Bizarro Box-Score Sorcery
Posted by John Autin on June 29, 2011
[Please humor me for a moment; I'll get to the point shortly.]
Here's a fun exercise with the Play Index: Try to find a game where both starting pitchers had the exact same basic box score line of IP, H, R, BB and SO. Here's a step-by-step guide for doing that search:
(To skip the search and see the results for a certain pitching line that I selected at random, click here.)
- In the Play Index, under Game Finder, choose "Player Pitching."
- Set the first option to "Find Number of Players Matching Criteria in a Game."
- Set "Pitcher's Role" = "Starter."
- Under "Select Additional Criteria," set each of the first 4 fields to one of the 5 basic stats (IP, H, R, BB or SO), and equal to any number you may choose.
- Set the 5th field (the relational field) for the remaining stat, and set it equal to 1.0 * [one of the other 4 stats]. This is a workaround to get a limited 5-stat search from a 4-field form.
- Leave the years at the default of "1919-2011."
- Click "Get Report."
When the results come up, look in the "#Matching" column. Does the first entry show "2"? I'll guess that it didn't. No matter what stats you choose, the odds are that you won't find any game where both starting pitchers had the exact same basic pitching line. It's very rare.
Now do another search just as before, but with the following criteria: IP=7, H=6, R=2, SO=0 and BB=R. (To skip the search and see those results, click here.)
At last, my point: This rare event occurred in Tuesday's White Sox-Rockies game. Opposing starters Gavin Floyd and Jason Hammel both had lines of 7 IP, 6 hits, 2 runs, 2 walks and 0 strikeouts.
Besides matching their basic stat line, both Floyd and Hammel induced 12 ground balls, yielded 7 line drives, and had a Win Probability Added of 0.169.*
Also, it was the first time in 30 years that both starting pitchers went at least 7 IP without a strikeout. And it was the first time in 10 years that Colorado collected 10 or more hits at home without an extra-base hit.
Ready for another fun exercise? Pick any two players and peruse their pages on B-R -- biographical info, career stats, etc. How many common traits can you find? By selecting from the vast array of data for each player, and presenting only only the things they have in common, can you make them seem eerily similar -- like those lists comparing the assassinations of Presidents Lincoln and Kennedy?
Here's what I came up with for Gavin Floyd and Jason Hammel:
- 28 years old
- Right-handed
- 6' 6" tall
- 7 losses this year, with a low-4's ERA
- Innings nearly identical -- 101 IP for Floyd, 100.2 for Hammel
- 10 wins last year
- Both are with their 2nd MLB team, and had 7 wins with their 1st team
I stopped there because, well, this is just an exercise. But you are now armed with all the tricks you'll need to dazzle your friends with tales of once-in-a-blue-moon events and coincidences ... no matter how meaningless they may be.
____________
* In case you're wondering, pitchers with the same stat line do not necessarily have the same WPA.
June 29th, 2011 at 6:13 pm
IP=7, H=7, R=3 for both starters:
<TH style="BORDER-BOTTOM: #aaa 1px solid; BORDER-LEFT: #aaa 1px solid; PADDING-BOTTOM: 2px; BACKGROUND-COLOR: #ddd; PADDING-LEFT: 2px; PADDING-RIGHT: 2px; BORDER-TOP: #aaa 1px solid; BORDER-RIGHT: #aaa 1px solid; PADDING-TOP: 2px" class="tooltip ranker sort_default_asc show_partial_when_sorting" onmouseover="" onmouseout="" onclick="" align=center tip="RankThis is a count of the rows from top to bottom.It is recalculated following the sorting of a column." sorttable_columnindex="0" table_number="1" jQuery16108896148938350992="43">Rk
Tm
Opp
<TH style="BORDER-BOTTOM: #aaa 1px solid; BORDER-LEFT: #aaa 1px solid; PADDING-BOTTOM: 2px; BACKGROUND-COLOR: #ddd; PADDING-LEFT: 2px; PADDING-RIGHT: 2px; BORDER-TOP: #aaa 1px solid; BORDER-RIGHT: #aaa 1px solid; PADDING-TOP: 2px" class="tooltip sort_default_asc" onmouseover="" onmouseout="" onclick="" align=center tip="A number in parentheses indicates which game of a doubleheader.Click dates for box scores of games or standings on this day." sorttable_columnindex="3" table_number="1" jQuery16108896148938350992="44">Date
<TH style="BORDER-BOTTOM: #aaa 1px solid; BORDER-LEFT: #aaa 1px solid; PADDING-BOTTOM: 2px; BACKGROUND-COLOR: #ddd; PADDING-LEFT: 2px; PADDING-RIGHT: 2px; BORDER-TOP: #aaa 1px solid; BORDER-RIGHT: #aaa 1px solid; PADDING-TOP: 2px" class="tooltip show_partial_when_sorting" onmouseover="" onmouseout="" onclick="" align=center tip="Number matching the criteria set" sorttable_columnindex="4" table_number="1" jQuery16108896148938350992="45">#Matching
1
CAL
MIL
1995-07-27
2
2
KCR
TOR
1993-07-17
2
3
STL
MON
1985-09-20
2
(Praying for the table to post correctly.)
