Hacker News new | ask | show | jobs
by caconym_ 2626 days ago
The title, "Review of Test Data Indicates Conservatism for Tile Penetration" is completely ambiguous to the point of being nonsensical. It is impossible to know in which direction the author of the slides thinks we should be conservative.

Ironically the slide you point out ("Damage Results From 'Crater' Equations Show Significant Tile Damage") seems to skew our interpretation of the next slide's title and content in the wrong direction. The former slide states that "'Crater' indicates that multiple tiles would be taken down to densified layer". That sounds bad. However, it also tells us that the "program" (assuming this means the Crater model) that generated this alarming prediction was "designed to be conservative". This would seem to downplay any concern generated by this result. Furthermore we are told that "Crater reports damage for test conditions that show no damage", further casting into doubt the predictions of the model.

And then, on the next slide: "Review of Test Data Indicates Conservatism for Tile Penetration". On the previous slide the word "conservatism" was used to tell us that the results from the "Crater" model may be on the high side, i.e. showing a problem where there is none, and that the test data show a much smaller degree of damage. This slide is about test data ("Review of Test Data"). On both slides we are told that compared to test data, the Crater results are inflated: "Crater reports damage for test conditions that show no damage", and "Crater overpredicted penetration [...] significantly".

Where does this leave us? The context of this additional slide makes the presentation even more misleading than the "asshole author", to use your delightful term, thinks it is. The title is not merely ambiguous; we are explicitly nudged toward the wrong interpretation of it. I have to thank you for bringing this additional context to my attention.

The author may in fact be an asshole (though I think your language is inappropriate in context), and his analysis may lack depth (I think it does to some degree), but in the simple matter of this slide being egregiously awful he's totally right. I don't know why you'd choose this hill to die on.

1 comments

This is a great point. The beginning of the disputed slide continues to build up a point—Crater (the prediction software) is overly conservative. The major caveat that the test data is far afield from the actual situation is buried at the bottom of all that.