Similarity Scores – How Many Clicks to Babe Ruth?
Posted by Andy on December 15, 2010
Reader Matthew Bohnert wrote in describing a game that I've actually played myself with the Baseball-Reference.com player pages. Matthew writes:
"One bizarre and silly suggestion for a time-waster. I am fascinated by Similarity Scores. I can’t really say why. But I am.
One mindless game I like to play is “How Many Clicks To Babe Ruth”? The game starts by picking any batter. Any - at random, from any point in baseball history. Click on their page. It could be Felix Millan. It could be Ted Williams. It doesn’t matter.
Then you start plotting a strategy whereby you click through the Top 10 Similarity Scores shown on the bottom of the page, linking your way to Babe Ruth in as few clicks as possible. Hank Aaron obviously takes only 1 link. Somebody like Enzo Hernandez or Johnnie LeMaster usually takes about 25 or 30. But it can be done. I have yet to find a player where it’s impossible."
I went back and asked Matthew to send me his click progression to get from Johnnie LeMaster to Babe Ruth and here's what he sent back:
"Here's the link progression for Johnnie LeMaster. It's not optimized....in other words, I just fumbled through it right now and you could certainly cut out some 3-4 link loops and probably find a faster path.
It's also fairly typical for a light-hitting infielder, whereby you mess around and don't get anywhere for about 10-12 links, and then you stumble upon a good player with some decent power totals, like Ryne Sandberg. Once you get traction there, you know it's just a matter of time before you link your way to the promised land.
I've also found that Vern Stephens pops up frequently as a critical link around steps 10-12, to connect underwhelming batting careers with the Sultan of Swat.
1. Johnnie LeMaster
2. Dick Schofield
3. Mike Gallego
4. Brian Giles
5. Gene Mauch
6. Jerry Dybzinski
7. Ron Gardenhire
8. Tony Pena
9. Kiko Garcia
10. Mike Phillips
11. Frank White (here's where we start to take off)
12. Bill Mazeroski
13. Roberto Alomar
14. Ryne Sandberg (now we know it's just a matter of time)
15. Lou Whitaker
16. Joe Morgan
17. Craig Biggio
18. George Brett
19. Dave Winfield (closing in for the kill)
20. Frank Robinson
21. Hank Aaron
22. Babe RuthSo 22 links from Johnnie LeMaster to Babe Ruth.
What a great game! You can have fun with your friends. Impress chicks. Wow business associates. All on Baseball-Reference.com!"
I imagine a lot of readers out there have played this game already. What have you discovered?
December 15th, 2010 at 7:58 am
I would think a tool like the Oracle of Bacon could be modified to provide the optimized answer to compare your "manual" attempt against.
December 15th, 2010 at 8:17 am
Just drove myself crazy trying to do this for Juan Pierre...
December 15th, 2010 at 8:28 am
I tried to find a faster route from Brett to Ruth but couldn't do it.
December 15th, 2010 at 8:29 am
and haha I was doing it for Alfredo Griffin! Went crazy and gave up.
December 15th, 2010 at 8:29 am
I was doing this last week...
December 15th, 2010 at 8:40 am
It took me 14 to get to Babe.
1. Johnnie LeMaster
2. Charlie O'Leary
3. Everett Scott
4. Roger Peckinpaugh
5. Luis Aparicio
6. Red Schoendienst
7. Billy Herman
8. Joe Sewell
9. Dereck Jeter
10. Pete Rose
11. Ty Cobb
12. Stan Musial
13. Lou Gerigh
14. Babe Ruth
Totally cool game! Congrats!
December 15th, 2010 at 8:40 am
Catchers are tough. I tried Bruce Benedict and couldn't get there. I thought that I had a path out, once I started finding guys who were C/1B, but it never went anywhere.
December 15th, 2010 at 8:43 am
Alfredo Griffin
Don Kessinger
Phil Rizutto
Jose Offerman
Adam Kennedy
Julio Lugo
Jeff Blauser
Juan Uribe
Jose Hernandez
Davey Johnson (was in the league 14 years and had 136 homers, but once had 43 in a year!!!)
