Okay, so I tested my JP AND US copies of Marvel Super Heroes.
JP:
No Cart = Cut animation, average slowdown.
Action Replay = Full animation, worse slowdown.
4MB Cart = Full animation, worse slowdown.
1MB Cart = Full animation, average slowdown.
The normal game cuts animation to run "smoothly", the 1MB cart adds animation and still runs "smoothly", while the 4MB/AR adds the animation and then has some sort of memory address issues (just like Metal Slug) and gets additional slowdown. It also accounts for all the Saturn users who have been playing the import version with an AR over the years. They use the AR for the import unlock, but they get the botched RAM support with it. End result is a slower game.
US:
No Cart = Cut animation, worse slowdown.
Action Replay = Full animation, worse slowdown.
4MB Cart = Full animation, worse slowdown.
1MB Cart = Full animation, worse slowdown.
This is what I don't get. Adding the RAM carts did restore the cut animation (I used Hulk for this and the JP tests), but not only did slowdown not improve across the different carts, it seemed universally worse than the JP version of MSH. This wouldn't be the first time that a US version of a game on Saturn had worse slowdown than the JP version (see Saturn Doom for an example), but I guess I expected the US and JP code for MSH to be the same.
Based on this anecdotal testing, I would say that the ranking of the ways to play MSH on Saturn goes like this:
1: The JP version with a 1MB RAM cart.
2: The JP version with any other type of cart.
3: The JP version without a cart.
4: The US version with any cart.
5: The US version without a cart.
Someone more familiar with the game than me might disagree, though. I probably don't have the best eye for this type of testing.
I saw a store locally that has a US copy of "In the Hunt" for $20 that I'm thinking about snagging. It usually brings about $40-50 on Ebay...
Make sure to not plug in your controller until after the Kokopeli logo or 75% of the time it will freeze.