Okay, I'll play.
Force = mass x acceleration. Power = work/time (work = force x displacement which is also equal to any change in the total mechanical energy of the system). There is also the factor of the angle between the force and displacement vectors.
The data in the article refers to force, although the discussion seems to use force and power interchangeably. No There are also no parameters given for the force measurements. Are they peak forces, averages over the range of motion of the swing, or what? Sinc no time parameter is given for the force measurements their relationship to power is unclear. I believe these are also ground reaction forces measured with a force plate, not measurements of forces acting directly on the kettlebell.
Since data presented is force and the stated conclusion is about power, it is not clear how the conclusion was arrived at. What are the relevant time parameters to determine power? The fact that the parameters of the force measurements are not clearly explained make it even more difficult to see the relationship between the data and conclusion.
But this just means that the connection between the data and the conclusions is unclear. There are also a few reasons I would question the accuracy of the conclusions.
Let's look at power as work/time. Although there may be value in looking at peak impulse (force/time) and I'll comment on that below, let's start by looking at the total time of a rep and the total work done during that rep. At the top of the swing, the KB is floating so it's total mechanical energy is all potential (m x g x h). At the low point of the swing, the potential energy is lower because the height is at its low point, but the bell has a lot of kinetic energy (.5 x m x v2 -- velocity squared). At the top of the back swing, the KB is again motionless, so it's energy is all potential again, but less than at the top because the top of the back swing is lower than the top of the up swing.
So in looking at the difference in power between reps with different masses, the relevant question is, "How much different is the total time/rep between a lighter KB and a heavier one?" We know the differences in the masses, and the heights at the top and bottom of the swing are basically the same (although a couple of complicating factors are if we can't swing the heavier bell to the same height and the extent to which the lifter might be forcibly arresting the height of the up swing). So the heavier bell will require more power because of it's greater mass, but is this mitigated or outweighed by a lower average velocity/longer total time? I don't know the answer to this question.
The article mentions that the eccentric force increases with bell size. The eccentric phase of the swing is largely negative work because the force exerted by the lifter is in opposition to the displacement (although the angles are not straightforward) and acceleration is negative (the bell ends up at zero velocity at the end of the back swing). My question would be, if the negative work on the back swing keeps going up with bell size, why isn't that proportional to the postive work the lifter had to do on the up swing? I don't know the answer to this question.
Now I'd like to consider the force measurements in the article as if they are peak force measurements (speculatively, since the article does not make it clear that this is the case). One thing I've notice in moving up in bell sizes is that with a lighter bell, you can explosively INITIATE the hip drive and gun the bell out of the hole. With a heavier bell, you have to ramp up the power of the hip drive and explosively FINISH the hip drive for maximum power. So I can imagine that there is a sweet spot where peak impulse might be higher with a given bell size (depending on the strength of the lifter) and decrease with a heavier bell. But whether this could be defined as higher power would depend on the time frame you are using to define power. I also don't know whether this would be more a function of bodyweight or the strength and skill of the lifter.
Finally, the idea in the article that lower power is produced with heavier bells because of the need to stabilize seems extremely speculative, but is presented as a definite unqualified fact. How was this determined?
@Bill Been, do "internets" have cash value? Are they like Bitcoin?