I think it's almost impossible to prove whether horses are getting faster or not using speed figures. I know most people here wouldn't agree with that, but IMHO all figure makers have biases built into their interpretations of the results whether they think so or not. Not all those biases are the same either. Even though some figure makers try to correct for those biases, the methods are not perfect.

IMHO it's very obvious that a person's view on which factors impact final time and which don't can lead to a self-fulfilling movement (or lack of movement) in the figures over time because of how they interpret results.

I think the cushion depth idea would have a lot of merit if we had data from a lot of tracks over a long period of time.

I also think a study of turf racing would have some merit even though the races develop differently and that could impact the results to some degree.

In the end, I don't think it matters much because if they are getting faster, they aren't improving fast enough to impact our wagering.