Shoutouts to Stinger Mag for the roll-jump loop gif nobody else would have uploaded this shit
Playing Penny's Big Breakaway got me thinking about Looping Movement in platformers. It's when there's a very clear, simple, one sided flowchart the player can follow to build or preserve speed, and of course strong incentives to do just that.
Penny's does an ostensibly cool thing and gives you full, unrestricted access to your grappling hook. At first I thought this was a great way of handling it since it removes context sensitivity from your grappling hook, making it a universal player input-driven mechanic. But the problem is that said context sensivity is what makes movement less loopy by tying it more directly to terrain.
In games like Umihara Kawase & Getting Over It, for example, the strength of your main movement tools completely depends on what sorts of terrain you have around you (ignore rocket jumps for this analysis to work pretty pls). As a result, their application is extremely varied and doesn't really boil down to a flowchart, you're always doing all kinds of micro-adjustments to what you use as an attachment point, where exactly you attach to, etc. Where you swing also completely dictates your subsequent options.
Penny's isn't like that - the grappling hook can be summoned anywhere at any time, as a result the only meaningful considerations you have to make are momentum going in (either forward momentum or falling) and which direction you're going to move to. Your dash is even simpler since it doesn't act as an effective "momentum converter". What you end up with is constantly accessible and near-universally powerful movement tools dominating the flow of gameplay - looping movement.
Movement loops signal that the game you're making or playing doesn't have enough interdependency between the player's moveset and the environments the player moves through. It's only when your moves usefulness and properties change based on context that your game's movement can break away from this. The important part is to make the "default" type of terrain in your game and the default moves the player can do the slowest, least effective options compared to more situational, unique stuff.
Less loopy, more context sensitive movement makes for more meaningful levels with more design space to create really distinct movement challenges (& rewards) that won't just get overriden by the player's skill at the universal mechanics. And they make for smoother integration of routing and gameplay mechanics. It's just important to make sure these situational mechanics naturally build on the player's moveset instead of just being glorified boost pads.
Looping movement isn't necessarily bad though, and with enough granularity to each "step" of the loop, you can make an engaging game solely through the player's mechanics. Quake/Defrag are great examples of this - it's engaging even on a flat plane.