3 ms·

>In a 2010 article, Stanford law professor Mark Lemley, who is known for his careful statistical analysis of patent cases (and is cited amazingly often by the S

>In a 2010 article, Stanford law professor Mark Lemley, who is known for his careful statistical analysis of patent cases (and is cited amazingly often by the Supreme Court), noted that the win rate for patent owners in the Eastern District was 40% That statistic is meaningless without the base rate, which turns out to be 32.5%. I'd also point out that there's something very weaselly about Lemley saying the district "is not in the top five", when what he really means is "it's the sixth most". It's technically true, but looking at N and saying "greater than N - 1" is a way of increasing the rhetorical impact of a figure without actually lying about it. It implies that N could fall anywhere over that threshold, when in fact you've picked the threshold right under N. Perhaps a more objective way to say it would be to say that the Eastern District of Texas's patent win rate is in the 81st percentile.

VLM 13y ago
My strategic advice would be to research the shape of a graph of all states; is there anything peculiar about Texas? It might be possible to change the locale without changing the outcome. A similar analogy would be declaring poverty to be the bottom quartile of financial resources, then being surprised that as long as more than 5 people exist we'll always have a bottom quartile therefore always poverty. What I'm getting at is there may be nothing special about TX at all, there always will be an 81st percentile of course. For example if we sell TX back to .mx, perhaps in exchange for oil or something, then obviously TX would cease to be the 81st percentile but some place else would become the 81st percentile, probably with no real change in the situation other than where the bizjets fly to. Or maybe TX is corrupt as heck but no one knows it... why TX? Everyone knows Chicago is super corrupt, NYC is super corrupt, N.O. is somewhat corrupt, the whole state of Louisiana is corrupt, the whole state of Alaska is corrupt, the whole state of Illinois is corrupt... so why bypass those probably more fertile grounds for ... Texas of all places? On the big list of cities and states with proven documented corruption problems, east texas never seems to make it very high on the list. IF I were a crook who wanted to hang out with crooks, I see no reason to go to rural TX if I could simply go to Chicago with likely better results, for example. I mean, haven't the last three governors of Illinois gone to prison on corruption charges? It would appear if you're looking for crooks you'd go to IL not TX, but the trolls don't, which is weird.
- saryant 13y ago
  These judges aren't appointed by Texas, they're appointed by the President. We're talking about a federal district court.
  - dctoedt 13y ago
    > These judges aren't appointed by Texas, they're appointed by the President. Here's how it really works. From the names of those lawyers who have applied to be considered, a name is "suggested" by whoever is the state's senior senator of the same party as the president. Assuming the vetting checks out, the suggested person is then nominated. If neither senator is of the president's party (as is currently the case in Texas), then somehow the congressional delegation of the president's party gets involved; I forget the details. (The late federal judge Norman Black in Houston, appointed by President Carter, was both well-liked and well-respected. He used to tell a joke: A trial lawyer is someone who knows how to try cases. A litigator is someone who knows how to take depositions. And a federal judge is someone who knows a U.S. senator.)
- mistercow 13y ago
  I mostly agree with you. I was just pointing out that there was some bias in the way that Lemley characterized the situation, and that it's not very useful to say "40% win rate" without citing the average nationwide. In fact, it's fairly deceptive, since in the absence of anything else to go on, people may tend to assume that the base rate should be 50/50 (that would be maximum entropy, after all), and might think that 40% is low. In fact, it's quite high; but it I would agree that it doesn't seem so high as to be completely out of whack (except insofar as a 30% win rate on patents nationwide is probably pretty out of whack).

VLM 13y ago

My strategic advice would be to research the shape of a graph of all states; is there anything peculiar about Texas? It might be possible to change the locale without changing the outcome. A similar analogy would be declaring poverty to be the bottom quartile of financial resources, then being surprised that as long as more than 5 people exist we'll always have a bottom quartile therefore always poverty. What I'm getting at is there may be nothing special about TX at all, there always will be an 81st percentile of course. For example if we sell TX back to .mx, perhaps in exchange for oil or something, then obviously TX would cease to be the 81st percentile but some place else would become the 81st percentile, probably with no real change in the situation other than where the bizjets fly to. Or maybe TX is corrupt as heck but no one knows it... why TX? Everyone knows Chicago is super corrupt, NYC is super corrupt, N.O. is somewhat corrupt, the whole state of Louisiana is corrupt, the whole state of Alaska is corrupt, the whole state of Illinois is corrupt... so why bypass those probably more fertile grounds for ... Texas of all places? On the big list of cities and states with proven documented corruption problems, east texas never seems to make it very high on the list. IF I were a crook who wanted to hang out with crooks, I see no reason to go to rural TX if I could simply go to Chicago with likely better results, for example. I mean, haven't the last three governors of Illinois gone to prison on corruption charges? It would appear if you're looking for crooks you'd go to IL not TX, but the trolls don't, which is weird.

saryant 13y ago
These judges aren't appointed by Texas, they're appointed by the President. We're talking about a federal district court.
- dctoedt 13y ago
  > These judges aren't appointed by Texas, they're appointed by the President. Here's how it really works. From the names of those lawyers who have applied to be considered, a name is "suggested" by whoever is the state's senior senator of the same party as the president. Assuming the vetting checks out, the suggested person is then nominated. If neither senator is of the president's party (as is currently the case in Texas), then somehow the congressional delegation of the president's party gets involved; I forget the details. (The late federal judge Norman Black in Houston, appointed by President Carter, was both well-liked and well-respected. He used to tell a joke: A trial lawyer is someone who knows how to try cases. A litigator is someone who knows how to take depositions. And a federal judge is someone who knows a U.S. senator.)
mistercow 13y ago
I mostly agree with you. I was just pointing out that there was some bias in the way that Lemley characterized the situation, and that it's not very useful to say "40% win rate" without citing the average nationwide. In fact, it's fairly deceptive, since in the absence of anything else to go on, people may tend to assume that the base rate should be 50/50 (that would be maximum entropy, after all), and might think that 40% is low. In fact, it's quite high; but it I would agree that it doesn't seem so high as to be completely out of whack (except insofar as a 30% win rate on patents nationwide is probably pretty out of whack).