Charlie Hayes
Juan Encarnacion
Jaques Jones
Casey Blake
Preston Wilson
Pete Incaviglia
Jose Cruz
Jesse Barfield
Dean Palmer
Jay Buhner
Troy Glaus
Adam Dunn
Ralph Kiner
Albert Belle
Hank Greenberg
Albert Pujols
Jason Giambi
Carlos Delgado
Jeff Bagwell
Frank Thomas
Jimmie Foxx
Lou Gehrig
Babe Ruth
33 links.
December 15th, 2010 at 8:48 am
@6
I think you are mixing columns or something. There isn't a direct similarity link from Derek Jeter to Pete Rose.
Is there a way to edit posts here? I could've cut Albert Belle out of my Griffin-Ruth sequence lol.
December 15th, 2010 at 8:51 am
@9
Yes there is. Similar Batters Through 36.
December 15th, 2010 at 9:08 am
Great Time Waster. 15 links from Babe Ruth to Matt Stairs
Babe Ruth
Frank Robinson
Al Kaline
Billy Williams
Jim Rice
Joe Carter (1st non-HOF)
Gary Gaetti
Brian Downing
Sal Bando
Larry Parrish
Torii Hunter
George Bell
Ben Oglivie
Jeff Burroughs
Matt Stairs
Curiously, it doesn't work the other way. For example, Larry Parrish is in the Similarity Score list for Sal Bando but it isn't reciprocated. Also, you need Burroughs to get from Oglivie to Stairs though Oglivie is on Stair's Similarity Score list.
December 15th, 2010 at 9:12 am
@ #11 Indeed going backwards from Ruth is a very different game, since HIS 10 most similar players are not necessarily the same players for whom Ruth shows up on their lists. There are lots of cases of guys who do not show up on each other's lists.
@ #10 That's also a different game, using most similar batters instead of just overall similarity scores.
December 15th, 2010 at 10:13 am
I think a better strategy in 'getting' to Ruth, is to try and find dead ball batters who may have pitched a bit, and to look for an age similarity score.
December 15th, 2010 at 10:18 am
The game is a bit harder if you limit yourself to career comps. In the Lemaster example, the step from Gallego to Giles leveraged that fact that those two were similar at age 27 (both got late starts in MLB).
Using the age comps is sorta cheating in my opinion. Jumping from Luis Rivas to Paul Molitor or Rod Carew is not really in the spirit of the game. My two cents.
December 15th, 2010 at 10:24 am
Oops... wrong Brian Giles! Still... my point still holds.
Got to be a better way to go from Gallego besides wrong-Brian-Giles. Wrong-Brian-Giles is a big step backwards!
December 15th, 2010 at 10:26 am
Yes, I've played this time-waster before, and the best way to get to Babe Ruth is just look for the player in the "Ten most similar" who has the highest home run total, that'll lead you to Ruth eventually. Problem is, you have a vague idea of that for players with long careers, but for more obscure guys it's close to random chance.
For example, I tried this with Ray Oyler, one of the most infamous "good-field, no hit" guys in the last 50 years, but had to give up, since it was literally guesswork, and I didn't want to "cheat" and look up HR totals of possible candidates.
December 15th, 2010 at 10:46 am
Took me 55 links to get from Rey Ordonez to Babe Ruth:
Rey Ordonez
Jose Uribe
Marty Perez
Billy Ripken
Nick Punto
Jeff Reboulet
Chuck Hiller
Roberto Pena
Pat Listach
Carlos Febles
Ken Aspormante
Bret Barberie
Aki Iwamura
Alberto Callaspo
Jeff Keppinger
Asdrubal Cabrera
Alexei Ramirez
Ryan Doumit
Stephen Drew
Tulow
Pedroia
Kinsler
Hill, A
Phillips, B
Sabo
Hammonds
Grieve
Wilkerson
Swisher
Hidalgo
Hill, Glenallen
Davis, G
Morneau
Bay
Horner
Rosen
Wright
Holliday
Cabrera, M
Teixeira
Sauer, H
Buhner
Strawberry
Kiner
Belle
Greenberg
Pujols
Giambi
Delgado
Stargell
McCovey
Thomas. F
Foxx
Gehrig
Ruth
December 15th, 2010 at 10:58 am
A curious sequence climbing the LeMaster-to-Ruth ladder highlights a personal obsession:
12. Bill Mazeroski (HOF 2B)
13. Roberto Alomar (presumptive future HOF 2B)
14. Ryne Sandberg (HOF 2B)
15. Lou Whitaker
16. Joe Morgan (HOF 2B)
17. Craig Biggio (HOF 2B)
Yeah, I know it's a game, not a science -- but there's still something wrong with that picture....
