Testing and machine learning

With AI and ML, would it not make more sense that instead of doing just a ftp test, to do a protocol that tests each of your energy levels (like insyd, 4d, or the original allen/coggin tests ) and then the AI would adjust straight from the gun and have perfect intensities for the workouts? Seems like right now it has to fine tune from the zones created scaling from ftp, why not give it the info it needs?

I think the plan is to do away with testing altogether and to use the progression levels in each system or some other max effort/duration based calculation calculate it. It’s just not available yet.

I don’t believe you need AI or ML or all the other buzzwords to do what you’re asking. The power-duration curve already does this. It just requires maximum efforts at a few places along the curve. That’s the beauty of the PDC…it describes physiology.

Maximal efforts, in my opinion, are absolutely necessary in training. You don’t know what you’re capable of until you do what you’re capable of. Then you can plan based on your capabilities.

You can go submax all day, but that doesn’t describe what your max is. And using a percentage off a population of data is no good. Same reason canned plans are not that good.


You can already do a lot of this with WKO5 or maybe Xert.

The thing is, how does the AI become a top sports scientist and coach all rolled into one? If the science was clear on how much of what type of intensity to do at any time of the year for a particular athlete then we’d all be doing that now and there would be no questions.

I have WKO5. I used 1/10th of it’s capabilities but so I think it is money well spent.