
18. Link
The first step here is giving him the up-B hitbox that we gave to Young Link. It’s still good for escaping combos, but it’s weaker in Link’s hands because of his horrible jumpsquat, second-worst in the game after Bowser.
Link is a slow joke of a character. He can’t approach, he can’t punish approaches, and he’s combo food. He also can’t roll. His aerials are good, but his jumpsquat and landlag are terrible. But he has a fair to good punish game, and because of that giving him a strong aggressive neutral could create a monster.
In order to achieve skill transferability to some extent while still giving Link the revolutionary changes he deserves, we want to take advantage of what Link players already do, but make it better. Much better. Because we’re looking to help him take fewer hits, and a big part of not getting hit is moving away from your opponent, we can cut his traction way down, from 1 all the way to 0.45, just above Mewtwo. Wavedashing is now his fastest movement option, and he can slide away from ground approaches while cc’ing, but in most respects he’s still a slow turd. His wavedash isn’t broken-good because, remember, he still has the second-worst jumpsquat in the game. Plus, Link mains often rely on ledge-canceling land-lag in order to make aerial strings work, and now that got significantly easier (but it’s still hard!).
Link, like Young Link, has terrible roll frame data. But unlike Young Link, he doesn’t have a terrible roll length. So here we’ll put him even with Yoshi, since we already buffed his traction. I guess. 33 frames.
One last cosmetic change I like that isn’t terribly influential but is kinda fun is changing out his Hylian Shield for a Mirror Shield that reflects projectiles, but as before, only while standing or walking and only covering part of his body. Since he has particular trouble with spacies and Sheik, who are affected by this more than other characters, I think it makes sense.
19. Mewtwo
He just dies so easily. If you’ve ever played event 51, and you have, you know that Ganondorf was the Mikasa to Giga Bowser’s Eren, and Mewtwo was that kid Armin whose primary talent is dying. Most of the time you didn’t even remember killing Mewtwo, you were just trying to grab Poke Balls without being hit by Wizard Foot. Mewtwo is lighter than Popo. That’s absurd.
And of course he would die, because this low-weight floaty has one of the largest hurtbox profiles in the game thanks to his stupid tail. The first change we’re going to make is to delete his top tail hurtbox, which allows him to do at least some things safely (deleting both gives him huge stupid disjoints). The other change is to increase his weight to be equal to Mario, or maybe Link, because what kind of supervillain dies at 70%?
Mewtwo tends to have a worse time with fast-fallers than floaties; they can easily punish his airdodges and they’re good at killing him off the top. He usually depends on Confusion to combo them; it’s one of his best responses to an oncoming SHFFL, but it’s so punishable that it’s nearly useless at a high level, so we’ll reduce the move’s length from 55 frames to 41 — not safe by any means, but pretty ordinary for how it works. We’ll also move the start of the reflector up from frame 12 to frame 6 so it’s not utter trash, since Mewtwo doesn’t fare well against Young Link or Peach either.
“Hey, SDRemix did all that and more” — SDRemix Mewtwo is overpowered.
20. Roy
Roy is a clone of Marth, but he uses a pool noodle instead of a sword. He gets run over in neutral because his long range is easily ignored by top tiers who simply “bulldoze” him by using high-priority hitboxes to which he has no real answer. However, he does hit pretty hard when he gets one, and his combo game depends in some ways on his ultra-weak sourspots, giving him a unique playstyle that I wouldn’t want to destroy. In fact, one of the more famous modifications to Roy strengthened all his sourspots, provoking complaints that his punish game had gotten worse and the new character wasn’t really Roy.
Instead I want to simply give Roy an answer to an approaching Fox nair that he knows is coming, other than shieldgrab which sucks. So I’ll buff Roy’s nair specifically, increasing the sourspot knockback slightly and the damage to 9%, while increasing the mid-spot to 12% and the sweetspot to 15%. Nair is still a poor approach, but it works well in retreat, which is what we were looking for.
But he’s too slow to use his toolkit properly: Roy needs to get closer to his opponents than Marth does to get a punish string going out of neutral, yet once he’s got the opponent in the air he is free to swipe repeatedly with his sourspots. SD Remix buffs his airspeed, which helps his punish game, but it’s still too slow to be useful for approaching in neutral — even Falcon (airspeed 1.12) has trouble with that. Instead, we’re going to give Roy what he always deserved: a better dash-dance, buffing his dash speed up to 1.72 from 1.47.
On defense, your typical glass-cannon character tries to tech away, but Roy’s fall speed is too low for that, so we’ll boost it to 2.7, just below Fox. Similarly, we increase his gravity to 1.34. This makes him a true fastfaller, which I’m emphatically in favor of: fastfallers are 3 out of 26 characters in Melee, but roughly half of the playerbase uses one, and that’s because they’re arguably the most fun. And because his dair is now such a risk, we’ll make the sweetspot a spike, but keep the sourspot the same. (I think easy-hit meteors are jank, and try to remove them)
To compensate for his now-terrible recovery, which was already one of the worst in the game, we let neutral-B cancel Roy’s horizontal momentum in the air (when he’s knocked horizontally and bounces), we add a walljump, and increase Blazer’s height gain by about a third or so. Yes, it’s another walljump (our third and last), but put it in context: all three of the other fast-fallers have walljumps. I’m not buffing side-B (looking at you SDRemix), it’s already too good, because it beats projectiles. I want more complex recoveries that reward learning, not free ****. Blazer is now good, but it’s a risk, and you have to trick your opponent a little for it to work. That’s good Melee.
21. Pichu
Anyone reading this from start to finish — all three of you — remembers that I said the low tiers had “obvious, fundamental flaws” that could be easily corrected. In Pichu’s case, everyone already knows what I’m talking about. Except they might be thinking about the wrong thing.