Three matching games — no big deal. The bizarro part is that one of the starters of the 1985 game took the loss in relief in the KCR-TOR game.
June 29th, 2011 at 6:17 pm
Kahuna buffoona. Sorry about that.
June 29th, 2011 at 6:27 pm
Dukeof flatbush was on the Chris Hansen show!
June 29th, 2011 at 6:48 pm
John: I ran the search for IP = 7, H = 6, R = 2, SO = 0 and BB = R. My search returned 15 results compared to your 53. I extracted the box score for one of your games that did not appear on my list, the game of 6-27-22. The pitcher, Dixie Leverett, had 7 IP, 6 H, 2 R, 0 SO and 1 BB, not 2. Perhaps you hit >= instead of = for the BB search. That is the only game I checked. Perhaps you could clarify.
BTW, in answer to a prior question of yours on another post, Connie Marrero is the oldest living former player.
June 29th, 2011 at 6:56 pm
@3
I just noticed that there is no >= nor a <= selection for the fifth "Choose a stat".
June 29th, 2011 at 7:26 pm
rox/sox was a great game to watch, one of the fastest-paced games i've ever watched
June 29th, 2011 at 8:59 pm
@3, Richard Chester -- You're absolutely right; I used > instead of = in the relational field. Lucky me, it didn't affect the one game I was targeting.
Good point about the lack of >= in the relational field. When I want that function, I work around it by dropping a tiny fraction off the multiplier -- e.g., 0.999 instead of 1.0.
And thanks for the followup on Connie Marrero.
June 29th, 2011 at 9:25 pm
At the risk of going political- who's with me on:
JOHN AUTIN FOR PRESIDENT 2012
I'll register the party "Sabercratican" if someone else can get the bumper stickers printed up.
June 29th, 2011 at 10:23 pm
Gavin Floyd is one of two major league players (so far) who were born on Jan. 27, 1983, and both have pitched for the Phillies. The other is Mike Zagurski. The Phillies have been shuttling relievers between the majors and AAA so much this year that I don't know if he's with them now or not.
June 29th, 2011 at 10:27 pm
Bad news, Rogerbusby -- I inhaled.
June 29th, 2011 at 11:12 pm
What two players had...
* Equal hits
* Both switch hit
* Played 20 seasons
* Had an OBP within 1 point of each other
* Never finished in the top 10 in MVP
* Made the HOF
* Played in New York...
* ... for a team no longer in New York
* Never played in an All-Star Game
YET were more than 40 WAR apart for their career
(This probably isn't TOO much of a mystery but I thought I'd play with JA's idea)
June 29th, 2011 at 11:15 pm
Of no connection to anything, just saddens me.
Mark Reynolds' June ON BASE streak was snapped tonight.
June 30th, 2011 at 12:02 am
On July 27, 1948, Ellis Kinder of the Boston Red Sox picked up a victory over the Tigers with 9 IP, 3 H, 0 R, 0 ER, 3 BB, 4 SO, 0 HR, 33 BF & 82 GSc. One year, to the day, July 27, 1949, Red Sox starter Mel Parnell picked up a victory over the White Sox with the same exact line.
