Sports Statistics

[cbloom]

Baseball and basketball have decent statistics now.

Football is still a huge problem because the teams don't play enough games, and also there are way more unknown variables you have to factor for.

eg. in baseball if you want to factor out a hitter's raw performance you only factor out a few other factors - the pitcher, the defense, the ballpark, and his position in the lineup. In football if you want to factor out for a WR's performance, you have to factor out the QB, the running game, the O-line, the defense, which DB he gets, the other WR's, the style of play calling, etc. etc.
There are too many unknown variables and not enough data points (eg. games) so you can't solve the system.

[Dids]

Zod,

There's a lot that happens between 98 and 03 that's meaningful. I'm pretty sure a few folks from BP were hired (Keith Law by Toronto specifically) prior to 2003. Moneyball didn't spaw those hires, it just reflected them.

[Aloysius]

[ QUOTE ]
John Hollinger and David Berri are doing some interesting work on the NBA in terms of statistical analysis. John Hollinger is on ESPN and David Berri wrote Wages of Wins.

[/ QUOTE ]

There's alot of good statistical work being done by analysts and NBA teams. I think more relevant stats and analysis are emerging, and in time a stat culture (not on par with baseball and football, but closer) will emerge in the basketball community.

Sports Forum Thread Discussing Relevance of Basketball Stats

-Al

[El Diablo]

Jason,

Others have piped in with stuff about sabermetrics and other more advanced stats. Here's the fundamental problem, though.

Most people who are into sports are casual fans and very results-oriented. So, stats about "last minute comebacks" and "touchdowns thrown" and "points per game" are going to always be more prevalent than stats that are actually more significant to the actual outcome.

[Kneel B4 Zod]

[ QUOTE ]
Zod,

There's a lot that happens between 98 and 03 that's meaningful. I'm pretty sure a few folks from BP were hired (Keith Law by Toronto specifically) prior to 2003. Moneyball didn't spaw those hires, it just reflected them.

[/ QUOTE ]
yes, true. Law was hired in 2001. and the Red Sox:

When John Henry bought the Red Sox in 2002, he appointed Theo Epstein's former boss at the Padres, Larry Lucchino, as President and CEO. At the end of the 2002 season, Lucchino appointed Epstein to replace interim GM Mike Port.

Under the regime of Henry, Lucchino, and Epstein, the Red Sox have stressed the disciplines of sabermetrics, the analysis of baseball through objective evidence and methods. In 2002, they hired the father of sabermetrics, Bill James, to be a special advisor to the team, and also hired statistical analysts such as Eric Van and Voros McCracken (famous for his groundbreaking work on BABIP).

[ALawPoker]

[ QUOTE ]
Jason,

Others have piped in with stuff about sabermetrics and other more advanced stats. Here's the fundamental problem, though.

Most people who are into sports are casual fans and very results-oriented. So, stats about "last minute comebacks" and "touchdowns thrown" and "points per game" are going to always be more prevalent than stats that are actually more significant to the actual outcome.

[/ QUOTE ]

Indeed. It's sort of like the government. Our elected officials could be much more efficient, but they have to cater to the masses who have preferences of their own even if they lack the knowledge to make their preferences an objectively good solution.

[Eagles]

From what I have heard Wages of Wins is not good. There was a long thread in the sports forum about this but there were numerous flaws with it. I'll try and find it.
Link
Link

[cbloom]

[ QUOTE ]

Most people who are into sports are casual fans and very results-oriented.

[/ QUOTE ]

So true. People have no concept of how huge variance is in sports. Even on this forum with people who should be smart & know poker, the Sporting Events forum is full of insane result-orientation. The reality is that two teams or two players could be nearly identical in true ability, but through variance one of them becomes a "clutch performer" and an all-time legend and the other becomes a "choker".

[punkass]

stats that kill me are the ones like "They are 12-0 when they score more than 120 points in a game." "Green Bay is 24-0 when Favre doesn't throw an interception."

I am really expecting a baseball stat like "The Braves are undefeated when they allow no runs."

[suppasonic]

[ QUOTE ]


I am really expecting a baseball stat like "The Braves are undefeated when they allow no runs."

[/ QUOTE ]

Well, they are.