See, Pichu, who is smaller and weaker than Pikachu, should, according to standard fighting game logic, have better defensive options and be able to move faster. The problem is, he has worse defensive options and he moves slower. You probably thought it was the self-damage. Actually, self-damage does matter on a couple moves: thunder jolt, forward air, and agility, so we’ll remove it on those. The other moves that hurt Pichu are used infrequently enough that we can keep it as an Easter egg. Oh, and instead of decreasing self-damage on pummel, we’ll increase enemy-damage to 5. That’s more fun.
So to properly complement his small body and lowest-weight-in-the-game, we’re going to boost Pichu’s dash to 2.1 and his run to 2.25, in both cases between Fox and Falcon, and reduce his gravity to 0.08: Sheik still gets follow-ups off of downthrow at low percent, but at high percent Pichu can DI up. In general, Pichu’s ability to DI out of aerial combos has improved, and he takes no damage from zapping to ledge or surrounding himself with fair.
The last thing we want to look at is down-smash, which is Pikachu’s best grounded move and one of Pichu’s favorites as well, but Pichu suffers a whopping 38 frames of endlag whereas Pikachu only waits for 26. So we’ll want to decrease that to about… let’s go with… 26. After all, down-smash has invulnerability, and even though I’m buffing something that’s already one of his best tools don’t forget that it’s still mediocre because after all this is Pichu we’re talking about. You probably remember that Pichu became infamously too good in SDRemix, so with the changes here going in a similar direction it’s interesting to ask why I don’t think that will happen again; first of all, Pichu in native Melee has the least landlag on wavedash of anyone (Landfallspeciallag = 1 frame), and SDRemix massively buffed his wavedash by reducing his traction from 0.1 to 0.06. I’ve kept Pichu as a high-traction character here because having a long lagless wavedash and invulnerability on downsmash is just immoral, but I do think that Pichu should be able to defend himself against approaches. I also just don’t like the trend of making really good attacks which are “balanced” because “it’s very punishable” if it whiffs, aka half the PM cast. Instead we’ve reduced endlag to make it a safer defensive option, while not allowing it to be used as an approach because it’s a kill move (semispike) after all.
The last bit is to slightly modify the length of the first Agility zap so that it exactly reaches the top platform on Yoshi’s Story starting from a teeter at the ledge. I don’t know if this makes it longer or shorter, but it’s basically Pichu’s version of the Pikachu Battlefield zaps. Yoshi’s is terrible for Pichu so he deserves a little help here. But if we expect it to be dropped from the legal stage list, maybe forget this change.
Oh, did I mention? Pichu already has a walljump. So there’s no way for me to give him one.
22. Mr. Game and Watch
Uh, you should probably give him the ability to L-cancel.
23. Kirby
Probably fix his grabs and get-up attack. Other than that, I’m out of ideas about what to do with the rest of the characters, including these last two.
————————— end %%%%
Anyway I was of course making fun of PM and SDRemix a lot while making many similar decisions to what they did, but it’s only in the spirit of improvement, I find both of those mods to be a lot of fun and if I didn’t think they were good I probably wouldn’t still think modding Melee is a good idea. The main difference you see above is that there are many fewer changes but they tend to be stronger. In SD Remix, Pichu got a buffed wavedash, dashdance, and aerial mobility; here, he only got the second one, but his run speed is now notably higher than Fox (which was intentional). I think this makes it easier to know what to fix if one of the characters turns out to be meta-breaking; there are fewer changes, and their impact is more obvious.
I don’t really expect the above changelist to be the final changelist (assuming this project gets off the ground). It’s just too likely that I did something stupid across 50 or so different major changes to character mechanics. The characters I left out might still be fixable in a way that doesn’t make them really obnoxious, although on that note I do like the idea of increasing Bowser’s weight.
So the changelist is:
Peach:
* Wink-eye turnip gains additional upwards knockback (kills Jigglypuff on the ground at 120 on FD)
Ice Climbers:
* Nana AI uses side-B and airdodges to recover
* Nana AI catches projectiles with probability equal to Popo’s % (or a fixed % if that won’t work lol)
Captain Falcon:
* can walljump out of side-B
Pikachu:
* f-air landlag reduced from 26 to 18
Samus:
* charged shot beats some other projectiles (needles)
* gets a larger shield
Yoshi:
* Egg Lay hits frames 15-18, total 26 frames
Dr. Mario:
* Super Sheet hits on frame 10, ends frame 28, can be used 3 times to slow falling
* Up-air hitbox slightly larger in top-front corner
Luigi:
* Fall speed 1.6 -> 1.83
* Can walljump, including up-B-cancel
Ganondorf:
* tech roll length increased ~15%
* grab range extended downwards slightly
* down-throw base knockback increased slightly (cg easier on fastfallers, harder on others)
Mario:
* Cape hits on frame 10, ends frame 28, can be used 4 times to stop falling
* Aerial fireball animation autocancels when Mario lands
* Fair knockback angle 280 -> 240 (reverse spike)
Young Link:
* Aerial Up-B gets a frame 1 fire hitbox behind YL (where he holds his sword to charge)
* Back-air has invulnerability frame 10-17.
