Reader Challenge: Similar Squared
Posted by Raphy on March 4, 2010
One of the interesting pieces of information available on the bottom of each of the player pages here at Baseball-Reference.com is the list "Similar Players". These lists,which utilize Bill James's similarity scores detail the the top 10 players to have comparable statistics to the player in question. For players who spent a significant amount of time batting and pitching there are 2 lists - one for hitting and one for pitching.
Today's challenge is to find 2 players who were so similar that they appear both of the other's lists, both as hitters and as pitchers.
March 4th, 2010 at 12:37 pm
Tim Keefe and Warren Spahn? Both sporting sub-.200 career averages, but they both must have been swinging for the fences - 47 HR combined
March 4th, 2010 at 12:39 pm
I'll go with Burleigh Grimes and Ted Lyons.
March 4th, 2010 at 1:04 pm
Rick Reuschel and Jim Bunning meet this criteria.
March 4th, 2010 at 1:04 pm
Bob Forsch and Vern Law.
March 4th, 2010 at 1:06 pm
Rick Reuschel and Jim Bunning.
March 4th, 2010 at 1:09 pm
Jerry Reuss and Jim Bunning.
March 4th, 2010 at 1:10 pm
Larry Jackson and Dutch Leonard
Tommy Briges and Virgil Trucks
(I guess it isn't a super-rare occurrence, is it?)
March 4th, 2010 at 1:11 pm
(That should be "Tommy Bridges")
March 4th, 2010 at 1:59 pm
Wow. Wboenig, you're right. There are a bunch of these.
Since we've already essentially linked them I added Reuschel and Reuss to the list.
Here is what we have so far.
Looks like Jeff's Forsch and Law combo is in the lead.
March 4th, 2010 at 2:08 pm
Is anyone #1 on both each others lists?
Candy Cummings and George Zettlein are close. Candy is #1 on both George's list, but George is only #2/#1 on Candy's lists. There's got to be a mutual #1 pair in there.
It'd be fun if there were some players that weren't exclusively pitchers, too. George Sisler types. That would make it much harder to find a similarity pair though.
March 4th, 2010 at 2:21 pm
Wow. Nice find David. I was hoping to find that as well.
In terms of totals Cummings and Zettlein are 916+978=1894.
March 4th, 2010 at 2:22 pm
Oh. We're switching to max similarity (with both on each list)?
Mike Scott and Ray Culp: 957 + 988 = 1945
March 4th, 2010 at 2:24 pm
Not switching. I just thought that it might be fun to keep track.
Scott and Culp take a sizable lead.
March 4th, 2010 at 2:35 pm
Two more in the "1900 plus" category:
Herb Pennock and Waite Hoyt (930 + 978 = 1908)
Jesse Haines and Larry French (920 + 986 = 1906)
more as they are found ....
March 4th, 2010 at 2:40 pm
Al Hollingsworth and Eddie Smith (952 + 984 = 1936)
No more postings from me unless they can beat 1936 (all CAPS if they beat 1945!)
March 4th, 2010 at 2:48 pm
A NEW LEADER ON THE BOARD:
HUGH MCQUILLAN AND GENE CONLEY
969 + 985 = 1954
March 4th, 2010 at 2:48 pm
Just out of curiosity:
How are you guys coming up with these players? Are you using a particular strategy?
March 4th, 2010 at 3:10 pm
Click->PageDown->PageDown->Click->PageDown->PageDown->Click->PageDown->PageDown.
Its pretty random. 🙂
March 4th, 2010 at 3:12 pm
Start with any pitcher who has both sets of similarity scores. Click on the link for the batter that has the highest similarity score, and go to his page. Look for names that appear on both the pitching and batting list. If none, repeat.
I use the batting list because those scores are invariably higher than the pitching similarity scores.
It also helps to have cookies enabled so that players whom you have already looked at will show up in purple rather than blue.
March 4th, 2010 at 3:13 pm
Eppa Rixey and Jack Quinn
866 + 968 = 1834
Not terribly similar, but I love the name Eppa Rixey
March 4th, 2010 at 3:20 pm
Jack Quinn also connects to Jack Powell (859/962/1821)
Eppa Rixey appears on both of Jack Powell's lists, but Powell is only on Rixey's pitching list (#1)
Also on Jack Quinn's lists twice is Waite Hoyt, but Quinn is only on Hoyt's pitching list (and Rixey is on his batting list but not pitching list)
With all that said, Jack Powell doesn't appear on either of Hoyt's lists and vice versa...odd
March 4th, 2010 at 3:27 pm
Ok, this really is very common. Just found two more immediately:
Old Hoss Radbourn and Kid Nichols (832/940/1772)
Old Hoss Radbourn and Mickey Welch (916/935/1851)
Incidentally, go read the sponsorship banner on Old Hoss Radbourn's page, it's a good one.
March 4th, 2010 at 5:14 pm
We had a discussion of this on the statlg-l mailing list in July, 2003. Deacon Phillippe was both the most similar pitcher to Sam Leever and the most similar hitter (and, of course, they were teammates for several years). Christy Mathewson had Pete Alexander as both his most similar pitcher and hitter. Gaylord Perry had Don Sutton as his most similar pitcher and his 2nd most similar hitter.
March 4th, 2010 at 6:12 pm
The amazing thing is that Jesse Tannehill and Jack Chesbro show up on the Phillippe/Leever sim lists too -- the four of them made up the great Pittsburgh rotation 1900-1902.
March 5th, 2010 at 9:29 pm
How 'bout Early Wynn and Ted Lyons?
March 6th, 2010 at 9:43 pm
I don't think that Wynn and Lyons make each others pitching lists.