There's always been a certain kind of tension in the sports world between stat heads and traditional analysts. Every sport has a set of traditional knowledge, things that people who have played the game just know and understand as part of the game and how its supposed to be played. These insiders and analysts don't take too kindly when people focusing on stats, "quants" as they're sometimes called, come along and dispel or at least call into question things everyone already knows is right, no matter what or how many numbers and data sets they have to back it up.
A famous example would be Michael Lewis's Moneyball, a book discussing Oakland A's manager Billy Beane and how he used advanced and non-traditional statistics to find undervalued production. In the book, on base percentage is preached over batting average, walks are king and sacrifice bunts and stolen bases shouldn't be bothered with. There's a lot of detail and logical arguments supporting these ideas, mostly based around the fact that the most important thing on offense in baseball is to avoid making outs. Of course, you still see teams move guys over with bunts and send runners, and batting average is still prominently mentioned.
Who knows if there will ever be a perfect coexistence between the box score people and the "watch the game" people or even what that would look like, but there's no question that you need both. Just looking at stats often ignores elements that don't show up in any box score such as momentum, leadership, and a general understanding of the flow of a game. Not taking into account stats at all leaves someone vulnerable to small sample sizes and misinterpreting fluke occurrences as patterns. Neither is necessarily better than the other, rather, they inform and complete each other.
While Moneyball might be the most famous example, advanced stats are making a push in basketball as well. Often, these stats can validate or disprove certain things that observers notice through watching the game.
One of the biggest stories in the NBA this year has been the explosion of Chicago Bulls point guard Derrick Rose. He established himself as a solid point guard in his first two years in the league, but has really busted out this year, averaging about 5 more points and 3 more assists a game while also raising his PER (player efficiency rating) from 18.60 to 22.72. While he's exceptionally fast and has incredible athleticism, the knock on him has been an inability to hit jump shots. He seemed to have worked on it over the summer, and last night's game against San Antonio led ESPN's Ric Bucher to tweet, "I try like heck not to be reactive, but it's time to stop talking as if Derrick Rose doesn't have a J. Or three-point range. He has both."
And based on last night's game, that's true. He was 5-6 from 16-23 feet (long twos) and 2-4 from 3 point range. But again, we're dealing with a really small sample size (one game). While his 3 point shooting has seemed to improve (up to 34.2% from 26.7%), his improvement from 16-23 ft has been marginal (2%) and he's shooting far worse from 10-15 ft (30%, a huge decrease from 50% last year).
Is his long range shooting better? Yes. But it should be noted that he's still in the bottom half of regular point guards in 3 point percentage, and the shot he improved most (long twos) is the most inefficient way to score in the game.
Don't get me wrong. This is no knock on Rose at all. He is a phenomenal player. It just strikes me as kind of ridiculous to make a disclaimer against being reactive, then proclaim his jump shot to be fixed based on one night. On the other hand, his improvement and the impact he has on this Bulls team can't be stated in any box score or statistical summary.
And this is where watching the games and tracking stats have to meet.
Stats courtesy of Hoopdata