* Roll frames 37 -> 31 (even with most of the cast)
Donkey Kong:
* Up-B reflects projectiles
* Aerial up-B can be cancelled into a special fall
* Can walljump
* Charged f-smash gets additional upwards knockback (kills Puff at 20% on FD)
Mewtwo:
* Weight 85 -> 102
* Outer tail hitbox removed
* Confusion ends frame 41 instead of 55 if it whiffs (old data on hit), reflection starts frame 6
Link:
* Aerial Up-B gets a frame 1 hitbox behind Link
* Roll frames 37 -> 33 (even with Yoshi)
* Traction 0.1 -> 0.045
* Shield reflects projectiles (incomplete cover of Link’s body; does not reflect fully charged Charge Shot)
Roy:
* Dash speed 1.47 -> 1.72 (run speed unchanged)
* Fall speed 2.4 -> 2.7
* Gravity 1.14 -> 1.34
* Flare Blade stops Roy from moving when used in the air
* Nair damage 8/7/6 -> 15/12/9, sourspot knockback increased slightly
* Can walljump
* Blazer goes 30% higher, can grab ledge from behind
Pichu
* Dash speed 1.72 -> 2.1, run speed 1.72 -> 2.25
* Fair, Agility, Thunder Jolt no longer damage Pichu
* Gravity 1.1 -> 0.8
* Pummel damage 3->5
* Down-smash IASA frame 39 instead of 51
* Agility’s first zap is just long enough to slide onto the top platform on Yoshi’s Story by doing an upside-down L-shape from a teeter
With that I open the floor to anyone who wants to complain that this changelist is terrible and I should have done this or that or whatever. If you take away anything from this, I hope it’s the idea that the changelist can be developed from community knowledge and experience and do better than just trying random things and see if it works, since we can get the right ideas faster and with less effort if we know what to look for.
But in order to help the process reach a reasonable conclusion, I want to set the foundational philosophy, and the goal is to -- only being a little arrogant here -- create more actual high-level gameplay than anyone before us. See? Only a little. The name is “Sanded Melee”, from “sanding”, using sandpaper to remove imperfections in a surface. The fundamental question that I start with is: who in hell is going to play this mod, and why? "Who" is Melee players, and "why" is... well, to beat each other up. But more importantly, "why" is: occasionally. Not always. Most of the time, they are still Melee players. Most people who have played SD Remix are probably people who play Melee most of the time. Only a handful become devotees.
In order to do better than all of the previous mods, I don't need to be smarter than the people who made them, and I'm probably not. The advantage that I have over them is that I'm doing this in the future, the same way that every modern scientist knows more about the world than Einstein did. I want to draw on the experience we had with the previous mods, and the most common complaint from Melee players about SD Remix and Project M sounds something like this:
This is commonly said in so many words by luminaries like Mew2King and other people who know way more about Smash than I probably ever will. Unfortunately, it's absolutely impossible to test this theory. Nobody can magically train people to understand the ins and outs of a character in a mod the way that Lucky understands Fox without a lot of very talented people having to spend thousands of hours in front of a screen, grinding reads and reactions, and that takes a looooong time, even longer than writing assembly.
The closest that we can come is taking the skill that people already have with Melee characters, and using as much of it as possible to play our new modded characters. Only then can we begin to approach true balance. One of the first goals of Sanded Melee is going to be skill transferability.
Skill transferability will in many cases consist of the muscle memory that people have with their characters' moves and movement, and so in order to achieve that transferability, we want to keep as many moves the same as we reasonably can. Noted exceptions are moves which are easy to use the same way after being changed: for instance, removing self-damage from Pichu's moves allows them to be used in much the same way as before, although now Pichu can throw out more down smashes. Another exception is when you make something easier which people already did when it was difficult: below, we'll take a look at making it easier for Link to ledge-cancel his aerials. This is already an important part of the Link meta, and all of the old tech will still work.
The next thing we want to focus on is a complaint that many people outside of Melee have with Melee itself:
The secondary focus of our efforts should be to soften the worst matchups. In general, I've mostly considered this in the context of the majority of other Melee mods, which consist primarily of buffs to low tier characters. Most of the discussion and disagreement, to the extent that there is any, concerns a one-dimensional concept of "how good" the low tier characters should become. Should every character be as good as Captain Falcon? How about Mario? Why not Pikachu or Ganondorf or, **** it, Fox?
Instead, I'd like to take a two point approach:
* reduce aspects considered "depressing". Here "depressing" means things like "getting back aired over and over":
https://www.youtube.com/watch?v=-zCX6NP8eOI#t=0m27s
Depressing is bad because, unsurprisingly, it causes people to quit. Such people can include me, or Sethlon. Sethlon is probably more important than me.
* focus particularly on giving mid-tiers the tools to handle their worst matchups.
Mid-tiers are overall characters who we know a decent bit about because some people have mained them and done okay at tournaments, or even gotten onto a regional PR or two. Almost every mid-tier is understood kinesthetically by a top player as a secondary or former main: Luigi (Plup), Yoshi (Leffen), Dr. Mario (Shroomed), Mario (Mango, PPMD), Ganon (n0ne), Young Link (Armada, Axe), DK (Westballz), Mewtwo (Leffen).
A major dilemma is whether to buff the B tier (Peach, ICs, Falcon) and C tier (Pikachu, Samus). These characters could be solo-mained at a high level but for one (B) or a few (C) matchups, respectively. It's tempting to try to fix Peach-Puff, ICs-Peach, Falcon-Falco, but players who already play those characters may not want to use a modded version for fear of messing up their Melee skill.
In my initial list of proposed character modifications, I'm going to buff everyone below Puff. In the final list, they might not get changed. In fact, all of my actual suggestions could be thrown out. Who knows?
But we're going to need a different philosophy when looking at the low tiers, the truly bad characters, who cannot be saved by merely helping them monkeywrench their way through a wall of Foxy back-airs. In fact, my version of the changelist will be characterized by me forfeiting on fixing Ness, Zelda, Bowser and Kirby. These poor broken toys cannot be shaped by my creativity into reasonable Smash characters. For Link, Roy, Mewtwo and Pichu, there are some flaws that I consider obvious enough to fix directly, while reasonably hoping not to create a "monster" like Pichu from early versions of SD Remix. Or Game and Watch. You know.