December 15th, 2010 at 11:29 am
You can't get there from Cesar Crespo. Then again, he was apparently uncomparable...
December 15th, 2010 at 11:34 am
Maybe Casey Close already did this and therefore knew that you can get from Jeter to Ruth in no more than 10 steps, since Sandberg is in Jeter's top 10.
December 15th, 2010 at 11:45 am
glad to see i'm not the only one who does this from time to time.
i'll add that i've discovered that there are a ton of no hit backup catchers out there that i've never heard of, and if you are playing this game and get into a pool of them, it is hard to get out.
also, i don't think game can be done for pitchers with cy young or walter johnson as the target, if you stick to career comps only. i think greg maddux broke the link by being too similar to warren spahn.
December 15th, 2010 at 11:52 am
Using age-comps...
Jay Bruce
Barry Bonds
Babe Ruth
... that should open a few back doors.
December 15th, 2010 at 12:10 pm
@Tristan --- LOL, ya... you actually have to have comps for it to work!
December 15th, 2010 at 12:17 pm
The Holy Grail would have to be Bill Bergen to Ruth, wouldn't it? I hope someone else tries this before I have to....
December 15th, 2010 at 12:25 pm
I have drifted through b-ref this way before but never tried to link from one player to another - cool idea!
I got Juan Pierre:
Pierre
Matty Alou
Enos Cabell
Steve Finley
Chili Davis
Dave Winfield
Frank Robinson
Hank Aaron
Babe Ruth
December 15th, 2010 at 12:27 pm
@Anon - that was surprisingly easy for Pierre.
December 15th, 2010 at 12:29 pm
Here's 1 route from Bill Bergen to the Babe:
Bergen
Gabby Street
Charley O'Leary
Johnnie LeMaster
I won't repeat the rest of the sequence from above but we know you can get from LeMaster to the Babe. . . .
December 15th, 2010 at 12:32 pm
@ Paul #26 -
"@Anon - that was surprisingly easy for Pierre."
LH hitting Outfielder - how hard could it be? 🙂
December 15th, 2010 at 12:42 pm
One Holy Grail coming right up! (seasonal shipping delays possible)
December 15th, 2010 at 12:55 pm
With Charlie O'Leary in the mix, you can start playing games with *really* old comps.
Bill Bergen
Gabby Street
Charley O'Leary
Nick Altrock
Minnie Minoso
Jim O'Rourke
Jesse Burkett
Shoeless Joe Jackson
Stan Musial
Hank Aaron
Babe Ruth
December 15th, 2010 at 1:06 pm
How about Archibald 'Moonlight' Graham?
Not bloody likely!
December 15th, 2010 at 1:10 pm
I managed to connect Nolan Ryan with Babe Ruth using just top 10 similarity scores...took 22, and hey! Jamie Moyer!
Nolan Ryan
Warren Spahn
Tommy John
Burleigh Grimes
Dennis Martinez
Jamie Moyer
Red Ruffing
Sam Sam Jones
Earl Whitehill
Hooks Dauss
Wilbur Cooper
Stan Coveleski
Babe Adams
Urban Shocker
Eddie Rommel
Rip Sewell
Pat Malone
Vick Raschi
Sal Maglie
Jeff Tesreau
Carl Lundgren
Babe Ruth
December 15th, 2010 at 2:11 pm
@31
You have to play a certain number of games before bb-ref will post a set of sim scores for you apparently. So no starting point for Moonlight Graham. I found that out while hoping to do John Paciorek.
