In Part I, I suggested the “How” and “Why” as to what organizations might be missing. In Part II, I will give you the “What” – the specific findings that underpin an unprecedented opportunity for an organization to capture a significant competitive advantage.

At the major league level, approximately 20-30% of players suffer from significant swing path issues. Since path issues are likely the single largest factor in player failure and underperformance, an organization with a systematic approach to “cure path” would be at significant advantage relative to the remaining teams. Not only would the club benefit directly through improved offensive production, but having an effective cure to path would provide superior insight into “true talent” as the largest remaining factor.  There are also logical extensions into how this could be further monetized in potential areas such as player arbitrage, draft selection, etc.

In summary, the comprehensive solution can be simply stated as follows:

1) The optimal swing paths that exist in the muscle memory of the best hitters can be quantified and visually represented to players performing below potential.

2) A systematic process can be built around the above for significantly improved performance.

Before getting into the key findings, I would like to talk about process using another parallel to investing. There is a popular view that all hitters are different and a “systematic process” cannot be utilized for fixing the ones performing below potential. Successful investors are also presented with a significant amount of uniqueness in their decision process. They have all developed a process that effectively deals with uniqueness, not avoid it all together. So if a process isn’t effective in getting hitters to potential, it’s not that you need to abandon process, you just need a different one. Swing path is a core mechanic that can be systematized and as you will see, there is plenty of room for customization within a systematic process.

The findings below are presented with an extremely high level of conviction based on several years of research including a patent filing in 2013. History will be the ultimate judge but based on communication with a handful of organizations, there is a strong possibility that many clubs will be unwilling to consider such non-traditional sources of value.

The Findings – Quantifying the Optimal Swing Path

Variables:
1) X Axis – Swing Loft
2) Y Axis – Bat Angle (vertical)
3) Z Axis – Swing Direction (Very small changes – Not considered here)
4) Timing (technically, horizontal Bat Angle but we’ll just call it timing for simplicity)
5) Ball Contact Point (Relative to ball equator – Not considered here in terms of loft)

Constraint – The Bat Angle for any given pitch height represents a straight line from the ball to the chest area such that the intersection between the body and the swing plane is a point in the chest area not the mid-section or waist.

Nothing major yet – just some visual representations of the Variables and the Constraint.

Although not a major finding, many are still of the opinion that the bat should be relatively level resulting in a much lower swing plane/body intersection as in the illustration above. The importance of Bat Angle will become more clear shortly. For now, I’ll use two extreme Infield Fly Ball Rates (IFFB%) as a general proxy for a path quality. Below, you will see the Bat Angle (and plane/body intersection) for one of the lowest IFFB rates (Joey Votto) and one of the highest (Kevin Kiermaier). Note the significant 10° difference between the Bat Angles for the same height pitch.

Looking more broadly, the average Bat Angle on low-middle pitches for the players with the five lowest IFFB rates (2015 and 2016) was 34° while the average Bat Angle on low middle pitches for the players with the five highest IFFB rates  was 26°. An example application is screening for players with high IFFB rates on low pitches. Since a relatively flat bat is required for an infield fly ball, this type of screen can highlight players with consistently insufficient Bat Angle.  However, this is only a small part of a more comprehensive approach to identifying players with poor paths as discussed below.

Timing is a Separate Loft Factor

To illustrate Timing Loft, consider the model below in which Swing Loft has been set to zero, thus 100% of the loft is a result of Timing Loft (Bat Angle set to theoretical maximum to illustrate the point).

Conversely, for a high pitch (Bat Angle set to theoretical minimum), 100% of the loft comes from Swing Loft while Timing determines the direction of the hit – pull, center, or oppo (below).

Loft Goals, Angle Mix, and Loft Contribution

Given the principles previously illustrated, optimal angle combinations can be constructed for each pitch location. Each location will have different loft contributions from Timing and Swing Loft based on the height of the pitch. So ball height determines Bat Angle which determines the mix of loft contribution. We will discuss customization shortly. For now, let’s just assume a launch angle goal of 15 degrees for a particular player.

*Note – A middle to low pitch was used to illustrate a 50/50 mix of loft contribution which occurs at a 45° Bat Angle. A true middle pitch has approximately 30-35° of Bat Angle

Thus, the optimal Swing Loft for a low-middle-height pitch for a player with a 15 degree loft goal is only 7.5 degrees – the other 7.5 degrees comes from timing

AND Swing Loft is not the major loft factor for low pitches – which is a lot of them.

Extending these concepts, the optimal loft goal and loft contribution mix can and should be adjusted for different hitter types. Pull hitters will have a greater relative contribution from Timing Loft while opposite field hitters will have a greater relative contribution from Swing Loft. While adjusting the loft goal higher for more powerful hitters is an option, there are potential drawbacks that should be carefully considered and are discussed below.

