Obv his ability to violate punishments is really salt inducing, he's probably the best at it perhaps alongside a few other characters like Yoshi/M2 imo but I'm not really here to talk about them. So him being faster falling isn't a bad change, but I would like that if it's done, his jumps be retouched so he can still do stuff like SH double fair, full hop gyro shot->catch gyro. Give him the same airtime I guess for those.
I'd make his throws less centralized on dthrow (less BKB so it can't combo at 0, more KBG so it also combos for less percent, but also so it kills later than current uthrow) and slightly worse in general (fthrow/bthrow much slower so that you can react to the grab+throw, but one of them mixes up with a reworked dthrow or uthrow, not sure, all of them being weight dependent) but improve his ability to get a raw grab (I've been wanting a boost grab for the longest time, idk why he doesn't have one when his dash attack canceled into an upsmash gets him momentum while canceling into grab on the same frame gets no extra momentum).
So basically since he couldn't dthrow combo at low/0 percent, he'd have to uthrow for a combo or bthrow for positioning, with them being a DI mix-up where bthrow has current angle but good DI for uthrow sends you further out on bthrow and good DI to not get sent far on bthrow gets you uthrow combo'd. Uthrow I guess would be a quick toss upwards, with a fair bit of endlag, so that the only way you get a combo out of it is if they don't DI/DI in for the bthrow. Or uthrow could DI mix-up with fthrow. It makes a difference depending on where you want ROB to be highly advantaged on the mix-up in relation to where he grabs the opponent. The uthrow would probs be like 75 degrees, meaning DI in puts you straight above him, DI out is like ~60 degrees depending on how precisely they DI. So the uthrow ends up having something like current dthrow's knockback scaling, but a higher release point, more endlag, and a steeper angle. Alternatively current uthrow could be reworked to be the combo throw that doesn't work at 0 or mid-high percent, with dthrow being the new throw that's quick and mixes up with fthrow or bthrow.
Next, gyro has increased gravity and throws with less speed. Its KB scaling would be compensated for its loss in speed so KB remains the same, but it becomes less of the cross-stage pressure tool it currently is, since it would only go a bit more than half as far. I think everyone benefits, including ROB mains, if ROB focuses less on gyro to dominate and gets something else in return. Part of it was the boost grab. This also isn't totally a nerf, as increase gravity+reduced speed on gyro means he can cover some edgeguard angles that previously he needed a charged gyro shot to do before. The gravity increase also coincides with his own gravity increase so that he can do a lot of the same gyro tricks he could before.
Dair landing hitbox and an aerial that auto cancels like the first 3-4 frames of its animation. Bear with me. Dair's hitbox would be super weak with low knockback scaling. Dair currently has this weird autocancel hitbox to where he never gets landing lag if the hitbox comes out (except for the first frame which is clearly unintentional and should be fixed as part of this suggestion). There's really not even a use for the dair landing animation, so this idea came out initially to give him that. Anyway, the hitbox is weak enough to jab reset at low percents, but nothing else, and you only get it if you land before. Bumping up the landing lag of dair to 40/20 landing/lcancel'd means the only time you get a good follow-up from this is if you use it for the reset on a knockdown, OR they are high percent enough for the low KBG to get them enough hitstun. Obv it would be unsafe on shield. It could be a meteor or not; if it's a meteor, you get a follow up once they pass the percent threshold for ~24 frames hitstun, if not a meteor you only get a guaranteed follow-up between 24-31 frames hitstun if they don't CC, and after that it causes knockdown if they CC or pops up for a combo, which would be like, 100+ damage. The early autocancel on one of his aerials (doesn't really matter which so long as it's not dair) is early enough that he can only really use it if he boosts very close to the ground. It can't be auto-canceled late enough to be used for SH->boost->auto cancel at any distance, instead it has niche uses in platform movement, being able to boost onto a platform and land in place with less lag then a waveland usually would, and at an earlier time.
