May 24, 2007

MLB 2007 salary vs performance interactive chart.: See which teams are getting the most bang for their buck, over the course of the season (created by ben fry, via kottke.org). You'll also find links to charts for the 2006 and 2005 seasons at the bottom of the page. I'd like to see one for the EPL, dammit!

posted by worldcup2002 to baseball at 05:02 PM - 5 comments

Previously.

posted by justgary at 05:38 PM on May 24, 2007

Ha, yes. That was me, two years ago. Good times, good times.

posted by worldcup2002 at 10:50 PM on May 24, 2007

Still very interesting.

posted by justgary at 11:36 PM on May 24, 2007

Fairly impressive what the Indians are doing so far with their "miniscule" payroll. Well, as miniscule as $61 million can be considered.

posted by The_Black_Hand at 06:27 AM on May 25, 2007

Grum's posts in that previous thread are quite wise- I agree, it's a nice notion, but a better one would be money/win, adjusted for average- not minimum- salary and likely minimum wins. I also think adding in the wrinkle of run differential (Expected WL in the ESPN expanded standings) and the Pythagorean theorem for scoring would be the real indicator, at least when looking at an in-progress season. Baseball has many elements of chance, but as a mathematician's delight, it has two great equalizers:

  • It is played over a long season (162 games) allowing trends such as batting average or winning streaks to regress to the mean
  • The offense and defense are all but completely isolated from each other, allowing the success of one to not impact the success of the other
As a result, you can really ask "If a team spends X dollars to field a team, and scores N runs and allows P runs... are they:"
  • Overall winning the amount of games you'd expect for their runs scored and runs allowed (likely the case)
  • Differing from the league averages in payroll by the same margins as they are the league averages in runs scored and runs allowed?
This means a team could be well above average in payroll, yet have a mediocre W/L as a result of a balanced runs scored and allowed, because they have a fantastic pitching staff- costing a lot of money- and not enough offense, resulting in a team ERA of ~4.00 but with an average runs scored of about the same. And since the pitchers and hitters are basically separate personnel, you could analyze "Did this team overspend on a top-notch staff, which is doing it's job, but with the result of a .500 record because the remaining money bought the offense that we're seeing?" If a team is doing poorly in relation to overall payroll, but can be seen to have overspend on one side of the roster, they might not be getting poor "bang for their buck"- they might just have spent too much of their money on the wrong things! For example, right now the Red Sox are the best team in baseball (yay!), with a 34-15 record. However, the runs scored and runs allowed predicts that their record should actually be... 33-16. They have the best ERA in the league, while they score the second most runs behind only Cleveland- unsurprising that they are in first, so clearly their payroll is being well-spent. The question then would be, "Are they spending more $ than they'd need to for this level of success, which is predicting a 108-win season?". Most teams in the AL are at 1-2 games off their "expected" pace, although 3 are underperforming by 4-5 games: the Yankees, the A's, and the Rangers. At least one of these 3 teams will likely regress to the mean over the remaining 110 games, and as a result put together a very nice winning streak. If NY stays behind their expected winning percentage, we can say that "well, they are the third-highest scoring team in the AL... and the third-highest runs allowed. Their hitters are great, their pitching sucks." Granted, anyone watching the Yankees could tell you this, but this means if we evaluate "bang for buck" we should evaluate them as two separate measures: the Yankees might be getting a steal on their hitting staff for what it produces on the whole, yet doing poorly on the pitching side of the equation.

posted by hincandenza at 11:51 PM on May 27, 2007

You're not logged in. Please log in or register.