Reconciling the Swing Up / Swing Down Views

It is interesting to consider the above in light of baseball ignoring Ted Williams for so many years. Usually, there is some rationale for significant movement in a particular direction. In hindsight, it is not too difficult to consider that the “swing down” movement came about (in part) because the bat path / ball path matching issue can’t be solved with a simple one-factor (i.e. Swing Loft) solution as Williams proposed (specifically, page 67 of The Science of Hitting where he shows front knee bend to hit low pitches). While Swing Loft should never be negative, one can certainly appreciate how minimal Swing Loft and relying on Timing Loft for low pitches might feel down to a player. A few takeaways:

1) I am convinced that the failure of Williams’ teachings to fully catch on as well as the highly variable success rates today in adding loft is because some players are able to intuitively arrive at the correct mix of loft contribution while others are forcing too much Swing Loft.

2) Since “timing adjustments” are clearly required to achieve high levels of loft using an optimal mix of angles, excessive loft goals are problematic for many hitters, particularly those who are not natural “pull” (early timing) hitters.

3) Given the significant number of consistent hitters in the 13-15 average LA range, this indicates a Swing Loft on average, of 6.5 to 7.5 degrees for a low-middle-height pitch – nowhere close to what many believe is required to become a card carrying member of the “Fly Ball Revolution.”

Visually Representing Optimal Swing Paths to Hitters

I am by no means an expert in neuroscience; however, it seems relatively straightforward that the more information hitters can transfer out of conscious thought into subconscious/muscle memory, the better they will perform. Hitters don’t want to (and shouldn’t) think about complex angle combinations. Consequently, a visual representation of the optimal combination of angles tailor made for each considering their power (i.e. to determine goal loft) and pull vs oppo tendencies can quickly correct a consistently poor swing path. Yes, “Keep it Simple” but solve the complexity first.

Below is the device that allows a hitter to train for optimal angle mixes through seeing and feeling the optimal paths for different pitch locations. The ball joint allows flexibility for any mix of angles while the angle guide provides compound angle settings based on a hitter’s customized loft goal and pull/oppo preference.

I will keep the product plug to a minimum. Additional information may be found here.

Finding Great Paths in the Data

Statcast data confirms that the best (most consistent) hitters hit the ball significantly flatter (in terms of bat/ball contact, not launch angle) than average. You can read more about the details of this here. I refer to estimated spin impact as Mean Unexpected Distance (MUD).

In addition to hitting the ball flat, one of the best indicators of a great swing path is low variability (Standard Deviation) of a player’s launch angles. After witnessing significant reduction in launch-angle variability through focused training, I had a significantly high level of conviction that this was a key indicator of “quality of path” several years ago. The availability of Statcast data through Baseball Savant changed everything. The data not only confirmed the benefits of flat contact previously discussed but also proved that data combined with video analysis can assess  “quality of path” with a very high degree of accuracy. Considering no other data than (low) MUD scores and (low) Standard Deviation,  the following hitters are returned (based on 2015 and 2016 data):

The average wRC+ and BABIP for the hitters above are 129 and .330, respectively. Given the return of Cabrera, Votto, Mauer, Trout, Freeman, J.D. Martinez, and several others considering no other performance factors, the benefits of low Standard Deviation of LA and flat contact seem relatively clear.

Putting It All Together – A Better Training Approach

I believe it would be a mistake to ignore the iterative training process that the best hitters have utilized up to this point. Hit a lot of balls, keep what works, discard what doesn’t and repeat the process over a very long period of time. In many cases, the only thing differentiating good from bad paths is that a player’s “filter” allowed something to get through that should have been discarded.

The drawback of the iterative approach is that it takes a very long time. If a player gets off track, it can take a frustratingly long time to go through the process to fix something that may have a very simple solution. Combining what we know from the discussion above, the variability of launch angles can be used in training sessions to quickly determine if a path is moving in the right direction or not before a player fully commits to a contemplated change.

In other words, path issues can be effectively addressed from opposite directions – static/device training  to improve path and dynamic (pitched balls)  training focused on reducing LA volatility. In training sessions with BP-type pitching, players with good paths are able to get down to a 13-14 degree standard deviation.

Implications From The Findings

Part I  pointed out the first step in the research process (as I have known it) is “Identify the Key Drivers”. I believe it is fairly safe to say that most, if not all organizations, believe that path is a key factor. This leads to the logical question of – Why didn’t MLB organizations “go deep” on one of the largest factors of player performance?  I believe there are two primary reasons:

1)  They assumed there was nothing of significance to be found – they concluded before they considered.

2)  It was outside the scope of responsibilities for employees on both the data/analytics side and the player development side of the organization.

Looking more broadly, it would be my guess that most organizations would say they do not believe significant “Moneyball-size opportunities” exist. It is this type of thinking that suggests that they likely do. I’m currently looking into another key driver of performance that could possibly be addressed through a similar systematic approach. Until baseball organizations change their thinking (and possibly their organizational structure), it’s likely that these opportunities will continue to exist. The source is the same – the “Gap in the Middle” that was outlined in Part I.