December 15th, 2010 at 3:24 pm
Still seems like "cheating" to use the Similarity By Age list.
December 15th, 2010 at 3:46 pm
@18,
I noticed the same thing about Whitaker.
December 15th, 2010 at 4:37 pm
#4/ CRTYonker Says: "Still seems like "cheating" to use the Similarity By Age list."
Agreed, to quote from the intro "...you start plotting a strategy whereby you click through the Top 10 Similarity Scores shown...". This is different than the Most Similar By Age listing on the right-hand side. I noticed this when trying to duplicate Juan Pierre-to-Babe Ruth in #25.
December 15th, 2010 at 5:23 pm
Glad to see someone got Rey Ordonez. I'm stuck on Mick Kelleher.
December 15th, 2010 at 5:50 pm
Mick Kelleher
Nate Oliver
Larry Milbourne
Mike Gallego
... which gets you to the Johnnie LeMaster.
("cheating" with using the age-comps column of course)
December 15th, 2010 at 5:54 pm
Hey, DavidRF, my curse is that I'm Cubs fan.. so you know I'm not going to take any short cuts. And how depressing is it that Kelleher and Nate Oliver played for the Cubs?
December 15th, 2010 at 6:15 pm
No shortcuts?
Its much harder using just the career comps list, but I think I just did it.
Mick Kelleher
Jack Heidemann
Jim Anderson
Ted Kazanski
John Boccabella
Todd Cruz
Pedro Garcia
Jerry Kindall
Dave Roberts
Bobby Morgan
ted Lepcio
Gene Oliver
John Wockenfuss
Morgan Ensberg
Jose Bautista
Nick Esasky
Bo Jackson
Ron Kittle
Jim Gentile
Glenn Davis
Dick Stuart
Tony Clark
Cecil Fielder
Jeromy Burnitz
Darryl Strawberry
Ralph Kiner
Albert Belle
Juan Gonzalez
Jose Canseco
Willie Stargell
Willie McCovey
Harmon Killebrew
Sammy Sosa
Ken Griffey
Willie Mays
Hank Aaron
Babe Ruth
.... whew...
The strategy was just try to maximize HR at every step and occasionally trying someone else if I hit a dead end. It was very slow going until I hit Morgan Ensberg. Probably some shortcuts to be had in there. I figured it would go faster after I hit McCovey. Someone could probably improve on that.
December 15th, 2010 at 6:26 pm
Yeah, its probably faster using infielders to jump from 200 to 500 HR. The positional adjustment makes it easy to make some pretty large leaps. Hence the Vern Stephens comment in the original post.
I'm supposed to be working though. Maybe I'll try again later.
December 15th, 2010 at 6:51 pm
What about taking some slight power, no speed guys through the chain to Rickey Henderson? Maybe some of those catchers discussed, someone like a Greg Dobbs or Kevin Maas.
December 15th, 2010 at 7:56 pm
Took me 60 to get from 2 career HR Rafael Belliard to Babe Ruth:
Rafael Belliard
John Sullivan
Enzo Hernandez
Bob Lillis
Ruben Amaro Sr.
Ted Kubiak
Curtis Wilkerson
Abraham Nunez
Billy Ripken
...and then follow everyone after Billy Ripken in #17's post...
December 15th, 2010 at 8:14 pm
Kevin Maas was a first baseman. I always remember the start of his career. If you look in his career log you will see for his rookie year he had 13 home runs in his first 113 PA's! Around that time he hit his 15 th homer against my Blue Jays, and also in that series hit a massive moonshot that looked to be a monster homer but ended up settling down on the warning track. I thought he was gonna be the next Babe Ruth.
He hammered the ball for a .902 OPS that year, and looked to be a big up and comer for the Yankees. He never hit well again though.
December 15th, 2010 at 8:49 pm
@16
Good strategy. I wonder what the Dave Eggler to Babe Ruth chain would look like.
December 15th, 2010 at 8:54 pm
how comforting to know that I'm not the only one wasting endless hours in this way.
I HAVE NO REGRETS!!
December 15th, 2010 at 10:51 pm
#42 - I don't think Rickey shows up on anyone's top 10 similarity scores (though I like the idea...)