June 30th, 2011 at 12:13 am
@11
Max Carey and George Davis
June 30th, 2011 at 12:23 am
@14
Yikes, Richard, impressive! How did you do it/know it? The Brooklyn and NY Giants part was easy but it would have taken me numerous looks at their team seasons to try and home in on the players. ~~applauds~~
June 30th, 2011 at 12:27 am
@12
Duke, yes. I know you are a big Mark Reynolds booster 🙂 but for HIM to go almost a month with at least one time on base........ is nothing short of incredible.
Now for your next trick, can you get Adam Dunne on a streak of consecutive games without striking out?
June 30th, 2011 at 12:28 am
@13, Jack Fish -- That is a world-class coincidence! Did you find it tonight with the Play Index, or have you been keeping it up your sleeve for years, waiting for the right crowd to drop it on? Either way -- nice work!
And how 'bout Ellis "Old Folks" Kinder? 102 career wins, all after his 32nd birthday! He was pitching in class D through age 26. He and Parnell combined for 48 wins with the '49 BoSox.
Oh, and in the Parnell game you mentioned, he went 3 for 4 himself -- as many hits as he allowed. Kinder merely had 1 for 3 with a walk in his game.
Man, do I love hyperlinked box scores!
June 30th, 2011 at 12:42 am
@15
I certainly did not know it, I never even heard of George Davis. I went to the career hits leaderboard and scrolled until I came to two guys with the same number of hits. They are easy to locate because all you do is to look for a blank space in the Rank column. The first pair was Charlie Gehringer and Ivan Rodriguez and obviously they were not the answer. Next I came to Carey and Davis and compared their stats.
June 30th, 2011 at 12:47 am
Max Carey -- owner of the highest (verified) SB% in a 50-SB season: 51 steals, 2 CS in 1922. (Scored 140 runs that year, too.)
And wow, I'd forgotten that George Davis didn't make the Hall until 1998 -- 89 years after his last game.
Nice one, BSK!
June 30th, 2011 at 12:51 am
Nice work, Richard. I found them on the hits list and reverse-engineered it from there. I didn't think it'd be too hard to find, but wanted to throw a lot of stuff in there that would make ya verify it. Way to go!
June 30th, 2011 at 1:25 am
Bad news, Rogerbusby -- I inhaled.
C'mon JA, this is 2011. Mustachioed Eldridge Obama wrote a book about snorting coke and he run tings.
[Ellis Kinder] was pitching in class D through age 26.
I'm not sure how much those minor league class designations mean back then. For instance, B-R doesn't classify any leagues as AAA in the '40s. The average age of the players in Class D in 1940 is about 23, which is well above the average age in today's Low-A or Rookie leagues, which one might think D evolved into. I'm not sure how many of those teams were affiliated and how efficient the system was in filtering the best prospects to the upper minors. I don't doubt Class D was faaarrr from the majors but I'm not sure what it really meant in terms of a particular player's chances.
June 30th, 2011 at 1:33 am
Hopefully this one is a little bit harder...
These two players...
* are adjacent on the all-time HR list
* had 11 letters in the name they are most commonly known by
* played for the same team
* had 162-game averages that were exactly...
*** 0 SO apart (that is to say, tied)
*** 1 RBI apart
*** 2 BB apart
*** 3 2B apart
*** 4 SB apart
*** 5 points of OPS+ apart
YET one player has a WAR that is 46.5 higher than the other!
Who are they???
June 30th, 2011 at 11:50 am
"* In case you're wondering, pitchers with the same stat line do not necessarily have the same WPA."
This is correct, but there are a large number of cases where they do have the same WPA. In looking for other identical stat lines, I started looking for starting pitchers with the same IP, H and R. (There weren't a whole lot of them, so you can look at them by hand to see if the BB and SO are the same; this should be faster than checking all the combos of all five categories.)
Anyway, in most of the cases where the pitchers had the same IP, H and R, and the IP was an even inning (none with a 1/3 or 2/3 of an inning), they had the same WPA. The only exceptions were games where one of the pitchers started an inning, but didn't record an out (for example, they had IP=8, but in fact allowed a single in the ninth, so it was really 8+ IP). That obviously changes their WPA from where it was after the 8th inning.
In fact, two starting pitchers with just the same IP and R will always have the same WPA if they both stopped pitching after a complete inning. If you think about it, this makes sense: they have put their teams into equal situations (a tie game), so each has contributed equally to the win probability of their team.
