Originally Posted by
Kapusta
A sample size of one does not validate anything.
Are you purposely avoiding looking at the 5th chart (the one with the 25mm Continental GP4000s II tires)? The roller drum data is different that the real world data (even the brand new asphalt). So for one tire the drum data is the same as smooth asphalt and another tire it is different on smooth asphalt, and then VERY different on rougher asphalt.
How can this possibly be interpreted as validating drum data as a accurate predictor of real world results?
The authors make several conclusions in this article. What you are claiming is not one of them.
Are being purposely obtuse? The second plot applies a model to translate roller data to road. The sample size is not one, but three for number of data points taken on the road. The lower plot does not apply the model to the roller data and just shows the roller and road results as is.