I have had occasion to add a new loss to my list of failures. I convinced the district court, but the Fifth Circuit reversed. 

The good news: I told my client I wouldn't bill them if I lost, so they don't have to pay for the work. 

The bad news: I don't know why I lost, unless you take it the judges are just legal realists: you can't be a tug and a tow, so there's no need to parse the precise language of the complaint. Something I keep coming back to is how to contend with probability. I won in front of the district court. You'd think that would mean I'd have a great chance on appeal, for two reasons: first, district judges get some deference - particularly, those experienced in the subject matter; second and more importantly, federal judges usually get the right answer. Then I lost, and the panel didn't ask me a question. 

So this article about probabilities is significant. 

When we think probabilistically, we are less likely to use adverse results alone as proof that we made a decision error, because we recognize the possibility that the decision might have been good but luck and/or incomplete information (and a sample size of one) intervened.

Maybe we made the best decision from a set of unappealing choices, none of which were likely to turn out well.

Maybe we committed our resources on a long shot because the payout more than compensated for the risk, but the long shot didn’t come in this time.

Maybe we made the best choice based on the available information, but decisive information was hidden and we could not have known about it.

Maybe we chose a path with a very high likelihood of success and got unlucky.

The problem, for which I don't have an easy answer: with such a small sample size, how do you figure out which is which?