June 30th, 2011 at 11:52 am
BTW, I've only checked IP = 8, 8.1, 8.2 and 9 (with various H and R combinations), but no other identical stat lines have shown up yet. A few have matched four of the five stats, missing only in BB or SO.
June 30th, 2011 at 12:09 pm
Another interesting tidbit that exemplifies how pitchers how being used differently than 30+ years ago: six times both starting pitchers in a game have allowed 4 H and 0 R and have been removed after 8 IP. The earliest was in 1990.
June 30th, 2011 at 1:08 pm
@23, Whiz -- That is a good point to make. It reminds me that I wish the official listing of IP in a game would include a signifier for those who pitched to one or more batters in an additional inning but did not record an out.
June 30th, 2011 at 1:47 pm
No body wants to play my game??? ARG! Alright... he is a clue... one player still active, one already in the HOF...
June 30th, 2011 at 1:54 pm
@27
I am still working on it.
June 30th, 2011 at 2:05 pm
BSK -- Seeking clarification on a couple of clues:
-- Played for the same team -- Does that mean each played for just 1 team, or that they had at least 1 team in common?
-- 11 letters in the name they are most commonly known by -- First name, last name, or both combined?
And you wouldn't pull a trick question, right? -- like making it seem that they're hitters by giving all batting stats, but they're actually pitchers?
Thanks.
June 30th, 2011 at 2:05 pm
Richard, I will say that I don't think there is an easy way to find it. Certainly nothing using PI. Another suggestion I would make is consider what factors could make a guy with relatively similar career totals so wildly different in WAR...
June 30th, 2011 at 2:07 pm
JA-
1- They both played for the same franchise. I'll leave the rest up to you. 🙂
2- First and last name any reasonable person would use for both players.
3- No.
June 30th, 2011 at 2:08 pm
JA-
I realize that answer #1 still won't suffice.
If you made a list of the team(s) these guys played on, they would not be mutually exclusive. Any better?
June 30th, 2011 at 2:09 pm
CRAP! There is an error in my stats listed... it should say OPS+ and not OPS. I simply typed it in wrong. Triple-bonus-points to anyone who would have had it if not for my error.
June 30th, 2011 at 2:15 pm
@33, BSK -- Don't sweat it. By the "power" vested in me, I've corrected your original post. I doubt that anyone was close enough to be thrown off by that difference, anyway.
June 30th, 2011 at 2:25 pm
BSK -- Which of these is meant by "adjacent on the all-time HR list"?:
(a) No player has a HR total that falls in between the totals of our two mystery guys; or
(b) They are physically adjacent to each other on the B-R career HR list (such players could have the same HR total, while other pairs might be separated by just 1 HR but not be physically adjacent on the list).
http://www.baseball-reference.com/leaders/HR_career.shtml
June 30th, 2011 at 2:25 pm
@30
I am not using PI. The difference in WAR could be that one guy is(was) an awful fielder and the other guy is(was) good. I have copied the WAR leader boards and pasted them into an excel spreadsheet and added a column which calculates WAR - 46.5 and then tried matching but I have not come up with an answer. I am throwing in the towel because I am missing the Yankee game.
Good luck John.
June 30th, 2011 at 2:29 pm
Thanks, Richard.
BTW, I can think of 2 other possible reasons for a vast WAR difference despite similar rate stats:
(1) Big gap in offensive context of their careers; or
(2) The active guy hasn't played all that much.
June 30th, 2011 at 2:43 pm
Richard-
You probably would have gotten it, had the active player's WAR not changed since last night. In light of that, should I just give the answer?
JA-
To your question, I don't really see an appreciable difference between your (a) and (b) but when I used the term, I was thinking directly in terms of (a).
June 30th, 2011 at 2:44 pm
WAR gap is now 46.3. I'm not sure if enough can change in one game to change the other stats, but it is possible the numbers no longer hold...
June 30th, 2011 at 2:51 pm
@38, BSK -- Glad it was (a).
The difference in my scenarios would be, sometimes a bunch of guys have the same HR total -- but if you look at the career HR list on B-R, they're not all physically adjacent.
