Sorry, buddy, as a scientist I can tell you that one experiment performed by a couple of people doesn't make something "proven by science", no matter how many capital letters you use. It only demonstrates that, given the specific assumptions they make and their experimental setup, a certain result might be expected.

Among the issues with their research methodology: I'm guessing it was done under dry conditions? Short trials? On a trainer? No incline? They performed the test under conditions that would be maximally beneficial for using platforms.

To be fair, one should consider other scenarios. One problem with platforms is that, in general, none of them are perfect at giving good traction between foot and pedal, and the more they try to ensure good traction, the more it hurts when they fail. I'm betting those studs don't feel good against calf or shin.

Additionally, as I read the abstract it did show a 2% advantage in output, which is a decent amount. Look at it this way, that's almost certainly more than the difference between Sora and Dura-ace components, or similar arguments for other componentry (wheels, frames, etc). Put yet another way, if you're going to go to the trouble of buying a bike for more than $600, you should get clipless pedals too or you're wasting your money.