Remember me

Introduction to Matching Aces

August 6, 2021


Matching Aces




         I had a question in "Hey, Bill" about a week ago, the essence of which was "are there fewer matchups between star pitchers than there used to be?"   Let me see if I can find the question and quote it correctly:

I'd be interested in knowing if some pitchers really did face off against the best of the best over and over again. That's the way I remember it being years ago, but maybe my memory is faulty.



         A quick answer is "Maybe", but there is a three-day trail of research behind that, and most of this article is about the trail.  But to move toward the answer as directly as I can.  ..


         The terms "Number 2" starter, "Number 3 starter", etc., came into common usage in the early 1980s.  I think they were derived from the term "Ace", which used to be more common than it is.  If the Ace is #1, then the second starter is "the number 2", etc.  

         Like most terms, these variants have multiple definitions.  The term "Number 2 starter" can be used either to mean a specific team’s second-best starter, or a pitcher of the quality that you would usually expect to be the second starter on a team.  Usually it means the latter.  Scouts and front office guys use the terms constantly when discussing players we don’t have.  A scout will be discussing a college player who has a so-so fastball but very good pitchability—you’ll hear the term "pitchability" a thousand times in draft prep—and the scout will say "I can see him as maybe developing into a #3 starter."  That’s not a compliment; that’s a way of saying he’s a 10th-round guy.  He’s NOT going to be a #1 starter or a #2 starter, but with some luck he might become a #3.  His future is limited. 

         The Red Sox started this season with a whole bunch of #3-#4 starters.  They have exceeded expectations, but at the start of the year everybody was saying "there is no #1 here, Eovaldi’s an OK #2, the rest of them are back-of-the-rotation guys."   If you say "he is OUR #2" or "he is THEIR #2", that refers to the specific team, but if you say "he is A #2" or "he is A #3", that’s assigning him a slot on a generic team, based on his usual expected performance. 

         So for this study I needed to designate pitchers as #1s, #2s, etc., so the first question I had to answer was "do I want to designate them based on their specific role on a team, or based on the generic standards?"   I decided on the latter.  I decided on that because (1) that’s the more common use of the term, and (2) the other approach leads to some ridiculous-seeming classifications.  In 1963 Don Drysdale pitched 315 innings with a 2.63 ERA and a strikeout/walk ratio of 251 to 57, but was nonetheless the #2 starter on the Dodgers.  Roger Craig pitched 236 innings with a 3.78 ERA, a strikeout/walk ratio of 108 to 56 and a won-lost record of 5-22, but was nonetheless the Met’s #1 starter.  It seems ridiculous to call Drysdale a #2 and Craig a #1, and it is better to avoid saying in numbers things that are objectively ridiculous. 

         That done, I assigned a number to every starting pitcher since 1921; actually, from 1921 to 2018, since the data file I used for this study, derived from Retrosheet, only goes to 2018, and I don’t actually know how to update it.  Anyway, every pitcher was classified as a #1, a #2, a #3, a #4, a #5 or a #6.  A fifth starter is a guy who bounces in and out of the rotation, pitches when you need him, you skip him when you don’t; a sixth starter is an emergency starter, more or less.   Of course, in modern baseball teams use five-man rotations, and the Red Sox have unaccountably managed to keep their five guys in rotation and on schedule through 100+ games, but that’s rare; ordinarily a fifth starter is in and out of the rotation, pitches a couple of bad games and you try somebody else. 

         Anyway, everybody from 1920 to 2018 was assigned a position in an order, in this way:

         The best starting pitchers in the league were assigned the number 1, up to a count equal to the number of teams in the league.  In other words, in an 8-team league (before expansion) there were eight #1 starters in the league.  In a 15-team league, the top 15 starting pitchers are #1s. 

         The next best group is assigned the #2, up to a count which is two larger than the number of #1s in the league.  In other words, in an 8-team league there are ten #2 starters.

         The next best pitchers are assigned the #3, up to a count which is two larger than the count of #2s.  In an 8-team league, there are 12 number threes.

         Etc.   So, in an 8-team league there are 8 #1s, 10 #2s, 12 #3s, 14 #4s, 16#5s, and then everybody worse than that is assigned the #6.  There are always #6s, from 1920 to 2018; they are always there. 


         To be assigned a high position means not only that you pitched WELL, but also that you pitched A LOT.   You pitch 20 games, 120 innings, no matter how well you pitch that doesn’t make you a #1, because that’s not what the term means.  Being a #1 pitcher means you’re The Man.   You anchor the rotation.  We’ll outline it better later in the series, but #1 starters made an average of 32 starts.

         I assigned the pitcher ranks by formula, Season Score, because I was dealing with several thousand pitcher/seasons and you can’t do them all by the eyeball test or the smell test, but then I noticed that Bob Gibson in 1967 was assigned as a #2.  Gibson was injured in ’67, a had broken bone in his foot or something like that, but he pitched great when he was on the mound and dominated the World Series, so it is kind of ridiculous to say he was a #2 starter.   As I mentioned before it is better to avoid saying things which are ridiculous, so I arbitrarily changed Gibson in 1967 to a #1 starter.  Then I noticed that the formula had Bob Feller in 1945, when he got out of the Navy late in the season, as a #3.   That’s ridiculous, too, so I changed that.  Then I got to wondering how many cases like that there are in history, where somebody who was an obvious #1 was wearing a different designation, so I undertook an organized search for those cases, and promoted players like that to their deserved designation. 