Clearly, all of the findings presented here are “known” in the muscle memory of the best hitters with great paths.  To this point, however, this muscle memory knowledge had not been understood or quantified in a way that could be systematically transferred to other players. By separating the loft factors, quantifying optimal paths for each location, and presenting a simplified visual representation of the optimal combination of angles, hitters can correct path issues with a very high rate of success.

Going forward, it will be interesting to see how organizations change in regard to considering non-traditional sources of value such as the “Gap in the Middle” previously discussed. At least one, the Houston Astros, announced in March (two days after Part I but likely just a coincidence) that they were moving their lead analyst, Sig Mejdal, to get “on-field experience.” This move, combined with his title of “Process Improvement,” indicates that they might be ahead of the pack in terms of considering new ideas. If you are aware of other clubs moving in this direction, please indicate in the comments. Based on my communication with a few organizations, I believe several are going to be late to the party as they appear unwilling to challenge existing assumptions.

Given the possibility of more “meat on the bone” for the findings above, I will likely take another short break before publishing the next article. However, for those interested in considering opportunities where data and mechanics intersect such as what has been presented above, there is considerably more material for your future consumption.

Aaron Judge is good.  Some might say he is great.  The front-runner for AL Rookie of the Year and MVP is the face of MLB for 2017, but the face of MLB for the future?  Unfortunately, maybe not.

It’s hard to find something negative to say about the New York Yankees right fielder, but in order to play devil’s advocate and not get our hopes up too high about Aaron Judge, just in the event that he has a down season, I was able to find some rather unflattering comps for the slugger.

First, there’s his minor-league career.  Aaron Judge was a pretty good prospect ranking first in the Yankees’ system in 2015 and 17th in baseball according to MLB Pipeline.  However, just because a prospect is ranked highly does not mean they are without flaws.  Judge would strike out in at least 21 percent of his plate appearances in all levels in the minor leagues.  This article from 2016 even identified Judge’s proficiency to strikeout:  

Judge’s Triple-A debut at the end of 2015 did not go well. He slashed .224/.308/.373, well below both his career levels and expectations. More alarming, he struck out a career high 28.5-percent of the time (74 times in 260 plate appearances). [The 2016 season] has been more of the same. His batting average is a bit deceiving sitting at .284 (heading into this weekend), considering he currently has a nice .354 BABIP compared to last seasons .289. His plate discipline is troubling.

Perhaps the lofty expectations of Judge have him pressing. You simply can’t overlook the fact that his strikeout rate is nearly identical to the small sample size of last season’s Triple-A numbers (27.2-percent). It has to be at least a slight bit worrisome that this is a trend and not a slump. His walk right is dropping daily to a new career low (6.8-percent or eight walks in 103 plate appearances).

The article seems to point to his plate discipline as his main flaw — as other evaluators have — but is overall positive with his prospect status.  But his strikeout tendency should not be overlooked.  He has failed to improve on that statistic in his short major-league career, where he has struck out in 32 percent of his plate appearances between his call-up in 2016 and now.  However, because he also takes his walks, his walk percentage is rather high, which puts him in exclusive company.

Since 2000, there have only been four players with at least 300 plate appearances who have struck out in over 29 percent of their plate appearances and walked in at least 16 percent of them: Jack Cust (2007, 2008, 2010, 2011), Ryan Howard (2007), Adam Dunn (2012), and Aaron Judge (2017).  All of these seasons resulted in wRC+ well above 100, which means that they were productive players; however, these player were known to be the embodiment of the “three-true-outcome” hitters.  Dunn had five consecutive seasons of 40 or more home runs, but also led the league in strikeouts four times; Cust led the league in walks once and strikeouts three times; and Howard led the league in home runs twice and strikeouts twice.  Admittedly, these comps are not encouraging.  Although these players were not horrible in the simplest definition, their careers were short-lived and their production sharply declined.  For Cust and Dunn, it forced an early retirement, and Howard a well-publicized and sad end to an illustrious career.

But it’s not just Aaron Judge’s strikeout and walk percentage — it’s also his raw strikeout numbers.  Judge is on pace to strike out over 200 times this season.  While it’s already been established that he is strikeout-prone, it does not serve him justice that the 200-strikeout threshold is upon him.  No player who has struck out 200 or more times in a season has had a very high average.  As the legendary Pete Rose noted, the highest single-season average for a player with 200 or more strikeouts was .262 (Chris Davis holds that honor).  The short list of 200 single-season strikeout players is a whopping five players long: Mark Reynolds, Adam Dunn, Chris Davis, Chris Carter, and Drew Stubbs.  Kris Bryant had 199 in his rookie season (he was called up late to the bigs due to service-time considerations, so it’s likely that he would have joined this club), and Ryan Howard had 199 twice and Jack Cust had 197 once.  Dunn, Howard, and Cust again…

I love Aaron Judge, and I love 500-plus foot home runs, but we also have to be realistic and rational in our love and praise for the slugger.  The worst thing that the New York sports world can do is rattle this kid if, and when, he goes from being an All-Star to the 25th man on a roster.  There is nothing I want to see more, as a Yankees fan and a baseball fan, than Judge succeed; it’s good for the sport.  But I also don’t want to get my hopes up too high, because nothing stings more than a player of his caliber going down the path of Adam Dunn, Jack Cust, or Ryan Howard.