June 30th, 2011 at 2:52 pm
JA-
Gotcha, I see what you're saying.
June 30th, 2011 at 3:15 pm
OK, I have the answer. I don't want to print it unless you're ready.
P.S. I used Richard Chester's method, but had to resort to eyeballing the totals that were *close* to 46.3 WAR apart, but not exactly. This points up one of the little irritants of the Play Index: It doesn't always produce the exact same decimals as are shown on the player pages. So, whereas one of the mystery guys has a round number of career WAR on his player page, it shows up as _.9 in the Play Index.
But anyway ... good challenge, BSK!
June 30th, 2011 at 3:17 pm
OK, back to the original topic for a moment.
No guarantee these are the only ones (although I would maybe place a bet on it :-)), but here are two other games since 1919 with two starting pitchers that have the same stat line (IP-H-R-BB-SO) and ER were the same (not one of the requirements JA mentioned):
1964-06-06, SFG@PHI, Jack Sanford vs. Dennis Bennett, 7-5-2-2-4
1968-04-20, ATL@CIN, Pat Jarvis vs. Gerry Arrigo, 7-6-1-1-4
Of course the Game Scores are the same since it uses only these stats.
Honorable mention to this game, in which the ER were different (it matches the criteria JA used):
1974-09-04, SFG@LAD, Mike Caldwell vs. Geoff Zahn, 6.2-7-1-3-4
These Game Scores are different due to different ER.
June 30th, 2011 at 3:34 pm
@42
I have the answer also.
June 30th, 2011 at 3:37 pm
@44
Can we please see it?
I started to toy with this thing last night at work. Before I knew it, I had been at it for a couple of hours. I gave up, shaking my head at how long I had been working on it... then came home from work and worked on it for another hour.
I am ready for the answer at this point.
June 30th, 2011 at 3:41 pm
Paul Konerko and Johnny Bench.
June 30th, 2011 at 3:41 pm
@45
Paul Konerko and Johnny Bench.
June 30th, 2011 at 3:55 pm
@46 and 45
Thanks. I looked at Johnny Bench closely because my first instinct when trying to find a reason for the WAR discrepancy was to think defense. I would have had the answer sometime around 4:00am if not for one little mistake in my reading of the original:
I read "Teammates" and not "played for the same team." So I was looking for two players who played for the same team at the same time.
Well... sanity is restored now.
June 30th, 2011 at 4:19 pm
Hmm, I found them too, but their WAR is 46.4 or 46.3 apart (depending on which pages you look on) and their 2B rate per 162 games is 1 apart.
In looking for the answer, you find sure some interesting stuff. Like two players with the same number of HR and very different WAR:
Cobb = Rod Barajas + 152
Wade Boggs = Ed Kranepool + 84.6
Joe Morgan = Gorman Thomas + 83.6
Roger Connor = Earl Williams + 81.4
Musial = Stargell + 70.3
And Rickey Henderson has 90.2 more WAR than Aramis Ramirez, despite having 1 less HR (as of yesterday).
June 30th, 2011 at 4:20 pm
@43, Whiz -- Nice finds! You are ... well ... eponymous!
June 30th, 2011 at 4:58 pm
@49, Whiz -- My favorite accidental discovery from this search was that Rogers Hornsby and Stan Musial both had 127.8. (Though perhaps I am the last to realize this?)
June 30th, 2011 at 4:58 pm
Hornsby & Musial had 127.8 WAR, that is.
June 30th, 2011 at 6:51 pm
There has been over 300 times where a player hit 40 or more Home Runs in a season. Only once did two of those seasons have identical Batting Average and HR and RBI totals.
...and it happened in consecutive years.
...and it happened by the same player.
In 1996 Vinny Castillo had 40 HR, 113 RBI and batted .304
In 1997 Vinny Castillo had 40 HR, 113 RBI and batted .304
June 30th, 2011 at 10:25 pm
@50
"You are ... well ... eponymous!"
JA, not to be confused with any Greek mythological figures.
June 30th, 2011 at 10:40 pm
@53
Castilla, not Castillo.
June 30th, 2011 at 11:09 pm
Onomateponymous: The mythical half-man, half-aardvark whose name was derived from the sound he made while snuffling out delectable box scores.
(It must be later than I think....)
