But he's not against abstract takes on sports stats. Like when he dreamed up the so-called Correlated Gaussian evaluation, which he says is a sort of extension of the Pythagorean expectation that relates team scoring and points allowed to winning percentage: "It came in my sleep. I remember waking up and having the picture in my head."

That'll happen. Well, at least it happens to Oliver, a Caltech point guard during the early years of the 310-game conference losing streak that the college broke last week. He went on to get a doctorate in environmental engineering — then worked on issues like toxic waste containment — while trying to figure out how to engineer a sports career.

He did. After the Seattle SuperSonics in 2004 made him the first paid NBA team analytics consultant, Oliver went on to a similar job with the Denver Nuggets— then last month became ESPN's first director of production analytics.

Meaning? He's supposed to come up with new stats, for various sports, for ESPN.

While a bit secretive about what's in his laboratory Petri dishes, Oliver suggests some common stats will become largely extinct. Like baseball batting averages, which he says should be replaced by a combination of on-base and slugging percentages. And current quarterback ratings should broadened to how much quarterbacks keep their own defenses off the field. Wide receivers catch stats could also include drops.

And the typical tack of just judging hoops teams on how many points they average and give up per game just tells you mainly how fast they play — "and it's not how fast you run that counts, it's how you take advantage of possessions."

Sports talk, he says, will have to evolve. "We have to introduce new language. You can get lazy using convenient language. Like with common discussions of basketball players' per-game points, rebounds and assists. That has to be replaced. How many team wins does he contribute? How often do his passes lead to buckets and free throws?" But he concedes it's easier to find flaws in stats than invent ones: "RBI, in 10 years, will have gone away. What's more of a battle is what it's replaced with."

Oliver would also seem to have potential to help TV networks present their own TV ratings — "as an environment engineer, part of my job was to see how people lie with numbers and I know how to do it" — but now he's devoted to "telling the truth with numbers" because they get around more than any expert: "The numbers see all the games."