Early this spring I did a writeup on dScore (“Dominance Score), an algorithm that aims to identify early on pitcher “true talent.” That article reviewed RP performance for 2016.

Here’s a quick review of dScore and how it works:

dScore takes each pitcher and divides them up into a bunch of stats (K-BB%, Hard/Soft%, contact metrics, swinging strikes; as well as breaking down each pitch in their arsenal by weights and movements). We then weight each metric based on indication of success–for relievers, having one or two premium pitches, missing bats, and minimizing hard contact are ideal; whereas starters tend to thrive with a better overall arsenal, minimizing contact, and minimizing baserunners. Below is a breakdown of the metrics we used in our SP evaluations:

Performance metrics: WHIP, K/BB%, Soft%, Hard%, GB%, Contact%, SwStk%, Z-Contact%, O-Contact%

Pitch metrics: wPitch, vPitch (where “Pitch”= FA, FT, CU, SL, CH)

Our current weighting for SPs is a bit more subjective and complex than our RP weighting system, but I’m looking to implement a similar weighting system to the way we weight RP metrics in this evaluation in the near future.

dScore has been around for a year or so now, and one thing I was asked when I initially posted was whether or not it has any “predictive” tendencies. The answer is a pretty clear “no”–BUT what it does do very, very well is validate performance. There’s a fine line between saying “the numbers say pitcher X’s going to stay good” and saying “pitcher X has been good, and this confirms he’s been good”. The problem with the metric is it uses per-pitch statistics, rather than Fielding-Independent metrics. What that means is at a technical level, dScore views the pitcher as directly responsible for everything that happened after a pitch is thrown. There’s been a few outside cases that I’ll get into in a later article; but generally if a pitcher’s been bad, he’s generally viewed as having been bad, or vice versa. It seems particularly bad at projecting regression from underperformance, although I haven’t been tracking pitcher movement as well as I should. I’ll look to implement some sort of evaluation by next year.

The Studs

The top eight guys are really a who’s-who. Scherzer, Wood, Kluber, Sale, Kersh, Keuchel, Syndergaard…Only guy I’m touching on here is Thor, who’s close to begin throwing again. Lat injuries are a whole lotta “?????” for pitchers, but he’s certainly worth a buy if someone is (stupidly) wanting to sell.

The Loaded Teams

Astros – Dallas Keuchel (6), Lance McCullers (8), Brad Peacock (24) / McCullers has broken out. Consider him a stud going forward.

Diamondbacks – Randall Delgado (9), Zack Godley (10), Zack Greinke (18) / Delgado is likely more of a bullpen option at this point. Godley had an awful first outing off the break, but dScore really believes in him.

Dodgers – Alex Wood (2), Clayton Kershaw (5), Kenta Maeda (17), Rich Hill (21) / Come on, really? Give some other team a chance!

The Young Breakouts

Zack Godley (10) – I touched on him above. Although I’m pretty sure he’s due for regression, dScore continues to think he’s got premium stuff. Continue to roll with him.

Luis Castillo (14) – He’s 29 innings into his big-league career, but that’s also 29 innings vs. the Nationals (twice), Rockies (once, in Coors), and the Diamondbacks (once, in Chase). All three teams rank in the top five in the NL in runs scored. BUY. / FUN FACT: The Rockies rank third in runs scored, but are tied with the Padres for dead last in the NL in wRC+ at 81.

James Paxton (16) – He is who we thought he is.

The Still Believin’

Kenta Maeda (17)

Masahiro Tanaka (22)

Danny Salazar (23)

Tanaka’s been god-awful. dScore agrees with his 3.73 xFIP though, and says he should’ve been significantly better than he is. Salazar has somehow been worse, but once again dScore sides with his 3.57 xFIP and says BUY when he comes back from the minors, although I feel like that’s what Salazar’s always been. Every metric says he should be significantly better than he actually is. In 10 years I feel like his career is going to spawn the ultimate sabermetric “what could have been” from FanGraphs.

The Just Missed

Jacob Faria (26)

Jose Berrios (28)

Mike Clevinger (29)

Jordan Montgomery (30)

Chris Archer (31)

A whole bunch of kids and Archer, aka the pitcher we all want Danny Salazar to be.