The last directive I'm working with is a slightly more specific attitude towards buffing moves. I apologize to whoever said this for the disdain I'm about to show for the phrase "no bad moves". There will always be bad moves. Marth's fsmash is bad if you use it wrong. Furthermore, if you try to make sure a character has no bad moves, you're vastly more likely to give them too many good moves, resulting in a character who is not only overpowered but very hard to master, meaning it'll take a long time in order for people to realize how overpowered they really are. Mew2King has said as much in regard to PM Link, as I recall. That's a metagame disaster waiting to happen, not to mention eternally frustrating for anyone who's trying to stick it out with Fox amidst the jank stampede.
Actually I want to leave most of the bad moves alone. What I want to do is to give characters some new good moves and mechanics, with a focus on use-cases, e.g. how is this (modified) move going to help Young Link's godawful out-of-shield game or his inability to edgeguard. We want the new moves to be easy to learn, so we want to have as few new moves and mechanics as possible. I'd also prefer to depend on fundamentals, like walljumps and ledge-cancels, rather than adding new superpowered recoveries and f-air finishers.
But wait, there's more. There are also some things that I do not think need to be goals of a mod:
* Canonicality. The characters' new mechanics do not need to be canonical. It's cool if they are, but gameplay should be central, and poetic license is a good thing.
* Preserving the "play style" of low tiers. The play style of low tiers is to lose in predictable and not-fun ways. We want to destroy the playstyle of low tiers completely, while drawing on the micro-situational knowledge of the people who play them. Link, for example, will probably get a significantly buffed defense. That's okay!
* Pointless changes to moves that don’t actually make them any better. I’m not going to increase the damage on Pikachu’s Thunder from 10 to 12. I mean, like, why? Just why?
* I’m not making a goddamn YouTube video. It’s insanely annoying to quote and analyze and critique a video. A video implies a finished product, which is absurd because I’m a drunk in his apartment, not a game design team with degrees and ****. I expect you to tear me apart, and in order for that to be possible you’re going to need the copy and paste functions on your keyboard.
So our final set of goals is:
* skill transferability, i.e. Melee ability leading to Sanded Melee ability
* focus on bad matchups/dynamics rather than how "good" a character is
* a short list of use-targeted changes to moves and mechanics
Never mind the ********, here's the sex pistols
I wouldn't be posting this thread if I didn't have my own changelist burning a hole in my metaphorical pocket. There's probably one or two people in the room at every tournament who have tried to come up with ideas for a mod; it's not a new concept, after all. If I'm lucky and not just deluding myself, this list will also be a tutorial example of how to implement the above goals in terms of actual changes to Melee characters.
I'm probably deluding myself.
6. Peach
The only thing I care about is Jigglypuff. Peach can fight Jiggs, but when percents get high she just can't pop this pink balloon. The hardest thing to do is change something about Peach in a way that makes her better against Jigglypuff, without making her even more absurdly good against everyone below the C tier.
The best thing I can do is give the wink-eye turnip (6.9% odds) upwards knockback with high knockback growth. Other characters face being popped up and combo'd, but the same turnip hit would in vanilla Melee lead to an edgeguard situation, which for every character except Jigglypuff is worse. Giving Peach the opportunity for a kill off of turnip farming evens up the odds considerably.
The idea here is: it's okay that she still loses to Puff as long as it's winnable. The rest of the list will follow this general idea.
7. Ice Climbers
As we all know, Peach kills the crap out of Nana, but as Chudat has shown the world, Sopo can survive for a surprisingly long time even against a determined edgeguarding Peach. I think the trick is to give Nana a significantly improved AI only when she's offstage, helping her take full advantage of ICs side-B and airdodge recovery options. Other characters, who rely on getting Nana to higher % in order to kill her, will be less affected by this change.
A more targeted change is giving Nana a certain chance of catching an item thrown at her in certain situations. Because the Links are getting buffed, and they're trouble for ICs as well, this might actually be a good idea.
8. Captain Falcon
His recovery sucks, but he gets a lot more out of it by DIing up, which works against every character that isn't a blue bird with a laser that can easily zap all of that height away. There are a couple of options: we can let him recover below the stage by canceling side-B into walljump, or we can add a little projectile armor to down-B or side-B to allow him to recover through some lasers. I'm actually not sure what exactly to go with, but unfortunately you probably can't help the Captain respond to Falco's up-tilt without also making him way better against the characters he already wrecks.
The other option would be, instead of messing with his recovery, end his down-throw animation one or two frames earlier to make it easier to reaction tech-chase spacies. I don't think you want to do both. But do people really enjoy long RTC strings?
9. Pikachu
"Can't kill floaties" is the simplest description of Pikachu's trouble with high level matchups. He does well against floaty characters at mid-level, but once they get used to his very few up-smash setups, you basically can't hit them with it, and there are no real alternatives. Puff lives to 160% until you can kill her with nair and for Peach it's even longer. There were two changes I thought of that might give him a better endgame: either allowing him to jump out of dash attack after about frame 35 or so (so a lot earlier than the animation ends), leading into possibly jump-cancel upsmashes or a well-placed nair, or reducing the land-lag of fair so that Pikachu can follow up with upsmash or uptilt.
I think I prefer the second option, reducing fair's land-lag to around 9 frames with an L-cancel. It also makes his neutral less one-dimensional without making it too good, and Marth and Falcon in particular can easily beat his fair, so it won't affect them too much.
10. Samus
The most depressing thing for a Samus player is when you shoot a charge shot at Sheik, she replies with uncharged needles, and her needles not only block the charge shot but go through it and hit you. Charge shot should beat needles, but only when it’s fully charged.
She’s also very easy to shield poke, because even though she has a large shield she also has a large… suit of armor, which is a hurtbox because Melee is like that. So… we give her an even larger shield in order to deal with that. She already does okay against all of the characters with ultra-safe shield pressure (Fox/Falco/Peach) so this way we shouldn’t affect those matchups too much.
