@M8R I think something people aren't considering here is the use of the vision sensor. In theory if you can utilize the vision sensor correctly and manage to keep the color codes tuned (just using signatures definitely doesn't work consistently enough, color codes kind of work, but everything seems to require re-tuning in any new lighting scenario), flywheel is infinitely better. Imagine sitting anywhere on the field and all you need to do is aim.
The real unseen design is double flywheel with individually powered sides. First, single flywheel loses a lot of energy to ball spin and double flywheel does not have that issue to nearly the same extent. Second, individually powering the sides allows you to curve the ball laterally and basically hit anywhere, so as an extension of flywheel + vision sensor, I think it's quite possible to create an insane double flywheel configuration to hit anywhere on the field from anywhere else on the field.
From my experience it doesn't seem to matter too heavily whether you have double catapult or flywheel so far. My double catapult seemed a bit faster than flywheel robots at recent competition, but considering how much time I needed to spend loading and aiming, it's about equal in the end.
You have some good points regarding the double flywheel, but I would like to mention some things after a season with a double flywheel:
The real unseen design is double flywheel with individually powered sides.
The double flywheel has this major benefit with the design, in which...
individually powering the sides allows you to curve the ball laterally and basically hit anywhere
But I simply must say that you are potentially over-estimating the precision of the vision sensor. The V5's Vision Sensor, as what I've been hearing, has had some major accuracy problems, and even after filtering it probably would have problems too. Considering these inaccuracies, I am near certain that you would be better off knowing to shoot in close-mid range compared to far range if you do use the vision sensor. And in this range, I ought to believe that a simple flywheel design that is simplistic and straightforward would probably outweigh advanced-calculating programs that can alter the curve of a ball to hit the flag. Another heads up I can say is that having two gearboxes for a flywheel is outright heavy, bulky, and friction will not be equal on both sides. A double flywheel is much larger than a single flywheel and it leaves little-to-no room for a cap-scoring mechanism. Although I was a middle-schooler during that time, I regret having a double flywheel and I should have gone simpler and did a single flywheel.
First, single flywheel loses a lot of energy to ball spin and double flywheel does not have that issue to nearly the same extent.
Considering that you require the same compression and more distribution with motors, and because of motor distribution is having to be split 1/2 and 1/2 on both sides to share similar to the same exact wheel mass as a single flywheel on both sides but only with 1/2 the torque, I will almost guarantee you that a single flywheel would have a faster recovery time, less burnouts, and faster ball speed capabilities than a double flywheel.
My double catapult seemed a bit faster than flywheel robots at recent competition, but considering how much time I needed to spend loading and aiming, it's about equal in the end.
Regarding the 2-Ball Catapult you have. According to what you're saying about time spent loading and aiming, you simply need more driving practice.