Physics-based platformer movement is essentially just shmup scoring made less abstract - from how it feels like to engage with to how its designed. The main difference is that platformers use the most intuitive feedback & the most future proof metric - real world physics simulation & time. Both use very similar design patterns, and can benefit to learn from each other. Thinking about it this way can help demistify scoring system design.
I made a video talking about scoring systems, and there I mentioned a few distinctions which can be directly applied & carried over to movement-based games. I'm assuming you watched the vid :^) and will apply these concepts directly to movement :
Linear vs Exponential Scoring - This is very directly applicable since some games like Quake/Defrag let you build speed continuously with no cap, making you go exponentially quicker the more you can go without fucking up. Burst-style platformers like Megama Zero, on the other hand, are very linear, even a major fuck up towards the end doesn't necessarily matter because you can recover from it with better performance right after.
Positive vs Negative Scoring - This concept is arguably even clearer when applied to movement than scoring. Positive movement is about actively gaining speed via your inputs or actions, dashes are the most obvious example. Negative movement is about minimizing or preventing speed loss, such as making sure you're going downhill after hitting a boost pad, or avoiding obstacles.
Reversed Risk vs Reward - Directly applies to platformers as well. If the biggest speed gains come from the most obscure technical tricks then you're essentially forcing the players to engage with the full extent of your movement right off the bat. Difficulty isn't a good disincentive, it'll just make people bash their heads against the wall harder.
Depth vs Clarity - This is essentially my Burst vs Continuous Movement post. However, movement-based games have an advantage here because of how intuitive they are by default, compared to the abstract nature of scoring.
Additionally, I also have articles talking about different types of movement in platforming (click on the terms). And I think the two compliment each other very well, so lemme apply them to scoring :
Burst vs Continuous Movement - Like movement, scoring gain can also either happen in bursts because of specific actions/tricks you take (think screen clearing bombs), or can be continuous and extremely granular (think grazing, as long as we're not talking about sitting-on-emitter strats). In general, the games tend to trend towards granularity because scoring gain is almost always tied to enemies & bullets rather than player mechanics per se. But the way games end up playing can still end up being very burst-like, especially when tricks boil down to catching bullets as they're spawning from an emitter, or the famous BIG JUICY CANCELS.
Looping Movement - Just like with movement, shmup scoring can become pretty loopy if it's not context-dependent enough. Just like with burst vs continuous, this is less of a problem in shmups. Still though, as granular as for example Psyvariar is, it does often boil down into a simple dive into emitter -> motorboat bullets kinda loop.
If you view velocity as a type of resource that can be converted (and it is, on a programming level) via different types of moves, it makes the games more directly comparable to shmups with resource-like multipliers such as Ketsui's chip system. In Ketsui, you build up chips by killing enemies with your shot, then spend chips by killing enemies with your laser, applying the number of chips you have as a multiplier to their base value. Essentially you gain chips, then convert them into score by considering the enemy base value and planning the areas where you'll be laser-ing vs shooting.
In movement games, you build up velocity as a central resource, and often you can convert it into either jump height or different types of forward velocity (like by swinging in grappling hook games). Each stretch of the game also has a "base value" of sorts - the speed with which you can clear it. And it's up to you to figure out how to allocate your resources to optimize your speed through each section.
Scoring is future proof because it reincorporates good strats back into itself allowing players to optimize things more and more, but because scoring is tied to ingame rules there can be exploits like infinite milks. Time is a lot more future proof because there's almost no real way to break it - even the most busted skips will allow you to save a few miliseconds here and there.
Beyond these design patterns, it's also worth it for shmups to borrow the general idea of racing games which is trying to make something that's inherently abstract more intuitive to understand by tying it to real world stuff like physics. Shmups already do this with stuff like coins, creating kinaesthetic incentives for scoring, but this can go further.
I know this might feel like a stretch, but I promise you it makes more sense the more you think about it & compare the two. Hopefully you find this way of looking at scoring/movement design useful. I know a lot of devs don't really know how to approach scoring design, and this makes it a lot more grounded, I think. It's been really helpful for me personally cause I have always used racers as a comparison when talking about scoring systems.