I initially created Baseball Graphs to develop new and better ways of presenting baseball statistics. The 2003 season was my first full-season attempt, and it turned out pretty well. I now spend a lot of my time creating and formatting the baseball stats at The Hardball Times.
Along the way, I've also picked up the writing bug. I spend most of my energy writing for THT, but sometimes I like to write about more geeky stuff, baseball research and things like that. That's what this blog is for.
I'm very interested in graphics and other information display, so I'll probably post links to creative stuff like that. I also plan to write historical articles that complement the historical graphs. And I have inexplicably developed an interest in baseball win-related stats, such as Bill James's Win Shares and something called Win Probability Added. So you can expect a few posts about that stuff too.
If you're interested, I've created a complex spreadsheet full of WPA logic, which you can download at ftp.baseballgraphs.com/wpa. If you'd like to read more about Win Shares, try these links:
Win Shares Intro
A couple of months ago, I posted some work regarding Win Shares replacement level. I won’t go into all the details of why I’m semi-obsessed with this, but it’s mostly due to a sick personality disorder. Ask my therapist.
Thanks to some excellent feedback (mostly pointing out my bad math), I updated the study for relief pitchers and position players, but never posted my final conclusions regarding starting pitchers. And I use the word “final” loosely.
I delayed the starting pitcher results because the methodology is more complex. Originally, I had found that starting pitchers have an extremely low replacement level of 40%. But Tango pointed out that starting pitchers are much more likely to be replaced during the year, compared to position players. And this created problems with the results.
So I went back to the data and placed pitchers in the first group (starting starting pitchers), based on how many games they had started as of the end of May. These guys represent the true top level of starters as of the beginning of the year, and all other starters represent the starting pitcher “bench”. I included pitchers in the “bench” group if he started at least two games, or one game if that was the only game he pitched.
And this change had the expected result, as the replacement level rose to 57% in the American League and 59% in the National League—about ten points below regular players and relievers.
So my final conclusion is that a baseline Win Shares level is 60% for starting pitchers, and 70% for everyone else. I plan to use these two figures in the 2005 Hardball Times Win Shares. As a reminder, these figures are multiplied by each player’s expected Win Shares to obtain a specific baseline for that player.
And I’m going to call it Win Shares Above Bench (or Baseline—just WSAB) instead of Win Shares Above Replacement. Replacement levels have become a controversial term, and “Bench level” is a more appropriate description of the analysis. Thanks to Tango for the idea.
Interesting. But why compare to “bench”? Don’t average bench players have value, relative to the worst bench players or the minor lg players who are even a bit worse than that?
Tango can probably give an accurate answer, but I would guess that a marginal bench players is 3 to 5 runs worse than avg bench player.
As long as a player has been called up to MLB, he has value to someone.
Studes’ metric seems to be “value above average bench”. I don’t see an issue with setting that as a zero baseline, any more than setting average as a zero baseline in a metric called “value above average MLB player”.
Whether you want to take issue with the relevancy of the metric, that’s a different story.
That’s my thought, too, Tango. My thought was to re-think what I call the baseline, because replacement level has so many associations in people’s minds.
What I would personally like to do with this is use it for salary analysis, though other people might want to use it for something else. And my stance would be that a major league team should be able to obtain an average bench player for very little—maybe just a bit above the minimum salary.
I haven’t thought that all the way through, but that’s my idea.
Ok. Out of 25 slots per team, I get about 10 that could/should be termed bench players. So, the midpoint is 5, the avg bench player. So 5/25, or 20% of slots are “below bench”. I’m not comfortable with that as a “general” baseline. I prefer to assume that most of those players have value. Of course, that depends on the availability of talent. If there is such a surplus that the worst 10 players on an avg team are pretty much of the same ability level, then so be it. I suggested that the worst bench players are 3 to 5 runs per 162 G worse than the avg bench player. That makes “sense” to me, but I don’t really have any evidence. But if that is true, I don’t see why you would not want to use that as the “general” baseline.
That makes sense, David. As I said, I’m more interested in a baseline that intuitively feels like it should be worth the major league minimum, or could be adequately filled by someone making the major league minimum. I don’t think that would be the last guy on the team.
However, I’m sure this is more complex than my simple math.
Why not simply keep the baseline as the “average bench player”, and simply say that this baseline is worth 750,000$ (or whatever it is)?
A player that is 0.2 wins above minimum is worth 400,000$ above minimum, or 700,000$. If you want to say the avg bench player is 1 million$ to correspond to 0.3 wins above minimum, that’s fine too.
Really, it doesn’t matter.
(All number for illustration purposes.)
| <<Previous Article: Introducing Baseball Analysts | Next Article: Baseball's Timelines>> |