Data Offers Deeper Legal Insights On How Judges Might Rule

The data may not be as important as the interpretation. JL

Sara Randazzo reports in the Wall Street Journal:

New tools, mined from millions of court documents, offer lawyers statistics on the likelihood of a lawsuit’s being dismissed, for instance, or the average wait time until a trial. Lawyers say the data can help temper client expectations, influence courtroom decision-making and even save money by flagging strategies unlikely to succeed. (But) “when you tell lawyers sophisticated mathematics are an important part of their arsenal, that’s not universally applauded.”
It’s like “Moneyball” for judges.
Lawyers looking for an edge in court are increasingly turning to hard data to predict how judges might rule, in some cases long before the judges put pen to paper.
New tools, mined from millions of court documents, offer lawyers statistics on the likelihood of a lawsuit’s being dismissed, for instance, or the average wait time until a trial. Lawyers say the data can help temper client expectations, influence courtroom decision-making and even save money by flagging strategies unlikely to succeed.
“Everybody knows litigation is wildly expensive and risky, but the fact is, all of those risks can be quantified,” said Kirk Jenkins, a Chicago lawyer who blogs about the behavior of the highest state courts in Illinois and California based on years’ worth of data analysis.
Traditionally, judges’ reputations have spread through anecdotes, often gathered by lawyers’ sending a firm-wide email to colleagues. This has led to broad insights, like knowing that the Eastern District of Texas is a plaintiff-friendly patent court, or that a certain judge is short-tempered or gives out light sentences.
Now, a slew of services, from companies including Lex Machina, Ravel Law and Bloomberg Law, are offering far more granular information about judges.
For securities cases filed since 2009, the Northern District of California terminated cases in a median of nearly five months, according to Lex Machina, compared with just over a year in the Southern District of New York.

 U.S. District Judge Susan Illston in San Francisco, who has presided over cases involving Wal-Mart Stores Inc., Uber Technologies Inc. and others, granted motions to dismiss in full nearly 48% of the time over the past five years, and denied such motions in full 24.5% of the time, Bloomberg Law data show. Defendants looking to transfer a case away from U.S. District Judge Gregory Sleet in Delaware can see on Lex Machina that he has granted such requests 62% of the time in cases filed since 2000, compared with the 53% national average.
Mr. Jenkins, a partner at Sedgwick LLP, said his data has run counter to some common beliefs. While many lawyers and judges say it is impossible to predict the outcome of an appeal based on oral arguments, he found that at the Illinois Supreme Court, if the judges ask your side more questions, “your odds drop like a rock.”
At Ravel, which is digitizing Harvard Law School’s 43,000 bound volumes of U.S. case law, attorneys can find “specific language judges like to use,” said Daniel Lewis, the San Francisco company’s co-founder and chief executive. Ravel uses machine learning to spot connections between cases and patterns in how judges rule.
Denver litigator Eric Olson said that while using Ravel, he learned a judge “made it clear she does not like sports analogies,” which are rife among lawyers. “I definitely would never use ‘moving the goal post,’ ” in front of that judge, he said.
Steven Bauer, a patent lawyer at Proskauer Rose LLP, said that in a recent pitch to a company looking for the best place to file a patent-infringement lawsuit, he compared differences in speed and case outcomes from several courts.
A few clicks on Lex Machina showed that in recent years in the Northern District of Illinois, the median time to trial for a patent case is nearly four years, while in the Southern District of New York, it is roughly one year and eight months. The client is still deciding where to file the suit, Mr. Bauer said.
Judges don’t seem to mind being tracked, though some wonder if the data is attempting to give deeper meaning to decisions than actually exists. “I look at it with a little bit of a skeptical eye,” said Shira Scheindlin, a recently retired federal judge who now works as a mediator and arbitrator. Averages on how judges rule may be misleading, she said, without drilling deeper into the data to see if differences emerge between a securities case and a prisoner lawsuit, for instance.
Even the data tools’ creators caution that used incorrectly, aggregates can lead to false conclusions.
Lawyers say the tools have cut back on tedious research projects given to junior lawyers or law librarians, and go beyond what was possible from such assignments.
Scott Breedlove, a partner at Texas-based law firm Vinson & Elkins LLP, said he advised a client being sued for patent infringement not to file a certain type of motion to dismiss after seeing that a judge didn’t favor them. That has freed him up to pursue other strategies in the case, which is about to begin the discovery phase.
Mr. Jenkins, who has a background in economics, sees data as the way of the future for litigation. However, in the notoriously slow-to-evolve legal industry, he said, “when you tell lawyers that fairly sophisticated mathematics are an important part of their arsenal, that’s not universally applauded.”