December 15th, 2010 at 11:16 pm
The similarity line can be very frustrating! Who could've imagined how long it would take to go from Willie Mays Aiken to Willie Mays! Once I got to Dave Parker, I knew it was coming soon - Parker to Dawson to Winfield to Robinson to Mays. Willie Aiken was born two and a half weeks after The Catch. I wonder how many other players were named after major leaguers (not counting families, e.g. Griffey).
December 15th, 2010 at 11:24 pm
@48, Doug -- Larry Doby Johnson got to play for both the Indians and the White Sox, appropriately.
December 15th, 2010 at 11:26 pm
I got from Griffey Sr. to Griffey Jr. via Cruz Sr, Willie Davis, Pinson, Parker, Dawson, and Winfield.
Still not as good as Bill Swift appearing on Bill Swift's top 10 similarity scores (and vice versa)...
December 15th, 2010 at 11:58 pm
Here's a challenge, go from Wally Pipp to Lou Gehrig.
I don't have the attention span to try it.
December 16th, 2010 at 12:12 am
@51.
1 Wally Pipp
2 Frank Schulte
3 Curt Flood
4 Jim Piersall
5 Enos Cabell
6 Mickey Rivers
7 Matty Alou
8 Juan Pierre
9 Steve Brodie
10 Lance Johnson
11 Mookie Wilson
12 Dan Gladden
13 Chad Curtis
14 Lee Mazzilli
15 Gary Matthews Sr.
16 Scott Spiezio
17 Candy Maldanado
18 Pedro Feliz
19 Ed Sprague
20 Scott Brosius
21 Ty Wigginton
22 Hank Blalock
23 Michael Cuddyer
24 Ben Grieve
25 Brad Wilkerson
26 Nick Swisher
27 Paul Sorrento
28 Glenn Davis
29 Justin Morneau
30 Jason Bay
31 Al Rosen
32 David Wright
33 Matt Holliday
34 Miguel Cabrera
35 Mark Teixeira
36 Hank Sauer
37 Jay Buhner
38 Darryl Strawberry
39 Ralph Kiner
40 Albert Belle
41 Hank Greenberg
42 Albert Pujols
43 Jason Giambi
44 Carlos Delgado
45 Willie Stargell
46 Willie McCovey
47 Frank Thomas
48 Jimmie Foxx
49 Lou Gehrig
December 16th, 2010 at 12:34 am
Another fun one is getting from CJ Wilson to Cliff Lee as Wilson will replace Lee as the Rangers ace. I'll let you do Colby Lewis if you'd rather. Either could become the #1 and both will prove rather tricky. Good luck.
December 16th, 2010 at 1:23 am
@48
Some guy named Mickey Mantle, wonder if he turned out any good?
December 16th, 2010 at 1:26 am
@42
What about Russ Nixon?
December 16th, 2010 at 1:32 am
@48 "Once I got to Dave Parker, I knew it was coming soon - Parker to Dawson to Winfield to Robinson to Mays. Willie Aiken was born two and a half weeks after The Catch. "
Would have been interesting if Winfield would have been a Giant, considering when he was born.
December 16th, 2010 at 5:35 am
Is the path from Rick Ankiel to Babe Ruth quicker as a hitter, or as a pitcher???
December 16th, 2010 at 6:53 am
Thanks, DavidRF....Your effort is awe inspiring.
December 16th, 2010 at 7:46 am
@ 54 oh yeah, Mantle and Cochrane. I once had a friend who was a big Giants fan and whose last name was Willey. He insisted that he was going to name his daughters Mays and McCovey. He caved and gave them boring ordinary names though. Then his wife dumped him anyway. He should have stood his ground!
December 16th, 2010 at 8:22 am
I've played this game for years. It started for me when I tried to click from Babe Ruth to Horace Clarke. I got my brother playing this also.
The interesting thing about this game is Babe Ruth is kind of a pivot point for all players because of his great pitching numbers. As it should be, The Babe is the connecting tissue for baseball's history.