R.I.P

Nathan Karns (19) – Thoracic Outlet Syndrome. Well, it was a good idea for the Royals…

Notes From Farther Down

Newly-minted Cubs ace Jose Quintana is sitting at 76th. Remember how I said this metric was bad at projecting regression from underperformance? Quintana was sitting just inside the top 100 before his last start. Even though dScore agrees he’s been bad, I’m still buying Quintana in bulk. Old Cubs ace Jon Lester is still getting love from dScore, even after his absolute meltdown vs the Pirates. He’s at 39th. Fellow lefties Sean Manaea and Eduardo Rodriguez bookend him at 38th and 40th respectively. Manaea was sitting in the high-teens for most of the season, then seemed to lose feel for his slider and effectively stopped throwing it. That really hurt his hittability and K’s. It came back around last start vs. Cleveland. I’m continuing to buy him as a #2 ROS. Boston activated Rodriguez recently. Adam Wainwright (104), Julio Teheran (108), Jake Odorizzi (123), Matt Harvey (137), Aaron Sanchez (140), Cole Hamels (143) are a whole bunch of ughhhhh. I’m out on all but Hamels, who I’d argue to hold. His strikeouts disappeared before getting shelved with an oblique strain, then got shelled in his first start back vs. Cleveland. His last three starts have been vintage, and I’m anticipating dScore to catch back up.

There are a multitude of ways that we can judge pitchers. Most people look at earned run average to gauge whether a pitcher has been successful, while many old school announcers will still cite a pitcher’s win-loss record. ERA is a nice, easy way of looking at how a pitcher has performed at limiting runs, but it doesn’t come close to telling the whole story. In the early 2000s, Voros McCracken created the idea of Defense Independent Pitching Stats or DIPS, which credited the pitcher only with what he could actually control. Fielding Independent Pitching was born from this theory and only took into account a pitcher’s strikeouts, walks and home runs allowed. It turns out that a pitcher’s home run rate is not terribly consistent, thus xFIP was created by Dave Studeman to normalize the home run aspect of the FIP equation by using the league home run per fly ball rate and the pitcher’s fly ball rate.

In 2015, a new metric was developed by Jonathan Judge, Harry Pavlidis and Dan Turkenkopf called Deserved Run Average or DRA. This new stat attempts to take into account every aspect that the pitcher has control over and control for everything that he does not, thus crediting the pitcher only for the runs that he actually deserves. DRA, however, is still dependent on the result of each batted ball. If the batter hits a ball deep in the gap and it rolls to the wall, the pitcher is charged with a double, but if the center fielder lays out and makes a remarkable catch, the pitcher is credited with an out. When evaluating pitchers, why should it matter whether they have a Gold Glove caliber defender behind them or not? It shouldn’t, and that’s where Expected Run Average comes in.

Expected Run Average or XRA gives pitchers credit for what they actually can control. FIP attempts to do this as well but assumes that pitchers have no control over batted balls. While the pitcher does not control how the fielders interact with the live ball, he does have an impact on the type of contact that he allows. XRA is based on a modified DIPS theory that the pitcher controls three things: whether he strikes the batter out, whether he walks the batter and the exit velocity, launch angle combination off the bat. After the ball leaves the batter’s bat, the play is out of the pitcher’s hands and should no longer have any effect on his statistics. The goal is to figure out a way to measure, independently of the defense and park, how each pitcher performs on balls in play. Since 2015, StatCast has tracked the exit velocity and launch angle of every batted ball in the majors. Each batted ball has a hit probability based on the velocity off of the bat and its trajectory. The probability for extra bases can also be determined. These batted ball probabilities have been linearly weighted for each event including strikeouts and walks to give each player’s xwOBA, which can be found on Baseball Savant. This is the perfect way to look specifically at how well a pitcher has performed on a per plate appearance basis.

Once xwOBA is found, then XRA can be calculated. The first objective is to find the pitcher’s weighted runs below average. To do this, I used the weighted runs above average formula from FanGraphs except I made it negative since fewer runs are better for pitchers.

wRBA = – ((xwOBA – League wOBA) / wOBA Scale) * TBF

For example, Max Scherzer has had a .228 xwOBA so far this season and has faced 487 batters. After finding the league wOBA and wOBA scale numbers at FanGraphs I can plug these numbers into the formula.

– ((.228 – .321) / 1.185) * 487 = 38.22

Max Scherzer has been 38.22 runs better than average so far this season, but now I need to figure out what the average pitcher would do while facing the same number of batters. To find this I need the league runs per plate appearance rate and multiply that number by the number of batters that Scherzer has faced.

League R/PA * TBF = Average Pitcher Runs
.122 * 487 = 59.41

So a league average pitcher would have been expected to surrender 59.41 runs facing the number of batters that Scherzer has so far this season. Now that we know how the average pitcher should have performed we can find the expected number of runs that Scherzer should have surrendered so far this season by subtracting his wRBA of 38.22 from the average pitcher’s runs.

Average Pitcher Runs – Weighted Runs Below Average = Expected Runs
59.41 – 38.22 = 21.19

Based on Scherzer’s xwOBA, he should have only given up 21.19 to this point in the season. If this sounds incredible it’s because this is the lowest mark of any starting pitcher though the first half of the season. Finally, XRA is found by using the RA/9 formula by multiplying the expected number of runs allowed by 9 and then dividing by innings pitched.