11. Yoshi
His worst matchup is Peach, followed by all three fast-fallers. They are all particularly vulnerable to one of his weaker moves, though: Egg Lay. Peach’s float naturally has her moving horizontally in the air, which is the best-case scenario for Egg Lay, and the faster a character falls, the bigger threat Egg Lay is when Yoshi’s near the ledge. Also, the high damage on spacie back-air and Falcon everything makes it easier to send Yoshi into a lethal tumble animation, and Egg Lay is the fastest way to stop tumble (seriously).
So we just want to make Egg Lay better. I’d shorten the animation by a third, from 39 frames to 26, and move the hitbox up to frame 14 or 15, ending on 17 or 18. It’s still a very slow move, but it does a much better job of stabilizing Yoshi in the air and letting him air-dodge. It might also make offstage edgeguards possible with Egg Lay, which is pretty cool IMO. And last, we let Yoshi swallow projectiles (if he catches them with his tongue/mouth), because a tumbling Yoshi has generally been a sitting duck for turnips or charge shots (and probably Link’s bombs as well), and killing him clearly shouldn’t be that easy.
A different possibility I had considered was giving Yoshi a zair, so he could lick the ledge. This works well for Samus against spacie edgeguarding, for one thing. Modifying parry to work better against multi-hit moves (maybe extending parry by 1-2 frames each time the shield gets hit?) is another way of targeting the Fox and Peach matchups. I would only do one of these, or maybe zair + one of the other two.
12. Dr. Mario
In order to deal with a wall of back-airs, we want to give Dr. Mario a move with better range; in order to not make him too good, it shouldn’t be grab which is what some people have suggested in the past. The other thing that sucks about being the Doctor against a spacie is platform camping, and we can rectify all of that by extending his up-air hitbox in the top-front corner. We want it to either beat, trade with, or lose to bair depending on spacing and timing; a move that always beats Fox’s bair is stupid strong, but because the hitbox lasts much shorter than bair we can safely make it a lot bigger.
The other thing we want to do is make his recovery a little less predictable, and the easiest way to do that is to make cape a little faster. Contrary to popular belief, Doc’s edgeguard on spacies is pretty tough on reaction because cape is a frame 14 move and Firefox starts moving at frame 45 so assuming 15 frames of reaction time you have a quarter of a second to get to where you need to be and often you won’t. We can safely move the hitbox up to frame 10-ish and shorten the animation to 28 frames or so. And of course we’ll let him use it multiple times to recover.
13. Luigi
Whenever I start to think of adding buffs to a character, one of the first things that always comes to mind is “give them a walljump”. But Luigi really *does* deserve a walljump. Sheik, Falco and Marth can easily wall him off the stage and exploit his inability to get back on. In order to make below-stage recoveries a thing, we also want him to be able to cancel side-b into a walljump, probably whenever the horizontal momentum ends. Walljumps are less important when recovering against characters who edgeguard offstage (Fox + floaties) which is fine because Luigi does okay against them already. We’ll also give him Mario’s up-B-cancel-walljump, which actually might be an automatic consequence of him having a walljump, and which seems like a good compromise between buffing his up-B and not.
Or, more simply, because walljumps are good when you can’t grab ledge, and because being prevented from grabbing the ledge is Luigi’s biggest weakness, give him a walljump.
The other thing about Luigi is the “sitting duck” effect where Luigi in the air has actually no movement options that aren’t side-b. Since increasing his airspeed would be too much, instead I want to let him move downwards, increasing his fall speed to 1.83 (even with Ness). He still has terrible airspeed, but being able to fastfall lets him avoid Falco lasers a little better.
14. Ganondorf
Bad matchups here are Sheik, Fox, and Falcon; uniquely, Ganon actually does very well against almost the whole rest of the cast, and PPMD has gone on record saying that Ganon-Falco is even although I think 45-55 is probably more accurate. So here we really don’t want to create a monster. Sheik and Falcon are really good at tech-chasing, so the first thing to do is give Ganon longer tech-rolls, which will affect those matchups the most.
The next thing is… well, it shouldn’t be too easy for Ganon to get a grab, but Sheik shouldn’t become immune to grabs by merely crouching. So first we extend his grab hitbox downwards a little so that he can grab a crouching Sheik with the right spacing. Then we’ll also increase the base knockback on dthrow a little so that the Fox chaingrab starts earlier (aiming for about 30%), which has a nice side effect where Ganon’s chaingrabs on low tiers become a little easier to escape at high %s.
15. Mario
Of course I’m giving him the faster, multi-use cape. Did you even need to ask? It’s pretty much the same rationale as for Dr. Mario, too.
But he’s not getting the up-air. Mario has to worry about Puff and Marth even more than Doc, and is less concerned about spacies. Instead, noticing that his worst matchups are Jigglypuff, Marth and Falcon, the logical thing to do is… give him a better projectile. In order to keep things interesting Fireball will only get better in the air, so we’re splitting it into an aerial animation and grounded, similar to spacie lasers; similar to spacie lasers, aerial Fireball will autocancel on landing. In fact, that’s enough of a buff right there.
The other thing every smash mod does when they buff Mario is making fair stronger. But partially thanks to his fist of mass destruction, Mario is one of the most hated characters in PM, generally referred to as “cancer”.
SD Remix considered making it a spike, but ended up just making it faster, because a spike was too strong. I don’t agree with this. Let’s consider where fair happens: usually at the end of a combo. It’s relatively safe to throw out offstage, and particularly it’s much less of a risk than Falco or Marth’s dair. The meteor smash will kill anyone who isn’t used to it, but if you know the matchup, fair won’t kill you even at 150%. But this is exactly why people complain about Melee low tiers: they get kills because of jank, and now it’s even easier for Mario to jank low-level players right into the blast zone. Meanwhile, once you learn to abuse Mario, it’s useless, so ultimately the better fair is just more annoying.
