OK, I have arrived, finally, at what I really wanted to talk about here, which is matchups between #1 pitchers and other #1 pitchers. 

I designated every pitcher in the study as a #1, a #2, a #3, a #4, a #5 or a #6 starter; I said that a long time ago.  I’m just re-orienting you.  Then I figured the record, for each pitcher, when he was opposing a #1, a #2, a #3, etc.   This, for example, is Bob Gibson’s record against each opposing quality of starting pitcher:

There are a million things you can do with this data, and believe me, I’m not going to shut up until I have done most of them. 

Gibson had a losing record when facing #1 starting pitchers from the other team.  Don’t overreact to that; everybody does.  Not everybody; almost everybody.  Gibson’s winning percentage was .444 when he was going against a #1 starter, .538 when he was opposing a #2 starter, .630 against a #3, .645 against a #4, .720 against a #5, and .800 against a #6. 

Everybody’s record is like that.  It is unusual that it stairsteps up perfectly from #1 to #6, but in general, almost every pitcher’s record gets better as the quality of the opposing pitcher gets worse. 

The reason that very few pitchers have .500 or better records against #1 starting pitchers is that nobody is a #1 for all of his career.  Gibson was a #1 starting pitcher for 62% of his career starts, which is what makes him a Hall of Famer, but the people he was battling against on that line were #1 pitchers 100% of the time.  He’s at a disadvantage, 100 to 62.  Everybody is at a disadvantage, 100 to something, because nobody is a #1 starter for all of his career. 

So that’s not unusual.  As Gibson was stair-stepping up in winning percentage, he was stair-stepping down in the number of starts.  He faced #1 starters 113 times in his career, #2 starters 109 times, #3 starters, 89; #4 starters, 79; #5 starters, 54; and #6 starters, only 38 times. 

That is more unusual.  Gibson faced sixth starters only one-third as often as he faced #1 starters.  Randy Johnson, for a contrast example, faced #1 starters only 91 times, but faced #6 starters a whopping 142 times in his career. 

The Big Unit’s record is both better and worse than Gibson’s in this respect.  Johnson faced #6 starters far more often than Gibson did, but Johnson was one of the relatively few pitchers who was able to post a winning record (35-32) even when matched against a #1 starter.  Johnson’s winning percentage was much better than Gibson’s against ones and twos, much better, but was not nearly as good against fives and sixes.  I hope you’ll have a better understanding of that phenomenon by the end of the article. 

One of the million things you can do, which is actually very difficult to do by any other approach, is to figure the average quality of the opposing starting pitchers faced by each guy.  If you faced as many 1s and 6s, as many 2s as 5s, and as many 3s as 4s, the average would be 3.50.   The average, though, isn’t 3.50, because there aren’t exactly the same numbers of games started by pitchers at all levels.  The average is 3.19. 

Gibson’s average is 2.93, which. . . don’t know whether "low" or "high" is the appropriate term here.  He faced a lot of good pitchers.  He’s about the 93^rd percentile in terms of the average quality of the pitchers he faced.  Randy Johnson’s average is 3.59.  This is the bottom one percent.  It’s in the bottom half of the bottom one percent.  It’s in the bottom half of the bottom half of the bottom one percent.

I don’t want to overstate this, because this is unproven math which may prove, on further investigation, to be misleading..  But this is a true statement, not an editorial:  Randy Johnson faced a LOT of sixth starters in his career, and did not face off against #1s a proportional number of times.  The advantage that Johnson gained from this (as compared to Bob Gibson), according to my back-of-the envelope calculations, would be about 20 wins. 

In the world of Sabermetrics, 20 wins is a big number.  It’s two MVP seasons, we could say, or one-third of a Hall of Fame career.  I make that estimate in this way, which I will explain so that, if I have made a mistake, one of you can correct me, or if additional facts come in related to this, we can adjust. 

The difference between Bob Gibson’s quality of opposition and Randy Johnson’s is 0.66 "slots" in the rotation—2.93 as opposed to 3.59.  Gibson made 482 starts, Johnson made 603, so let’s say that we project that over 550 career starts.  To equalize Gibson and Johnson in their Quality of Opposing pitchers, you would have to move about 360 slots—550, times .66.  That’s actually 363; let’s say 360.

