asgelle

I have no hard data at hand, but I think I have a response to the question. What it comes down to is the conversion from work done as measured by a powermeter to calories consumed is based on some estimate (guess) as to the efficiency (used in the proper sense) of the rider. We usually take 25% as an estimate which is generally close enough. However, we know that efficiency isn't a static number for an individual. It depends on many factors and will vary with the balance between sub and supra threshold work. This could be important if a very precise value for calories consumed was needed, but then you'd be measuring the individual's efficiency over a range of parameters in addition to workload (e.g., cadence) and not relying on a population estimate in the first place.