Using the same retosheet gamelog data as before (2000-2008), I took all of the intraleague and interleague runs scored and allowed data and grouped it by year, home team league and visiting team league.

I then looked at the difference between runs scored and allowed based on whether the opponent was in the same league or not.

First, here are the runs scored and runs allowed differences between facing teams in your league versus the other.

Both numbers are set so that higher is better [RS is how many more runs per game you score in interleague play; RA is how many fewer runs you allow in interleague play].

To put it together, the final chart has the total RS+RA:
The NL stacks up well from 2000 through 2004, but there was a HUGE change in 2005 that has sustained itself over the last four years [and from the 2009 records, I'd guess the gap is still pretty similar].

Perhaps the next thing to look at is what particular thing or things happened between 2004 and 2005 that would have caused a shift like this.

[Sky: Great FanPost from Eric.  Bumped to the front page.]

I've often seen articles and blog entries online about how great interleague is for attendance.  And then some numbers get quoted that confirm that.  From 2000-2008, the average interleague game had an attendance of 33,284 and all other games 29,981.  That is +11%!  Wow, interleague play is a GREAT thing for the teams/owners.

Well, not quite true.  The distribution of games isn't the same.  66% of the interleague games in the sample were Friday through Sunday, compared to just 45% for the non-interleague games.  I ran the attendance figures by day of week and got the following:

Continue reading this post »

My first WAR graph... so be gentle... :)  I was reading something about Adrian Beltre recently and noticed that considering he's just 30, he seems to racking up a pretty decent WAR... Alot of that has to do with him starting young, but even his WAR/600 PA stacks up with better players than I think he is perceived as being.

Anyways, here is Beltre vs Delgado, Brock, Clemente and Juan-Gone, all through age 29 [all four were within 10% of Beltre's WAR/600 through age 29 and all of them accumulated atleast 4500 PAs through age 29]:

I took the last 20 years of player salaries and adjusted them to 2008 values; I did this by taking the average salary for each year and calculating the ratio between that and the 2008 average.  I then multiplied the salaries by the ratio of the 2008 average to their average.

While the average player's peak in terms of performance may be around 26-30, their peak in terms of getting paid is 33 to 37.  Breaking the players down into three piles of about 6000 player-seasons each:

18-26 $998K

27-30 $3.2M

31-50 $5.1M

Using OBP*TB as a quick and dirty measure of run value [and assuming that the hitters and pitchers pay distribution was similar], the old players [31+] provided, slightly less value than the middle group, yet cost over 50% more.   The young group provided 85-90% of the value of the other groups, but for 20-35% of the cost.

I'm pretty sure I knew that young players were a bargain [free agency rules and arbitration process pretty much ensure this], but I didn't think the other side of 30 was THAT bad of a bargain.

Inspired by the previous post on the 2009 Hall of Fame voting compared to the players' career WAR, I was curious how things looked from previous years.

I grabbed all of the BBWAA ballot results [through 2008] for players who debuted 1955 or later and had atleast 3000PA in their career. 

Just some quick factoids:

• Hal Lanier (-5.3 WAR), Jesus Alou (-1.6) and Sonny Jackson (-0.5) are the worst players that made it onto a Hall of Fame ballot.  It has to be an impressive feat to manage to be a below replacement level player for ten or more years and get your name on to the Hall Of Fame Ballot.
• Perhaps even more impressive...?  Bobby Richardson (5.8) had enough support to get himself onto THREE ballots.  The next lowest that were on multiple ballots are Manny Mota (2 ballots, 15.4 WAR), Bill Virdon (2, 17.4) and Leo Cardenas (2, 23.7)

I dropped all of the data into a scatter plot and with a linear trendline and it had a .247 R2.  I was thinking that voters HAVE to be getting better and better over time; Breaking it (roughly) into fifths:

• 1972-1984, R2 .285
• 1985-1990, .226
• 1991-1995, .139
• 1996-2002, .191
• 2003-2008, .413

Not entirely sure why they would have initially gotten worse and worse at voting, but definately looks like the age of sabermetrics has changed things over the last dozen years or so.

The players inducted on their first ballot (21) had a median career value of 72.7 WAR; the [elidgible] players not inducted on their first ballot were all below that.  That is a good sign atleast. 

Here is my data if anyone cares to look at anything else with it:

www.buttonsarenttoys.net/ESPN/attachments/WAR.xls