(9 * Expected Runs) / Innings Pitched = XRA
(9 * 21.19) / 128.33 = 1.49

Max Scherzer’s XRA of 1.49 is easily the lowest of any starter through the first half. The second best starter has been Chris Sale who has a 2.15 XRA. Of course these names are not surprising as they each started the All Star Game and are both currently the front runners for their leagues’ respective cy young award.

Here is a list of the top ten qualified pitchers:

And the bottom ten:

Full First Half XRA List

It is interesting to see that three members of the Giants rotation rank in the bottom seven in all of baseball. In fact, AT&T Park is such a pitcher-friendly park that once you park adjust these numbers, Moore, Cain and Cueto become the three worst pitchers in baseball. It’s not surprising then why the Giants are having such a disappointing season.

One measure of a good stat is whether or not it matches your perception. Therefore, while it is interesting to see Dan Straily as one of the best pitchers in baseball and Johnny Cueto as one of the worst, it is much more assuring to see Max Scherzer, Chris Sale and Clayton Kershaw as some of the very best in the sport. The numbers for relievers also reveal how dominant Kenley Jansen and Craig Kimbrel have been. This is all good evidence that XRA is doing what it is supposed to do, accurately displaying how good pitchers have actually been, independent of all other factors.

Another important characteristic of a good stat is how well it correlates from year to year. While ERA is the most simple and popular way to look at pitchers, it is not very consistent. XRA is much more consistent than ERA and FIP and also compares favorably with xFIP. However, it is not as consistent as DRA. DRA controls for so many aspects of the game that it should be expected to be the most consistent. However, being the most predictive or most consistent stat is not necessarily the goal of XRA. The real goal is to show how well the pitcher actually did, and XRA seems to do this remarkably. While not being as consistent as a stat like DRA, the level of consistency is extremely encouraging and puts it right in line with the other run estimators.

XRA is a stat that takes luck, defense, and ballpark dimensions out of the equation. When evaluating a pitcher, he shouldn’t be penalized for giving up a 350-foot pop fly for a home run in Cincinnati while being rewarded for that same pop fly being caught for an easy out in Miami. With XRA, no longer will people have to quibble about BABIP, since it is results-independent and removes all luck from consideration. A ground ball with eyes will now be treated the same whether it squirts through for a single or is tracked down for an out. Pitching ability will no longer need to be measured with an eye on the level of the defense. It takes a good offense, a good pitching staff and a good defense to make a great team, and with XRA we can finally separate all of these important factions.

Dodgers superstar ace Clayton Kershaw has already cemented himself as the greatest starting pitcher of this generation and could go down as one of the best of all time. Despite all his tremendous regular-season success, an ongoing narrative has haunted him throughout most of his career, a well-known theory that Kershaw chokes in the postseason and can’t pitch in big games.

But in reality, this actually hasn’t been the case, and the fact that so many people consider Kershaw to be a choke artist speaks more to his amazing regular-season dominance than any struggles he’s had in the playoffs. Through 282 starts in the regular season, Kershaw has an outstanding 2.35 ERA and 0.998 WHIP, so anything worse than that in the postseason is going to feel like a disappointment.

The main argument defending Kershaw’s postseason woes for awhile now has been lack of sample size. As Kershaw has reached the playoffs more and more this argument has weakened a little bit but is still relevant, as his 89 total postseason innings pitched is less than half of what Kershaw pitches in a typical regular season. It’s a large enough sample size that we can make some conclusions about how Kershaw has pitched in the playoffs, but not enough that we can judge his true-talent level. We have 1892.1 innings of regular-season data to judge his true-talent level.

Let’s start with the basic statistics. In 18 games (14 starts), Kershaw is 4-7 with a 4.55 ERA and a 1.16 WHIP. At first glance these numbers seem not horrific, but very underwhelming for what we’ve come to expect from Kershaw. This ERA is a mix of some very good starts and some not so good ones that evens out to a mediocre 4.55.

But as we start delving into the advanced statistics, Kershaw doesn’t look so bad. His FIP is a very good 3.13, with his xFIP about the same at 3.17. These stats take into account the things the pitcher can mostly control — strikeouts, walks and home runs — in an attempt to gauge a pitcher’s true-talent level in the sample size given, and are on the same scale as ERA. So in a sense, Kershaw has had some bad luck in the playoffs, and while the results still haven’t been as great as his regular-season results, he has still mostly pitched like himself.

But where does this FIP come from, and why is it so much lower than his ERA? FIP takes into account strikeouts, an area in which Kershaw has actually performed better in the postseason than in the regular season. In the regular season, he has averaged 9.88 K/9, while in the postseason, he has averaged 10.72 K/9. He has also kept his walks down in the playoffs, averaging 2.73 BB/9, which is only a little bit worse than his regular season 2.37. As a result, his 21.5 K-BB% in the postseason is nearly identical to his 21.2 regular season K-BB%. So the problems he’s had in the postseason haven’t had to do with walking too many hitters or not striking out any batters. In that regard, he’s still pitched like the Clayton Kershaw we know and love. So where have his issues come from?

