And Third Basemen

And Shortstops

            The process for second basemen, third basemen and shortstops is essentially simple, and is the same process for players at all three positions.   The players at all three positions are credited with Runs Saved based on three things:  Double Plays, Range, and Fielding Percentage.  

            The process for all three is so simple (now that we have explained its components, several times) that I am actually unsure whether I need to write it out in a string of formulas or not.  I am trying to explain EVERYTHING, explain the whole system in one continuous memo so that a programmer could take it and run the numbers, but I already have explained all of this.   There’s nothing new about it.   I’m not sure whether I would be helping the process by explaining it again, or just annoying people. 

            For Double Plays, 2B-3B-SS get credit for Runs Saved based on:

            The number of double plays that they participate in,

            Divided by the total double plays of all of the team’s pitchers, second basemen, third basemen and shortstops (which was Formula 17 in this series),

            Times the Team’s Runs Saved by Double Plays (which was Formula 16),

            Times .70. 

            For Range, 2B-3B-SS get credit for Runs Saved on the team level based on:

            Their Enhanced Range Number (which was explained in the Essay "What Is Range", and defined again in Formulas 32, 33 and 34, for second basemen, third basemen and shortstops.)

            Divided by the Team Total of Team Adjusted Range Outs (which was Formula 36),

            Times the Team’s Runs Saved by Range (which was Formula 38),

            Times .90.  

            Then the TEAM credit for Range is transferred to the individual fielders at the position based on the formula  PO + 2*Ast + DP – 7*Err, as was done for first basemen (Formula 56). 

            For Fielding Percentage, 2B-3B-SS get credit for

            Errors Not Made above the Baseline,

            Divided by the Sum of Errors Not Made for all players on the team,

            Times the number of Runs Saved by the Team by Error Avoidance. 

            This process was explained in Formulas 20 through 25, and was repeated for catchers as Formula 53, and was repeated for First Basemen as Formula 58.   It is exactly the same process at every fielding position.   I hope that is sufficient explanation.

SECOND BASEMEN

            We’ll start our reports on the Runs Saved by Second Basemen with the Leaderboards.  These are the leaders in Runs Saved in the "15 teams" part of the study:

And here are the leaders from the 1961 portion of our study:

OK, Bill Mazeroski was a really good second baseman; that may not be news, let’s move on.  When we are dealing with only 15 teams in one sample and only 18 in the other, players will make the Top 10 based on playing time alone; when we move on to the +/- portion of this report, three of the top 10 in each of the two groups I have given you will show up as below-average defensive players.  Why don’t you look over the lists, and see who you think the below-average guys will be? 

Mazeroski each season was credited with saving 26 runs on Double Plays and 19 runs by Range.   In 1960 he was charged with only 10 errors, so he was credited with another 17 Runs Saved by Error Avoidance.  In 1961, however, he made 23 errors, so his Runs Saved by Error Avoidance was down 30%.  This is the top 10 combined list, with a layout of where the Runs Saved were credited.  One of these Top 10 is still below average on a per-inning basis; why don’t you guess which one:

The +/- approach is fairly legitimate just for the 1961 data, because it’s all the same year. . . .two leagues, but at least the same year.   And Mazeroski is actually not the +/- leader for 1961.  It’s actually Frank Bolling:

Mazeroski played 26 more innings than Bolling and is credited with saving 0.5 more runs, so Bolling is +13, Mazeroski +12 in 1961.  Bolling and Mazeroski were kind of the same guy.  They were the same size, both right-handed hitters, basically the same hitter. . .six points difference in batting average, one point in slugging average, Bolling was 85 OPS+, Mazeroski 84.   Both very good second basemen; Bolling had won the Gold Glove in the American League in 1958. 

   Second basemen within our data—which is very preliminary, very temporary data covering only 33 teams—but second basemen are credited with saving .0336 runs per inning, or 0.302 Runs Per Nine Innings.   Somebody had asked that I present those numbers on a per-9-innings basis, because that is more familiar to us.  But you have to remember:  all of these runs are being saved in the same innings.  Based on the data we have seen so far—which is very preliminary—it looks like Catchers are saving about .40 runs per nine innings, first basemen about .20, Second Basemen .30, Third Basemen .20, and shortstops .35.   I’ll have to adjust the first basemen and third basemen to move the third basemen up an inch and the first basemen down an inch.  I don’t know how instructive it would be to show you a list with one second baseman at .31 and another one at .32 and another one at .30.   But remember, all of those things are happening at the same time.  There are about 4.50 Runs Per 9 innings being prevented—the same number as are being scored.  That’s the operating assumption of the system; offense equals defense, so Runs Prevented equal Runs Scored, not in every game, not for every team, perhaps not in every season, but overall.   So far we have accounted for about 1.45 Runs Saved per game among catchers, first basemen, third basemen, shortstops and second base weenies. 