July 1st, 2011 at 1:47 am
You guys got it! Sorry for whatever errors were in the original post and using a number that fluctuates on a daily basis AND appears differently. Nice work! Basically, I was intrigued by the guys who, at first glance, appeared statistically similar but had very different WARs or WPAs (originally, I tried to find guys with similar WARs and wildly different WPAs or vice versa, but that was too hard as those numbers generally link up). Konerko, playing 1B in a high offensive era, offered much less bang for his buck than Bench, playing C in a much lesser offensive era. So their similar HR and RBI and OPS+ numbers meant very different things. Fun stuff!
July 1st, 2011 at 12:50 pm
A little off topic, but while researching something else at baseballlibrary.com I learned that, in the second game of a doubleheader between Pittsburgh and Brooklyn on August 13, 1910, each team scored eight runs‚ had 13 hits‚ 38 at-bats‚ five strikeouts‚ three walks‚ one hit batter‚ one passed ball‚ 13 assists‚ 27 putouts‚ two errors‚ and used two pitchers. The game wound up being called by darkness.
July 1st, 2011 at 8:15 pm
@58, Kahuna -- I like to think that game was called not because of darkness, but to preserve that perfect symmetry for all posterity!
BTW ... what were you researching?
July 1st, 2011 at 8:22 pm
Late addendum to the collection of identical pitching lines:
Thursday in Wrigley, Matt Cain went 7 innings on 4 hits, no runs, 1 walk, 6 Ks -- the exact same line as his previous start. And each time, he left with a 1-0 lead, on an unearned run.
http://www.baseball-reference.com/boxes/SFN/SFN201106250.shtml
http://www.baseball-reference.com/boxes/CHN/CHN201106300.shtml
July 2nd, 2011 at 6:36 pm
For completeness, here are all the games I found (1919-present) with identical starting pitching lines, IP-H-R-ER-BB-SO. I only checked games with at least 5 IP before, and there were a couple of shorter ones.
1933-06-11(1), BOS@BRO, Fred Frankhouse vs. Sloppy Thurston, 4.1-6-4-4-1-0
1964-06-06, SFG@PHI, Jack Sanford vs. Dennis Bennett, 7-5-2-2-2-4
1968-04-20, ATL@CIN, Pat Jarvis vs. Gerry Arrigo, 7-6-1-1-1-4
1995-08-19, ATL@STL, John Smoltz vs. Donovan Osborne, 3-4-4-4-2-3
2011-06-28, CHA@COL, Gavin Floyd vs. Jason Hammel, 7-6-2-2-2-0
I also found a second game with the same IP-H-R-BB-SO (different ER), so there are two of those:
1961-05-24, WAS@KCA, Tom Sturdivant vs. Jim Archer, 7-6-4-3-2
1974-09-04, SFG@LAD, Mike Caldwell vs. Geoff Zahn, 6.2-7-1-3-4
Pretty rare!
July 2nd, 2011 at 11:05 pm
@61, Whiz -- Would you mind sharing how you performed these searches covering 6 pitching game stats, when the Play Index can only take 4 of those criteria at one time? Did you just find those that matched in 4 stats, and then eyeball them for the last 2 stats?
And did you really check every possible combination? I don't even want to think about how many there are ... but will you come and work with me? 🙂 (And if I ever do get a dollar out of this, you can have half!)
July 3rd, 2011 at 4:20 pm
I did 3-parameter searches, IP, H and R, and for games with two starting pitchers that matched I eyeballed the BB, SO and ER. So no special secrets, just grunge work. It was something to do while watching baseball games on TV 🙂
As for how many combos, there were 31 IP choices (0 to 10 in third-inning steps), and about 10 H and 10 R choices (for higher values of H and R the number of games peters out, at which point I quit looking), for roughly 3000 different combos. A lot of them had no pairs of pitchers to check, so they go fast (maybe a few seconds each). If there are matching lines with IP > 10, I wouldn't find them.
I also checked my Retrosheet play-by-play database as a check -- that found one that I had looked at but somehow missed the first time. Of course that doesn't help for games before 1950.
It was interesting that 5 of the 7 matching games had 6.2 or 7 IP. I wonder if the IP distribution for starting pitchers peaks there, at least lately.