In order to analyze fair I want to split the possible uses of fair in half. In one case, Mario has stage control; in the other case, his opponent has stage control. As it stands, when Mario has stage control, he can combo into a fair relatively easy, and it’s a likely kill on an unpracticed opponent. When he doesn’t have stage control, the opponent gets hit towards center stage and can easily tech away. This sucks particularly because Mario has a very hard time being cornered.
Ideally I want fair to be better when Mario doesn’t have stage control and worse when he does. I also don’t want it to be a meteor kill because that’s hella jank and annoying. So our change to fair is to make it a spike, but the knockback goes behind Mario, which means that if you’re comboing your opponent off of the stage, you have to get them pretty far to kill with fair or they can DI back on. It also makes Mario more dangerous when cornered because a surprise fair can turn the situation completely around. A newb doing survival DI against an aggressive plumber won’t get janked, but a Falco abusing his bair on a cornered Mario risks dying. That’s what I call “justice”.
To accomplish this we change fair’s knockback angle from 280 to 240. So here I have cured cancer.
16. Young Link
If you’ve watched Laijin play against spacies and enjoyed it, maybe Chris Hansen should check if there are any other forms of child abuse you enjoy. YL shields = he loses a stock. He also has the most useless moonwalk in the game, gets chaingrabbed by Sheik forever, and has absolutely no way of hitting Marth (bomb pulls can be punished like turnips, only harder). His matchups against other characters are actually decent, but every high tier has their own special way of wrecking this fool.
In terms of his OoS game his best options are nair and upair, but I don’t want to make his nair even faster because it’s already too good. This applies to his other aerials as well, which are generally his best moves. So the only other option is to buff up-smash or up-B, and for reasons of hype and coolness, plus versatility, we’re just going to make a slight change to his aerial up-B: a hitbox active on frames 1-2 which corresponds with the lens flare that appears next to his sword when he draws it. This still requires some tech to use, of course, but he gets a frame 5 hitbox out of shield if you do it right. This fast hitbox behind YL also gives him a way out of downthrow to downthrow to downthrow to downthrow to downthrow to downthrow to downthrow to fair. The appropriate knockback direction, of course, is straight up, because down would be too good and anything else would let him get punished for hitting, which is never cool. It’s still a super high commitment move, of course.
He also has nearly the worst rolls in the game, as bad as G&W and Bowser. But G&W and Bowser are at least floaty and heavy respectively, whereas YL is really easy to combo, so it might be worse for him. Most other characters have a 31-frame roll; Yoshi has 33, but the Links have 37. YL’s roll is even shorter than that of most other characters. So when I suggest giving him the same frame data as everyone else, it’s not like I’m giving him a great roll. I’m just giving him a fighting chance. YL deserves a 31-frame roll. Everyone is already used to punishing 31-frame rolls anyway; most YL mains in Melee don’t have to deal with people taking advantage of YL’s roll because he’s so unknown.
The other thing is: he can’t close stocks, and some high-priority moves just wreck him. Falcon is a bad matchup, too. But if we fixed every situation like this by giving the character a huge hitbox, we just end up with a new set of too-good top tiers. YL’s moonwalk hints strongly that he deserves a better bair, but how can we fix all of these problems without making something too good?
Instead of a bigger hitbox, the modification to YL bair that I like is giving him invulnerability (but still grabbable) in between the two kicks, so frame 10-17. Now, rather than being Swiss cheese, YL bair is pretty safe, and capable of trading with almost anything, although it beats almost nothing. It’s particularly useful against characters whose hitboxes don’t stay out for very long, i.e. Marth, who otherwise hard-counters YL by out-ranging him in every situation.
17. Donkey Kong
DK’s worst matchups are the characters who can abuse his recovery, particularly by way of projectiles. In PM this was fixed by putting armor on his recovery, which made it too safe. Instead I want to target projectile gimping specifically, by putting a shine-like projectile-reflector hitbox on his up-B. The other modification that seems to fit here is allowing DK to press B to cancel the up-B animation into a special fall (possibly with delay), which makes it harder for floaty characters to abuse its predictable trajectory.
The other particular matchup which causes a particular pain point is Jigglypuff, who has an easy time landing Rest. My favorite approach here is adding a lot of upwards knockback to charged f-smash, which generally should only ever land on Jigglypuff anyway. If Rest punish kills at 30% on Dreamland I think that’s good.
Also, because it’s too easy to get backthrown off of the ledge, it seems natural that the monkey should be able to jump off of the wall. Unless this would make grab-and-walk-off too easy for DK to survive… but I don’t think it will.
————————— low tiers%%%
OK, remember all that stuff about not making disorienting changes? I’m not following that as much anymore. The remaining characters are so bad that nobody can do anything with them, so they have to change significantly at least somehow.
18. Link
The first step here is giving him the up-B hitbox that we gave to Young Link. It’s still good for escaping combos, but it’s weaker in Link’s hands because of his horrible jumpsquat, second-worst in the game after Bowser.
Link is a slow joke of a character. He can’t approach, he can’t punish approaches, and he’s combo food. He also can’t roll. His aerials are good, but his jumpsquat and landlag are terrible. But he has a fair to good punish game, and because of that giving him a strong aggressive neutral could create a monster.
In order to achieve skill transferability to some extent while still giving Link the revolutionary changes he deserves, we want to take advantage of what Link players already do, but make it better. Much better. Because we’re looking to help him take fewer hits, and a big part of not getting hit is moving away from your opponent, we can cut his traction way down, from 1 all the way to 0.45, just above Mewtwo. Wavedashing is now his fastest movement option, and he can slide away from ground approaches while cc’ing, but in most respects he’s still a slow turd. His wavedash isn’t broken-good because, remember, he still has the second-worst jumpsquat in the game. Plus, Link mains often rely on ledge-canceling land-lag in order to make aerial strings work, and now that got significantly easier (but it’s still hard!).