The "Lasters" in the 1961 group, other than Chuck Hiller, were all old guys, 1950s players near the end of their careers.  Hiller was nicknamed "Iron Hands", so that’s not controversial.   The only one that could be controversial is Nellie Fox, who had won the Gold Glove in 1957, 1959 and 1960, but Nellie was 33 years old, and his numbers in 1961 crashed across the board. . . .38 points down in batting average, doubles and triples dropped from 24 and 10 to 11 and 5, Runs Scored dropped from 85 to 67 in the same number of at bats and plate appearances, double plays dropped from 126 to 97 and Assists dropped from 447 to 407. 

The three players who made the leaderboard in the first group although they were below average per inning were Jose Altuve, Mark Lewis and Roberto Alomar.  The three from the 1961 Group were Nellie Fox, Jerry Lumpe and Jake Wood.  The one player who made the combined-group leader board in Runs Saved although he was actually below average in Runs Saved per inning was Jake Wood.  Wood played every inning of the season at second base for the Detroit Tigers; I believe he was the only major league player in 1961 to play every inning of the season.   There was a lot of excitement about him before the 1961 season started.  He was very fast, had a little pop in his bat, and he had hit .300 in the minor leagues for four straight seasons.   The Tigers traded Frank Bolling to Milwaukee in the off-season to give the second base job to Wood—an unusual move, because Bolling was a really good player.  Wood as a rookie stole 30 bases, hit 11 homers, and led the majors in triples, with 14, but also led the majors in strikeouts, with 141.  He hit .258, he didn’t walk, and he wasn’t a great second baseman.  He seemed to be coming into his own as a hitter in 1963, but a broken finger ended his season in July, and his career was disappointing.   

The +/- system loses a certain amount of integrity when you apply it to players playing more than half a century apart, and assume that the same "center" applies to all of them.   But what the hell; let’s give it a shot. 

Knoblauch’s defensive problems are well known, and I wouldn’t expect anyone to argue with me about Mark Lewis or Scott Hairston.   Altuve in 2012 was basically a rookie, and the Astros lost 107 games.  Ben Zobrist is one of my favorite players ever—as Altuve is—but Zobrist in 2016 was 35 years old and dividing his time between the infield and the outfield, so. . . well, you can argue it if you want to.  This system does return lower numbers for fielders on high-strikeout teams, and the 2016 Cubs are easily the highest-strikeout team in the study.   At some point, further down the road, we can adjust the defensive innings for the strikeouts to get more accurate player-to-player comparisons.  See comment in Monday’s article about Clete Boyer.

We show Bolling and Mazeroski in 1961 as being +25 runs between them and as saving a total of 113 runs between them.  That would mean, logically, that average second basemen in their shoes would have saved 88 runs, which would mean that Bolling/Mazeroski (Bolleroski) are 28, 29% better than average.  That seems like a reasonable number.   You figure the best hitter at a position might be 25-35% better than the average hitter at a position in a season.  Doesn’t seem like a crazy number.   Anyway, that’s what bothers me about the traditional +/- form of fielding numbers; we say that a given fielder prevented 15 runs more than an average fielder at his position, but then when we ask "How many runs did the average fielder prevent?", the answer is "we don’t have any idea.   We don’t know anything about that."  How is that even possible?  If you don’t know how many runs the average fielder prevented, how can you know that this fielder prevented 15 more than that number, which you don’t know what it is?   Doesn’t that bother anybody but me? 

There are six Gold Glove Second Basemen who happened to wind up in our study.  All of them except one are seen by this analysis as being above-average second basemen.  The one Gold Glove season which is seen by this analysis as being below average was by Roberto Alomar in 1992:

The six Gold Glove second basemen in our study saved 294 runs among them, which is 14% better than average.