Dsmash is a move I've wanted reworked for awhile but haven't totally nailed down. A lot of people think it's good, and it is, but for the wrong reason. I played anther on netplay the other day. He was ASDI'ing up out of my dsmashes more than half the time, on netplay. The move becomes way worse when you do that, as you can just kinda escape with like 8 damage an no follow-up. Far as I know fox is the only exception to this, he can't escape at most percents without hard SDI. It's still good for tech chasing tho, since invinc usually makes them miss the first hit or two which helps them escape the launcher. But I don't think it's really smart design for a move to be so apparently amazing at low level since it kinda skews the perception of the viability of the move, then mediocre at high level where you have to really catch someone slipping up to get off CC dsmash. Anyway this change I could leave or take, I can't figure out how to fix the move without really making it stupidly good or bad or just changing it to a different move. But the idea is that the range and speed of it is reduced, making it harder to actually use for CCing since it loses to good spacing and low endlag more frequently, but give him more reward for landing it even if they escape with ASDI. So the first hitbox is moved back from frame 5 to 6 or 7, and instead of having his arms fully extended, they are bent upward, as with his charging animation. This would actually be a pre-charge start-up hitbox, with a weak inward meteor or spike trajectory, that helps set-up the placement of the move more consistently. The reasoning for this is that with the way the hitboxes are currently set-up, there are weird differences in when the launcher hitbox can connect on ASDI up based on really specific spacing. It's part of my issue with the current design. This start-up hitbox basically means, if you hit it, they're always in a certain range of of locations for the rest of the multi hits+launcher unless they try to DI side-ways, which prevents ASDI up to escape. The start-up hitbox would be the first of the multi hits, so he's down to 3 multi hits and the launcher. The multi hits all have current range, with the first one being frame 9 or 10, and have an increased SDI/ASDI multiplier so that you can still escape by holding up; however they also have increased knockback such that if you're near an edge when you ASDI out, you get sent off at their low, harsh angle (the current KB on the multi hits is low so you just kinda pop up a bit when you ASDI out, but at high percents you can see the low 15 degree angle they send at), and if you're mid stage, you get sent closer to the edge behind ROB with minor frame disadvantage (as opposed to frame ADVANTAGE which you currently have on ROB at most percents for ASDIing out). Lastly a change to the launcher to make it akin to over-the-should angles on other quick dsmashes like peach's or samus's, instead of outward. This is largely to compound its utility as a reversal tool, making it more threatening near the ledge but less so center stage. The endlag+base knockback would be increased to compensate for the new angle, meaning follow-ups would be about as likely as they are now with various angles of DI, since over the shoulder angles mean part of their travel distance they are getting closer to you and thus are easier to combo given everything else is the same. The increased knockback also means they wouldn't be able to CC tech the move anymore (currently this can be done at low percents) which is another thing that makes it fake-good, and instead means they have to rely on escaping upwards.
-TL;DR dsmash is slower with less reach on start-up, more endlag, but better if you get the hit and better if they ASDI out.
Also some minor quality of life buffs/fixes:
Fsmash hitbox lasts more than 2 frames. The ****ing animation looks like it lasts 9 frames. It doesn't need to last that whole duration but jesus ****
Upsmash launcher hits aerial opponents. It's stupid that it doesn't currently. It's already fake-good as an OoS option since you can CC/ASDI to knockdown the launcher to infinity, or ASDI+DI away and never get sent into the strong hit, but it also just straight misses airborne opponents period.
Fix upair GFX, it has four blasts but only 3 hitboxes.
Fix side-b gfx, you can't even tell when the hitbox ends
Make laser hitbox work like falco's in that it repeats a couple times across part of the length of the laser. Right now it's literally trying to hit stuff with a pea sized hitbox that stretches to maybe ROB's width, yet the laser graphics are like as long as FD. obv the hitbox doesn't need to be as long as FD, but maybe like two ROB's widths like falco's lasers are.
Can we please get some cool new animations too, all these ones I'm seeing on reddit are making me jealous
So overall, the tl;dr is
Fall speed increased so punishing him isn't quite as absurdly hard, but jumps reworked to compensate for when he's free to move
Very low percents and high percents, he gets nothing guaranteed off throws but has a good mix-up
Mid percents he still gets free follow-ups tho
His kill throw kills later (currently 120-180 depending on fall speed, should be more like 140 min-idk max), encouraging more mix-up and throwing for position at very high percent
But he gets a slightly easier time landing raw grabs
Gyro doesn't control the whole stage when its in ROB's hand, basically reduced throw distance much like Link/Tink saw for 3.5
Dair gets a new utility for knockdown resets, and kinda helps fill the high percent punishment gap left by old dthrow
Dsmash is harder to hit but more consistent and rewarding when it does