LAUSR

In his comment [1] on the Contact-Mechanics Challenge [2], Ciavarella reports his difficulties to reproduce one data set of the submission of one contributor, namely the gap–load relation by Persson. The comment could convey the impression that Persson’s model was not fully defined, which might imply by a worst-case extension that Persson’s data were fudged. After all, the term fudge factor appears three times in the comment. I respond to the comment as the corresponding author of the study, who designed the challenge, who collected all data from all contributors, who prepared all figures, and who ultimately wrote the paper with helpful feedback from many co-authors and steady encouragement from two Tribology Letters editors. First and foremost, I feel the desire to state that my own reference data had not been disclosed to anybody before I collected the results. In some cases, I noticed potential errors in the submissions, which appeared unrelated to the used methodologies themselves, but seemed to be unit-conversion problems or similar misunderstandings. In these cases, I told the contributors where I believed the errors to be, asked them to consider my findings and to resubmit. This was done without the passing on of any reference data or of hints if some predicted number was too large or too small. Persson was the second participant of the challenge to submit his data. There was no need to ask for resubmission, because his data matched my full Green’s function molecular dynamics (GFMD) [3]-based simulations within the small deviations that usually occur when both theory and simulations are carefully conducted. I can thus assure that there was no fudging in Persson’s or any other submission. As a personal note I may add that I used to be skeptical of the good match of his theory with experimental data myself. This is why I asked Persson in 2011 to make predictions on the Reynolds flow through the thin gap between a randomly rough surface (defined with mathematical precision) and an elastic manifold with equally well-defined elastic properties. During the write-up of this personal challenge, I did not share my data with him either, until I prepared the figures of the final write-up, which was published in Physical Review Letters [4]. This procedure between Persson and me is now established whenever we collaborate. In his comment, Ciavarella asks for the results to be completely accessible. As the corresponding author of the Contact-Mechanics Challenge, I rewrote all model descriptions such that anybody with access to the internet and my modest background in contact mechanics would be in a position to reproduce the gist of every contribution. In some cases, it took many iterations and perhaps some frustration on behalf of my co-authors until I felt to be in a position to (roughly) reproduce each submitted data set from scratch. The one contribution that I would not dare trying to reproduce is the stunning experimental work of Greg Sawyer’s group. As a two-left-handed theorist, I should not be allowed to enter a real laboratory. However, it should be clear to anyone that a paper such as that summarizing the Contact-Mechanics Challenge needs to remain within a reasonable page limit and that this cannot be achieved if every theoretical approach is described in such excruciating detail that every reported number can be reproduced to several decimal digits. While it might have been appropriate to state explicitly that the elastic energy expression from Reference [5] was taken, the description of Persson theory in the Contact-Mechanics Challenge closes with a clear reference to my own work with Wang [6], where we explain and extend Persson’s submission, for anyone interested in more details. Persson’s submission happens to be the only one that I recomputed (with my own, independently written code), i.e., the one that Ciavarella reports to be irreproducible. I was dissatisfied by the small number of data sets that Persson had submitted and wanted to find out for myself how well * Martin H. Müser [email protected]