What’s the value of a slot?  The overall winning percentage of quality pitchers against #1 starters is .413.  Against #2 starters, it is .487; against #3 starters, .531; against #4 starters, .576; against #5 starters, .608, and against #6 starters, .688.  We will note that Bob Gibson beats the norm at each level of opposing starters. 

We can estimate, then, that the value of each "slot" in the rotation is about .055 wins per game.  360 slots times a slot value of .055 is 19.8 wins.  We thus estimate that Johnson gained about 20 wins, as opposed to Gibson, by facing more pitchers of lower quality, and fewer top-of-the-rotation pitchers. 

Well, that was an interesting digression.  Alerting you again that additional research may spot some gross flaw in my approach. Let me generalize about WHY Gibson faced tougher opposing pitchers, as best I understand that. Two reasons.  First, Gibson’s data is relatively typical of Bob Gibson’s era, which was the era of stable four-man rotations.  Repeating myself, it was difficult to run a steady four-man rotation before 1960, because of (a) travel days and (b) rainouts, leading to (c) double-headers.   After about 1980, starting rotations switched from four men to five men, and also, I think, teams became more aggressive about calling up a sixth starter for an emergency start.

There was an interim period there where significant numbers of pitchers actually made 37 to 41 starts a year, and Gibson’s ratios of opposing starters are actually pretty common in his era.  Gibson’s quality of opposing starting pitchers is the sixth-highest ever for a Hall of Famer, behind Tom Seaver, Burleigh Grimes, Carl Hubbell, Juan Marichal and Whitey Ford.  Gaylord Perry and Catfish Hunter are close behind him. 

Randy Johnson faced a sixth starter more times than anyone else in history, but behind him on that list is an unbroken list of his contemporaries and Hall of Fame contenders:  Bartolo Colon, Roger Clemens, CC Sabathia, Greg Maddux, Andy Pettitte, Jamie Moyer, Kenny Rogers, Tom Glavine, Tim Wakefield.  I suspect that the cause of this is:  that we don’t have pennant races anymore. 

I couldn’t talk about this on Twitter without getting 200 "Get off my lawn" responses from the Mole People, but modern fans, coming to the game after 1990, have NO idea what a pennant race was. In a pennant race, you had to win every game you possibly could. 

In 2002, you could throw a sixth starter out on the mound and say, "Hey, kid, this is a no-pressure situation, you understand?  Nobody is expecting you to beat Randy Johnson.  Just relax and do the best you can." 

With five teams in a division, you typically have two competitive teams at the top of the division—and if you don’t win the division, you can still make post-season with the Wild Card.  Wild Card teams won the World Series in 2002, 2003 and 2004.  Not winning the division wasn’t that big a deal, if you had a really good team. I was working for the Red sox in those years; we didn’t worry about winning the division. We always believed that we would make the playoffs, even though sometimes we didn’t.

The 1960s, with the ten-team leagues, had more fantastic pennant races than any other decade, easily; again, that’s not a memory or an opinion, that’s an objective fact.  Not every year or every league, but in a lot of leagues, you had crowds of teams separated by a couple of games.  In the 1962 National League, three teams ended the regular season between 98 and 101 wins.  The National League in 1964 had teams with 93, 92, 92 and 90 wins.  And 88.  The American League the same year had teams that ended with 99, 98 and 97 wins.  The National League in 1965 had 97 and 95.  The National League in 1966 had 95, 93 and 92.  The American League in 1967 had 92, 91, and 91, and 89 just in case. 

There was no SPACE in there.  The pennant race contenders were packed together, and if you finished one game short, you were DEAD. Now, most divisions you’ve got two teams between 88 wins and 100, and if you finish second you’ve still got the Wild Card.  It’s not remotely the same. In Randy Johnson’s time you could back off and play for tomorrow a little bit.  In Bob Gibson’s time, you didn’t throw a #6 starter against Bob Gibson unless the wolf was at the dugout door, because that one game had a real good chance of costing you the pennant.  There was no breathing space.