I usually try to make connections by using only 5 clicks. For example, The Babe is only 2 clicks from David Segui and just 3 from Jay Payton! Also, because of the Babe, guys like Frank Robinson are only 3 or 4 clicks from guys like Dwight Gooden and Bob Welch.
December 16th, 2010 at 9:35 am
There was once a proof in a similar game that the number would not be greater than 6. Paul Erdos was a mathematician who published papers with a very wide variety of different colleagues. Someone (maybe Erdos himself, I don't recall) suggested that one could count a mathematician's (or anybody in related fields who published with mathematicians) Erdos number by how many people you had to go through to get to Erdos, where you can link through anyone you have published a paper with.
I believe that it was proven that all finite Erdos numbers were <=6 or some such fairly small number. There were a number of people who could not link to Erdos in this way, but everyone who did at all, could be linked within 6 or 7 steps.
This is related to the 6 degrees of separation concept, where it is astronomically unlikely that any given pair of people anywhere in the world cannot be linked in <= 6 steps through people where each link is between two people who personally know each other.
I would think similarity scores is a tougher nut, because we've artificially limited the space to top 10, while almost all people know many more than 10 people, and a fair number of mathematicians (for instance, Erdos) have co published with far more than 10 other authors.
But that said, 10^6 is around 1 million, which massively overwhelms the sample space of major league baseball players. So I'd hardly be surprised to discover that the largest defined Ruth number is fairly low. A lot of that meandering around with weak players until you find the key guy with decent stats can probably be cut to 2-3 links if we do a real analysis of the graph, and find those weak players who happen to link well.
Players with highly unusual stats should generally represent good link points in this game. In order to jump dramatically in quality, you need people who are much better than you to be more similar to you than most of the players who are about as good. If a player has highly unusual stats, there is a good chance that any player who is unusual in the same way will show up on their similarity list, even if they were much better or worse as a player.
December 16th, 2010 at 9:39 am
Pursuant to Michael's excellent comment at #61 and #1 about the Oracle of Bacon, we have here the Oracle of Baseball:
http://www.baseball-reference.com/oracle/
Where you can connect teammates. Just like Michael says, it rarely takes more than 6 links to connect any two baseball players.
December 16th, 2010 at 9:54 am
I got from Rafael Belliard to Ruth in 45 clicks
Belliard hit a grand total of 2 HR but still played for 17 seasons.
December 16th, 2010 at 10:40 am
@57
What would be cool is linking from Babe Ruth first as a pitcher/hitter, through someone like Monte Ward, and then back to Babe Ruth as a hitter/pitcher (the one that you didn't start with)
December 16th, 2010 at 10:41 am
@61
17K big leaguers is also not 6 BILLION people
December 16th, 2010 at 2:19 pm
Hehe! This is so funny, I've been doing this for years! I do sometimes cheat with ranked by age, but I swear I can get there faster than most of the others. Try some other games like trying to get from Ichiro to Sandberg, it's a blast!
December 16th, 2010 at 2:40 pm
Hey! This is a fun game. But I gave up on Enzo Hernandez after about 100 clicks--I did get it to Billy Ripken, who seems to be on most lists posted above-- thought that might be the way, but I just couldn't find it.
December 17th, 2010 at 6:34 am
[...] on to our fun with Similarity Scores, reader Rick Jennings wrote in with the following trivia [...]
December 18th, 2010 at 4:00 am
I wanted to connect Carlos Quintana to Babe Ruth, but I accidentally switched to pitchers when I clicked on Al Orth. I'm not sure if oscillating between similar pitchers and similar batters is within the rules, but I did stay away from the "Similarity By Age" lists; I know DavidRF considers that cheating. I was happy I included Terry Francona, too (who, apparently, hit about as well as dead ball era pitchers hit).
1. Carlos Quintana
2. Terry Francona
3. Al Orth
4. Red Donahue
5. Red Ehret
6. Jimmy Ring
7. Sid Hudson
8. Slim Harriss
9. Max Butcher
10. Jim Tobin
11. Ed Willett
12. Phil Douglas
13. Bob Rhoads
14. Charlie Ferguson
15. Carl Lundgren
16. Babe Ruth