Link, like Young Link, has terrible roll frame data. But unlike Young Link, he doesn’t have a terrible roll length. So here we’ll put him even with Yoshi, since we already buffed his traction. I guess. 33 frames.
One last cosmetic change I like that isn’t terribly influential but is kinda fun is changing out his Hylian Shield for a Mirror Shield that reflects projectiles, but as before, only while standing or walking and only covering part of his body. Since he has particular trouble with spacies and Sheik, who are affected by this more than other characters, I think it makes sense.
19. Mewtwo
He just dies so easily. If you’ve ever played event 51, and you have, you know that Ganondorf was the Mikasa to Giga Bowser’s Eren, and Mewtwo was that kid Armin whose primary talent is dying. Most of the time you didn’t even remember killing Mewtwo, you were just trying to grab Poke Balls without being hit by Wizard Foot. Mewtwo is lighter than Popo. That’s absurd.
And of course he would die, because this low-weight floaty has one of the largest hurtbox profiles in the game thanks to his stupid tail. The first change we’re going to make is to delete his top tail hurtbox, which allows him to do at least some things safely (deleting both gives him huge stupid disjoints). The other change is to increase his weight to be equal to Mario, or maybe Link, because what kind of supervillain dies at 70%?
Mewtwo tends to have a worse time with fast-fallers than floaties; they can easily punish his airdodges and they’re good at killing him off the top. He usually depends on Confusion to combo them; it’s one of his best responses to an oncoming SHFFL, but it’s so punishable that it’s nearly useless at a high level, so we’ll reduce the move’s length from 55 frames to 41 — not safe by any means, but pretty ordinary for how it works. We’ll also move the start of the reflector up from frame 12 to frame 6 so it’s not utter trash, since Mewtwo doesn’t fare well against Young Link or Peach either.
“Hey, SDRemix did all that and more” — SDRemix Mewtwo is overpowered.
20. Roy
Roy is a clone of Marth, but he uses a pool noodle instead of a sword. He gets run over in neutral because his long range is easily ignored by top tiers who simply “bulldoze” him by using high-priority hitboxes to which he has no real answer. However, he does hit pretty hard when he gets one, and his combo game depends in some ways on his ultra-weak sourspots, giving him a unique playstyle that I wouldn’t want to destroy. In fact, one of the more famous modifications to Roy strengthened all his sourspots, provoking complaints that his punish game had gotten worse and the new character wasn’t really Roy.
Instead I want to simply give Roy an answer to an approaching Fox nair that he knows is coming, other than shieldgrab which sucks. So I’ll buff Roy’s nair specifically, increasing the sourspot knockback slightly and the damage to 9%, while increasing the mid-spot to 12% and the sweetspot to 15%. Nair is still a poor approach, but it works well in retreat, which is what we were looking for.
But he’s too slow to use his toolkit properly: Roy needs to get closer to his opponents than Marth does to get a punish string going out of neutral, yet once he’s got the opponent in the air he is free to swipe repeatedly with his sourspots. SD Remix buffs his airspeed, which helps his punish game, but it’s still too slow to be useful for approaching in neutral — even Falcon (airspeed 1.12) has trouble with that. Instead, we’re going to give Roy what he always deserved: a better dash-dance, buffing his dash speed up to 1.72 from 1.47.
On defense, your typical glass-cannon character tries to tech away, but Roy’s fall speed is too low for that, so we’ll boost it to 2.7, just below Fox. Similarly, we increase his gravity to 1.34. This makes him a true fastfaller, which I’m emphatically in favor of: fastfallers are 3 out of 26 characters in Melee, but roughly half of the playerbase uses one, and that’s because they’re arguably the most fun. And because his dair is now such a risk, we’ll make the sweetspot a spike, but keep the sourspot the same. (I think easy-hit meteors are jank, and try to remove them)
To compensate for his now-terrible recovery, which was already one of the worst in the game, we let neutral-B cancel Roy’s horizontal momentum in the air (when he’s knocked horizontally and bounces), we add a walljump, and increase Blazer’s height gain by about a third or so. Yes, it’s another walljump (our third and last), but put it in context: all three of the other fast-fallers have walljumps. I’m not buffing side-B (looking at you SDRemix), it’s already too good, because it beats projectiles. I want more complex recoveries that reward learning, not free ****. Blazer is now good, but it’s a risk, and you have to trick your opponent a little for it to work. That’s good Melee.
21. Pichu
Anyone reading this from start to finish — all three of you — remembers that I said the low tiers had “obvious, fundamental flaws” that could be easily corrected. In Pichu’s case, everyone already knows what I’m talking about. Except they might be thinking about the wrong thing.
See, Pichu, who is smaller and weaker than Pikachu, should, according to standard fighting game logic, have better defensive options and be able to move faster. The problem is, he has worse defensive options and he moves slower. You probably thought it was the self-damage. Actually, self-damage does matter on a couple moves: thunder jolt, forward air, and agility, so we’ll remove it on those. The other moves that hurt Pichu are used infrequently enough that we can keep it as an Easter egg. Oh, and instead of decreasing self-damage on pummel, we’ll increase enemy-damage to 5. That’s more fun.
So to properly complement his small body and lowest-weight-in-the-game, we’re going to boost Pichu’s dash to 2.1 and his run to 2.25, in both cases between Fox and Falcon, and reduce his gravity to 0.08: Sheik still gets follow-ups off of downthrow at low percent, but at high percent Pichu can DI up. In general, Pichu’s ability to DI out of aerial combos has improved, and he takes no damage from zapping to ledge or surrounding himself with fair.