         It turned out that there were only 12 cases in history where it was necessary to change a designation.  Two of those were Gibson and Feller, and four were Ted Lyons at the end of his career, when he only pitched on Sundays but completed every start and had outstanding ERAs.  The other cases where I changed the designation were:

         Pete Alexander in 1924

         Whitey Ford in 1957

         Jimmie Key in 1988

         Dwight Gooden in 1989

         David Cone in 1996

Randy Johnson in 1996


There are lots of pitchers in history who pitched brilliantly for half a season but were hurt, but the rule was that, in order to be promoted back to #1 status, you had to come into the season as a #1, and pitch like a #1 while you were on the mound.  So, for example, Jim Kaat in 1972; Kaat made 15 starts and was 10-2 with a 2.06 ERA, so he pitched like a #1 in 1972, but if you look back, he was a #2 in 1971, a #2 in 1970, a #2 in 1969, a #2 in 1968, a #2 in 1967.  He doesn’t become a #1 in 1972 by pitching great for half a season.  If he had pitched great for the entire season, as he did in 1966, then he would have been a #1 starter. 

Nobody, and I mean nobody, is a number one starter for his entire career.  The designation comes and goes.  More on that tomorrow. 

Anyway, this enables me to address the question that AlbaNate posed:  are there fewer head-to-head matchups of aces than there used to be?

Well, actually what he asked is "If some pitchers really did face off against the best of the best over and over again?"  Obviously that is a lot different in a 15-team league than it is in an 8-team league, right?  We don’t need to study it to know that when you have more teams in the league, you see the same people over and over again less often.   What I took to be the intent of the question had to do with posting star pitchers against one another, like Marichal and Drysdale used to do, only the schedule is different. 


The answer is: the frequency with which #1 pitchers face off against other #1 pitchers has gone down a little bit because of the five-man rotations, rather than four-man rotations.  If you adjust for that, though, head to head matchups of #1 pitchers are not only not less common, they are a little bit more common than they were years ago. 

Think about it this way.  If every team used a true four-man rotation all season, then #1 starters would face off head to head in 6.25% of games, or one out of 16.  Team A uses their #1 starter one game in four; Team B uses their #1 starter one game in four.  They coincide one game in 16. 

When you go to a five-man rotation, this drops to 4%, or one game game in 25. 

The actual decline in the frequency of head-to-head matchups of #1 starters isn’t anything like that drastic, because, while teams used four-man rotations (mostly) until about 1975-1980, there weren’t actually very many pitchers who stayed in rotation the entire season.  Up until about 1960, there were many rainouts, and many scheduled double-headers to allow for travel.   Through the 1940s almost everybody both started and relieved; people would pitch 250 innings by making 30 starts and 20 relief appearances.  The schedule was relatively chaotic, compared to today, and this made it impossible for most teams to maintain a coherent rotation. 

About 1960 air travel replaced travel by train, and this reduced the need for scheduled off days.  Beginning in the mid-1960s, artificial turf swept baseball, its selling point being that you could vacuum it off and play after a hard rain, reducing rainouts.   About 1985 drain-through field technology began to replace artificial turf, but this had the same effect.   Rather than teams playing 154 games but playing them typically on 120 to 130 different days, with 25 to 35 double headers, teams began playing 162 games, but playing them on 155 to 160 different days, with few double headers. 

That left pitchers working on a more regular four-man rotation, and in that era you have dominant pitchers:  Koufax, Gibson, Seaver, Palmer, Catfish, Marichal, Bunning, Gaylord, Gibson, Ferguson Jenkins, Steve Carlton, etc. (although Gibson worked almost all of his career in a five-man rotation.)  Then we transitioned into a five-man rotation, so the net effect on the number of starts by a #1 pitcher was not all that large.  This is the percentage of games started by #1 starters, by decade:

1920s         23.0%

1930s         20.8%

1940s         21.3%

1950s         20.9%

1960s         21.5%

1970s         22.1%

1980s         20.3%

1990s         19.6%

2000s         19.2%

2010s         18.7%


As the percentage of games started by #1 starters has decreased a little bit, this has led to a small decline in the percentage of games in which a #1 faces another #1:


1920s         4.4%

1930s         4.0%

1940s         4.2%

1950s         4.1%

1960s         4.5%

1970s         4.7%

1980s         3.9%

1990s         3.8%

2000s         3.5%

2010s         3.6%


So yes, the frequency of head-to-head matchups between star pitchers HAS dropped over time, but not really.  It has dropped mathematically, but it would be difficult to say that it has dropped observably.  It’s 4%; it’s always been 4%. 

However, in regard to superstar pitchers, the highest quality of #1s,  there have in fact been some very notable changes in the ratio of head to head matchups.  

©2024 Be Jolly, Inc. All Rights Reserved.|Powered by Sports Info Solutions|Terms & Conditions|Privacy Policy