Analyze Pitchers

CLASSIC FINESSE PITCHERS

The 1960s and 1970s featured some amazing pitchers with excellent control.  These Cy Young Award winning pitchers all had ERAs below 2.00 and held batters to batting averages around .200 in the years selected for this analysis. These pitchers were also work horses each facing over 1000 batters in a season.

We’ll take these in chronological order starting in 1963 with Sandy Koufax.  Koufax held opposing batters to a batting average of .189 and achieved an impressive ERA of 1.88.  For reference, his OPS was at .501 with an excellent WAR of 10.7.  His BMP was .443 with 69% BBM, 29.4% BRBM, and 1.6% RBM distribution.  As one might expect, unlike batters, a high BBM distribution and a low BRBM distribution is desirable for a pitcher.  This indicates that most of the opposing hitter’s damage is being done without runners on base. 

We should probably set the BMP bar for an epic pitcher performance at .500.  With Koufax setting the pace at .443, let’s examine the amazing Cy Young winning performances for Bob Gibson and Denny McLain in 1968. McLain compiled and amazing 31-6 record facing nearly 1300 batters.  His ERA of 1.96 and opposing batting average of .200 were top notch.  Gibson, in the same year, compiled a 22-9 record with an unbelievable ERA of 1.12 and an opposing average of .184.  OPS and WAR have Gibson rated much better than McLain and BMP analysis agrees in this case.  Gibson achieved a BMP of .405 vs McLain with a BMP of .488. 

As with batters, the BRBM is typically the most important factor in achieving a good BMP.  In this case, both pitchers had excellent BRBM enabling low BMPs.  In this case they both had identical BRBMc (.129) indicating that both performed equally well against batters with runners on base. However, Gibson had a ridiculously low BBMc of .278 and a negative RBMc which made the difference between these pitchers. Gibson’s low BBMc is directly attributable to low SLG and OBP (especially with no runners on base) and his negative RBMc indicates that more runners were caught stealing than successfully stole bases.

Tom Seaver and Steve Carlton also battled for Cy Young Awards in the early 1970s. They both achieved sub .500 BMPs with very similar performances but did not approach Gibson’s mark.

If we rank these five pitchers, we can see that BMP and OPS are fairly well aligned this time.  Remember that OPS is somewhat representative of BBMc so this is not surprising since the BBM distributions are high. What is surprising is that Steve Carlton, with the third highest BMP, had the highest WAR with 12.1.   We need to remember that OPS and BMP only measure the opposing batter’s performance.  WAR takes into account other factors such as the average performance of other pitchers in the same year.   So, WAR may legitimately rank Steve Carlton higher than Bob Gibson but that is not apparent from the standard statistics.

MODERN FINESSE PITCHERS

Some of the top pitching performances in recent times are shown below.  Again, we see world class ERA and opposing batting averages for all these pitchers. We also notice that the number of batters faced is less than earlier years as teams became more dependent on relief pitchers. Pedro Martinez had a stellar year in 2000 which is recognized as one of the all-time best pitching performances if you google it. With a BMP of .406, he narrowly loses out to Bob Gibson’s 1968 performance by 1 point.  Let’s try to understand the reason why. Martinez had a much lower BRBM distribution than Gibson which gives Martinez a 29 point BRBMc advantage (.100 vs .129).  From what we have seen so far, the BRBM usually makes the difference. But not this time.  Martinez’s BBMc component was 22 points higher (.300 vs .278) indicating that Martinez essentially gave up more extra base hits with no one on base.  This nearly offset the BRBMc advantage.  However, it’s the RBMc that makes the difference with Gibson (-.003) beating Martinez (+.006) by 9 points.  In this comparison, handling base stealers was the key factor.

Next, we analyze the 1995 performance of Greg Maddux.  His ERA and opposing batting average are excellent, but they don’t stand out against Martinez and Gibson.  His OPS and WAR are not as good as Martinez and Gibson.  However, unexpectedly, he ends of with the lowest BMP at .401.  This one is not obvious and demonstrates the value of BMP analysis.  The reason is that he simply didn’t allow much base movement with or without men on base.  He quietly kept his BBMc down to .281 and his BRBMc down to .101.  These two components together set the baseline for an amazing BMP.  Unfortunately, Maddux had a poor performance against base stealers with a 4.8% RBM% yielding a very high RBMc of .019.  But even this poor RBMc did not knock him out of first place.  This case is a very good example of how standard statistics don’t capture the true base movement production that BMP does.

An excellent 1985 season for Dwight Gooden yielded an impressive BMP of .431.  Clayton Kershaw, in 2014, also came close to the .400 barrier with a .430 BMP.  Although his BBMc and BRBMc are nearly identical to Gooden, Kershaw got the lower BMP thanks to a -.009 RBM. 