The last thing we want to look at is down-smash, which is Pikachu’s best grounded move and one of Pichu’s favorites as well, but Pichu suffers a whopping 38 frames of endlag whereas Pikachu only waits for 26. So we’ll want to decrease that to about… let’s go with… 26. After all, down-smash has invulnerability, and even though I’m buffing something that’s already one of his best tools don’t forget that it’s still mediocre because after all this is Pichu we’re talking about. You probably remember that Pichu became infamously too good in SDRemix, so with the changes here going in a similar direction it’s interesting to ask why I don’t think that will happen again; first of all, Pichu in native Melee has the least landlag on wavedash of anyone (Landfallspeciallag = 1 frame), and SDRemix massively buffed his wavedash by reducing his traction from 0.1 to 0.06. I’ve kept Pichu as a high-traction character here because having a long lagless wavedash and invulnerability on downsmash is just immoral, but I do think that Pichu should be able to defend himself against approaches. I also just don’t like the trend of making really good attacks which are “balanced” because “it’s very punishable” if it whiffs, aka half the PM cast. Instead we’ve reduced endlag to make it a safer defensive option, while not allowing it to be used as an approach because it’s a kill move (semispike) after all.
The last bit is to slightly modify the length of the first Agility zap so that it exactly reaches the top platform on Yoshi’s Story starting from a teeter at the ledge. I don’t know if this makes it longer or shorter, but it’s basically Pichu’s version of the Pikachu Battlefield zaps. Yoshi’s is terrible for Pichu so he deserves a little help here. But if we expect it to be dropped from the legal stage list, maybe forget this change.
Oh, did I mention? Pichu already has a walljump. So there’s no way for me to give him one.
22. Mr. Game and Watch
Uh, you should probably give him the ability to L-cancel.
23. Kirby
Probably fix his grabs and get-up attack. Other than that, I’m out of ideas about what to do with the rest of the characters, including these last two.
————————— end %%%%
Anyway I was of course making fun of PM and SDRemix a lot while making many similar decisions to what they did, but it’s only in the spirit of improvement, I find both of those mods to be a lot of fun and if I didn’t think they were good I probably wouldn’t still think modding Melee is a good idea. The main difference you see above is that there are many fewer changes but they tend to be stronger. In SD Remix, Pichu got a buffed wavedash, dashdance, and aerial mobility; here, he only got the second one, but his run speed is now notably higher than Fox (which was intentional). I think this makes it easier to know what to fix if one of the characters turns out to be meta-breaking; there are fewer changes, and their impact is more obvious.
I don’t really expect the above changelist to be the final changelist (assuming this project gets off the ground). It’s just too likely that I did something stupid across 50 or so different major changes to character mechanics. The characters I left out might still be fixable in a way that doesn’t make them really obnoxious, although on that note I do like the idea of increasing Bowser’s weight.
So the changelist is:
Peach:
* Wink-eye turnip gains additional upwards knockback (kills Jigglypuff on the ground at 120 on FD)
Ice Climbers:
* Nana AI uses side-B and airdodges to recover
* Nana AI catches projectiles with probability equal to Popo’s % (or a fixed % if that won’t work lol)
Captain Falcon:
* can walljump out of side-B
Pikachu:
* f-air landlag reduced from 26 to 18
Samus:
* charged shot beats some other projectiles (needles)
* gets a larger shield
Yoshi:
* Egg Lay hits frames 15-18, total 26 frames
Dr. Mario:
* Super Sheet hits on frame 10, ends frame 28, can be used 3 times to slow falling
* Up-air hitbox slightly larger in top-front corner
Luigi:
* Fall speed 1.6 -> 1.83
* Can walljump, including up-B-cancel
Ganondorf:
* tech roll length increased ~15%
* grab range extended downwards slightly
* down-throw base knockback increased slightly (cg easier on fastfallers, harder on others)
Mario:
* Cape hits on frame 10, ends frame 28, can be used 4 times to stop falling
* Aerial fireball animation autocancels when Mario lands
* Fair knockback angle 280 -> 240 (reverse spike)
Young Link:
* Aerial Up-B gets a frame 1 fire hitbox behind YL (where he holds his sword to charge)
* Back-air has invulnerability frame 10-17.
* Roll frames 37 -> 31 (even with most of the cast)
Donkey Kong:
* Up-B reflects projectiles
* Aerial up-B can be cancelled into a special fall
* Can walljump
* Charged f-smash gets additional upwards knockback (kills Puff at 20% on FD)
Mewtwo:
* Weight 85 -> 102
* Outer tail hitbox removed
* Confusion ends frame 41 instead of 55 if it whiffs (old data on hit), reflection starts frame 6
Link:
* Aerial Up-B gets a frame 1 hitbox behind Link
* Roll frames 37 -> 33 (even with Yoshi)
* Traction 0.1 -> 0.045
* Shield reflects projectiles (incomplete cover of Link’s body; does not reflect fully charged Charge Shot)
Roy:
* Dash speed 1.47 -> 1.72 (run speed unchanged)
* Fall speed 2.4 -> 2.7
* Gravity 1.14 -> 1.34
* Flare Blade stops Roy from moving when used in the air
* Nair damage 8/7/6 -> 15/12/9, sourspot knockback increased slightly
* Can walljump
* Blazer goes 30% higher, can grab ledge from behind
Pichu
* Dash speed 1.72 -> 2.1, run speed 1.72 -> 2.25
* Fair, Agility, Thunder Jolt no longer damage Pichu
* Gravity 1.1 -> 0.8
* Pummel damage 3->5
* Down-smash IASA frame 39 instead of 51
* Agility’s first zap is just long enough to slide onto the top platform on Yoshi’s Story by doing an upside-down L-shape from a teeter
With that I open the floor to anyone who wants to complain that this changelist is terrible and I should have done this or that or whatever. If you take away anything from this, I hope it’s the idea that the changelist can be developed from community knowledge and experience and do better than just trying random things and see if it works, since we can get the right ideas faster and with less effort if we know what to look for.
Last edited:
