It's often said that most of the time on a drop bar bike is spent on the hoods, and that feels plausible, but I know with sustained headwinds I've ridden a long way in the drops (and much of the rest on the tops for a stretch on the rare sections out of the wind). To plug in to calculators, and for curiosity, how might I measure the proportion of a (long) ride spent in various positions? I'd assume a solo ride, so any solution would need to be mounted on the bike and not affect riding (too much) though a difference between solo and group rides would be particularly interesting and may explain some comfort issues. Alternatively published data would be of interest, though very limited interest if it relates to pro races.
I imagine one could build something with thin capacitive sensors under the bar tape (or the rubber of the hoods) and a microcontroller. Here is a board with 5 touch inputs you could simply wire up if you have some basic knowledge of electronics and microcontroller programming.
Alternatively you could record the ride on video (mount a wide-angle camera to the bike somewhere) and manually analyze it (or train some machine learning model to distinguish between the three main positions).
Maybe even data from an orientation sensor or accelerometer on the back of the rider would be sufficient to distinguish between the positions. Try putting a smartphone into your jersey’s back pockets and record the sensor data.
However, I doubt the end results would be very useful. Which hand position you pick highly depends on terrain, wind, intensity of the ride, geometry of the bike (and drop bars), the rider, your seating position etc.
I think generally one mainly adjusts the seating position for the hoods because the assumption is that you’d spend most of the time there. So “per definition” the hoods should be the main position. If they are not it means you should adjust your handlebar position.
Also keep in mind that even though there are only 3 main hand positions there are countless variations. For example there is a big difference in gripping the drops right at the end (feels like your shoulders are almost above the hands) vs. at the bend of the bar with elbows bent and a very aggressive riding position.
This answer addresses only how someone could measure the proportion of time spent in various hand positions. Leomo makes motion capture sensors that can be used in bike fit. I have never used them, and I don't even do bike fits, but my understanding is that right now they are still a niche item. In any case, it seems like it might be possible to use motion sensors mounted on your arms to automatically record the proportion of time spent in various positions. If Leomo hasn't already done this, one thing I would want to see is the user or Leomo validating that the sensors can correctly distinguish between body positions on the road - this would require a controlled test. I don't think these are practical for most individuals to use, unfortunately, with the first barrier being the cost.