In this group we see more variation between OPS, WAR, and BMP.  BMP has Maddox first, OPS has Martinez first, and WAR has Gooden first.  We’ve already dismissed OPS as less-than-valid statistic, but the WAR numbers are interesting.  Even though Gooden had the worst OPS and BMP, his WAR of 12.2 is very high. Based on pitcher’s WAR formula, this is most likely due to weak competition among pitchers that season. This highlights that WAR can be relative to a season.

POWER PITCHERS

This analysis would not be complete without taking a look at some fireballers.  The power pitchers in this section feature a high number of strikeouts which, one might think, could lead to some very good BMP numbers.  Although these pitchers have excellent ERA and opposing batter averages, only Roger Clemens has a BMP below the .500 mark.  Despite the strikeouts, these pitchers gave up a few more extra base hits and a few more walks than the finesse pitchers leading to BMPs nearly 100 points higher.  Clemens leads this group with a .489 BMP.  Clemens and Johnson had nearly identical BMP distribution with high BBM% and low BRBM% which naturally gave them the edge over Ryan and Feller.  They both had impressive negative RBM%.  Despite equal distribution, Clemens simply had lower magnitudes of BBMc and BRBMc to win this battle.

We can see that the best power pitching performances, although impressive, did not come close to challenging the finesse pitching performances previously reviewed.

OLD TIME PITCHERS

Data for old time pitchers is not available on Retrosheet before 1911 so we don’t have information from many pitchers in that era.  This was called the dead ball era where pitchers dominated until rule changes in 1920.  Standout pitchers from the 2010s are Walter Johnson, Grover Alexander, and Christy Mathewson. These pitchers have standard statistics (ERA, Avg, OPS) on par with the best finesse pitchers and generally result in a similar range of BMP.

However, Walter Johnson’s performance in 1913 stands out with a BMP of .368 beating all pitchers analyzed so far.  It can be seen that his BBM distribution of 76.4% and BRBM distribution of 27% is the primary reason.  Although his BBMc (.281) and his BRBMc (.099) is similar to Maddux, his insane RBMc of negative .013 creates a 32 point gap with Maddux.  In this case, Johnson’s control of base runners was the deciding factor.  Grover Alexander comes in with a BMP of exactly .400.  Although he matched Walter Johnson with BBMc and RBMc, his higher BRBMc resulted in a higher overall BMP.

RELIEF PITCHERS

Relief pitchers are generally not compared to starting pitchers since they pitch less innings per game and they pitch more often.  Ultimately, they face significantly less batters in a season. They often inherit runners on base who are considered the responsibility of the previous pitcher.  With BMP we can certainly compare relief pitchers against each other, but we can also compare them to starting pitchers.  With BMP is doesn’t matter who put the runners on base, the pitcher is responsible for the base movement of batters and the runners.  With ERA, the relievers get off the hook for allowing previous runners to score.  With BMP, the relievers do NOT get off the hook for previous runners advancing or scoring.

Let’s start with the all-time saves leader, the great Mariano Rivera.  Although not a memorable year, 2008 was probably his best performance.   His ERA was only 1.40 and his opposing batting average was an amazing .165. BMP distribution was favorable with a low BRBM% although he gave up a lot of stolen bases with 5.1% RBM distribution.  Despite having a high RBMc(.019), he had an excellent BRBMc (.108) and an unbelievable BBMc (.255).   As a result, Rivera ends up with a sub-400 BMP of .382.  His ability to limit batters with no one on base resulted in a ridiculously low BBMc and was the determining factor.

Trevor Hoffman is another world class reliever who had a great year in 1998.  On the surface, his MLB statistics were very similar to Rivera although OBP was a bit higher.  Hoffman’s BMP distribution was not as efficient as Rivera with a higher BRBM% and a low BBM%.  Although Hoffman ends up with an impressive BMP of .489, he loses out to Rivera on all three BMP components.    

Dennis Eckersley comes closest to Rivera with a BMP of .432 in 1989.  MLB statistics would indicate this might be a close race, but a much higher BMBMc component for Eckersley widens the BMP gap in favor of Rivera. 

Rollie Fingers, despite an incredible ERA of 1.04, failed to eclipse Rivera, Hoffman, and Eckersley and ended up with a BMP of .475.  This is a good example of a relief pitcher allowing base movement but not getting penalized for the runs scored that were attributed to the previous pitcher. 

BMP SUMMARY - PITCHERS

Of the pitchers examined, it can be seen that all categories can produce excellent BMP performances.  Most pitchers analyzed beat the .500 mark with a couple stellar performances coming in below .400.