The answer to that is a higher average on balls in play, a higher HR/FB%, and a bad bullpen coming in to relieve him. FIP also takes into account home runs, and he has allowed more home runs in the postseason, averaging 1.01 HR/9 (which is still good, just not Kershaw good) versus an outstanding 0.58 HR/9 in the regular season. It’s really not fair to criticize him too much for this since his postseason sample size is still less than half of a regular season. In fact, that 1.01 HR/9 is actually better than his 2017 regular season HR/9 so far, which is a very uncharacteristic 1.22 in a year where he’s been neck-and-neck with Max Scherzer for the Cy Young award. Kershaw has allowed more home runs in the postseason as a result of not only a slightly higher fly ball% but also a higher HR/FB%, 10.9 versus 7.7 in the regular season. While this doesn’t mean that he’s been unlucky, it does mean that his HR/FB% is likely to regress closer to his career norms. xFIP takes this into account and the number ends up being virtually the same as his FIP.

In addition to the extra home runs, Kershaw hasn’t been as lucky on balls in play as he has in his career. In the regular season, he’s held a .269 BABIP, which for most pitchers would be thought to be unsustainable, but Kershaw’s pitched for so long now that it’s become clear that he’s just that good. He hasn’t been quite as lucky in the postseason, where he’s allowed a .295 BABIP. And it’s not like Kershaw has allowed way more hard-hit balls in the playoffs than in the regular season, although he has allowed slightly more. He has a 20.1 line-drive rate in the playoffs, which is just slightly higher but very similar to his 19.8% in the regular season. Pitchers obviously try to prevent line drives, as they often result in hits, and Kershaw has prevented line drives from being hit about as well in the playoffs as in the regular season. So that’s not the problem.

Kershaw has allowed slightly more fly balls — 40.2 FB% versus 34.3% — and this, paired with the higher HR/FB%, makes for a bad combination and more home runs. He’s still allowed ground balls at a similar rate, only slightly less, at 39.7% versus 45.9%. So has Kershaw allowed more well-hit balls in the postseason than in the regular season? Yes, but only slightly, and not enough that he should be considered a choker. The only slight increase in line drives shouldn’t result in as big a gap in BABIP as it actually does, meaning that luck has not quite been on Kershaw’s side the way it has been in the regular season. He’s struck people out like regular-season Kershaw, he’s prevented walks like regular-season Kershaw, and he’s prevented balls from being well hit only slightly less than regular-season Kershaw. That, in addition to slightly more fly balls leaving the ballpark, has resulted in a really good pitcher that maybe is not quite as good as regular-season Kershaw, but still very good, and it certainly doesn’t warrant calling him a “choke artist.”

It can also be argued that Kershaw has been overused and over-pressured to do well. He’s been so ridiculously good in the regular season that the expectations are for him to be just as good in the playoffs and to do it practically every three or four days against the best teams in baseball. Anything less and he seem like a disappointment. People often overlook the great moments he’s had in the playoffs, like when he came out of the bullpen against the Nationals to save a tight game or when he dominated the eventual World Champion Cubs in Game 2 of the 2016 NLCS. As a result of high expectations and trust in Kershaw, he has perhaps been left in games slightly longer than he maybe should have.

An occurrence that has plagued Kershaw in the postseason a few times is going deep into games and then getting hit around before his exit from the game. He’s often left with men on base, and the relievers coming in after him haven’t exactly been kind to him, allowing nine of the 14 runners he’s left on base to score. Let’s say the bullpen comes in and dominates, stranding all 14 of those runners, and his postseason ERA drops from 4.55 all the way down to 3.64.

Also remember that in the playoffs, teams are in their full strength and effort, doing everything they possibly can to try and win. These are the best teams in baseball, the teams that had everything working well enough for 162 games to make it past all the other teams and into the playoffs. The offenses Kershaw has to face in the playoffs are going to generally be better than the average offense he might face throughout the season. It is not uncommon for great pitchers to have slightly worse results in the playoffs. Madison Bumgarner, a famous “postseason hero” for the Giants, has a postseason FIP only 0.02 better than Kershaw’s and an xFIP 0.43 worse than Kershaw’s. Luck can go in very different directions for some pitchers in small sample sizes, and this is a perfect example.

Look at Pedro Martinez. In more postseason innings pitched than Kershaw, he has a significantly worse FIP/xFIP (3.75/4.31) despite an unsustainable low BABIP of .257, lower than his regular season .279. And no one thinks of him as a postseason “choker.” Greg Maddux, another all-time great, also has a worse FIP/xFIP (3.66/4.45) than Kershaw in even more innings pitched (198). And nobody considers him a postseason choker. Roger Clemens is the same deal. 3.52 FIP, 3.91 xFIP in 199 innings pitched. These pitchers are still considered all-time greats despite having postseason numbers that are arguably worse than Kershaw’s.

This really goes to show just how good Kershaw has been in the regular season. He puts up godlike numbers and then when he puts up “only” good numbers in the playoffs, it seems like he’s bad in comparison. When you look at the aforementioned fellow all-time greats, it’s clear that Kershaw is not the first great pitcher to have a little trouble in the playoffs.

So has Kershaw been as utterly dominant in the playoffs as in the regular season? No. But has he been a choke artist who gives up eight runs every time he’s put under pressure? No, not at all. He has had some rough outings in the postseason, particularly against the Cardinals, where he hasn’t been able to dominate and take control of the game quite like normal, but he has also had plenty of good moments of great pitching and when he’s left with runners on base, his bullpen has mostly let him down. All he really needs is one great World Series run to erase this ongoing narrative once and for all. No matter what, these small hiccups in the playoffs shouldn’t diminish the legendary career that Clayton Kershaw is in the midst of.

The panel on ‘The Changing State of Sabermetrics: at the 2017 SABR convention in NYC with panelists Joel Sherman, Mark DeRosa, Vince Gennaro and Mike Petriello claimed that fewer balls are going into play and singles are actually down. They posed the question, “Are singles still a thing?”

With that in mind, we aimed to verify if these claims are true and what makes people feel that players are hitting fewer singles in today’s game.

We used data that’s current as of July 2, 2017.

NOTES:

• Data being visualized is from http://www.baseball-reference.com/leagues/NL/bat.shtml and http://www.baseball-reference.com/leagues/AL/bat.shtml on July 1, 2017.
• BABIP definition is taken from  http://www.fangraphs.com/library/pitching/babip/

Below you will see two charts illustrating the number of hits, home runs and strikeouts per game.

You can conclude three things from these graphs:

1. Over the past 10 seasons, strikeouts have been increasing dramatically — 1.94 K/Game in the AL and 1.52 per game in the NL.
2. Over the past 3 seasons, singles per game have dipped.
3. Over the past 3 seasons, HR per game have spiked higher than ever before.

To say that Kevin Siegrist has really struggled in 2017 would be an understatement. After allowing 15 earned runs in 31 appearances through June 22, he was placed on the DL with a cervical spine sprain. With an ERA near 5, Cardinals fans have been left wondering what happened to the player who led the league in appearances (81) and finished third in holds (28) in 2015.

At first glance, Siegrist has an obvious issue — a very clear and very serious velocity problem. Take a look at this graph.

The velocity of his fastball has decreased every year since 2013. It hovered around 95.8 mph at one point, but more recently it’s dropped well below 93 mph. That’s a significant decrease, as the steep slope indicates. And for the first time, Siegrist, who is a reliever, has a fastball velocity well below a league average that includes starting pitchers.

If you have ever looked at aging curves, for hitters or pitchers, then you know that skills decline with age. Certainly, pitching velocity is no exception to this rule. Still, Siegrist is an extreme case.

Velocity very clearly declines with age and Siegrist has fallen right in line with this trend. For the first two or three years of his career, his changes in velocity pretty closely matched the aging curve. However, for the last two years, there has been a marked decrease.

In case you haven’t gotten the point, here’s one more graphic that shows Siegrist’s velocity problem.

This slope looks more like something I would ski down than data you want to see from a pitcher’s velocity. Clearly, Siegrist had an excellent stretch in 2015 and he produced the numbers to back that up. Other than that, we see a pretty consistent decline.

So, is that it for Kevin Siegrist? A slow decline into oblivion? I don’t think so. I actually expect him to far surpass expectations in the second half of the year.

What if I told you, Siegrist has actually improved this year? He’s not telegraphing his pitches. He has improved his tunneling. (For extra reading, here are primers on tunneling from The Hardball Times, Baseball Prospectus, and FanGraphs.)

Essentially, tunneling is the ability of a pitcher to repeat his delivery with similar, if not identical, release points. If a pitcher is able to do this, a batter has less time to recognize the pitch and a lower chance of getting a hit. If a pitcher’s release points are completely different, say for his fastball and changeup, a hitter can more easily distinguish between the two and put a better swing on the ball.

These are Siegrist’s release points from 2015 (his most successful year).

And here are the release points from the first half of 2017.

Let’s keep in mind we’re talking about inches here, not feet. Still, the differences between these two years are significant. The release points from 2015 are more spread out than the data from 2017. Siegrist has improved his ability to replicate pitch deliveries. Unfortunately, due to his decreased velocity, this hasn’t resulted in any type of noticeable success.

In 2015, the changeup and the slider release points overlapped nicely, but the fastball release points stick out like a sore thumb. In 2017, with the addition of a cutter, there is much more overlap among the pitches. If he can keep this up, it should translate to long-term success.

Moving away from release points, pitch virtualization data confirms the same hypothesis: that Kevin Siegrist has improved his ability to replicate his delivery.

This is the data from 2015. To the average viewer, and even probably to you and me, this doesn’t look too bad. At the 55-foot mark, the pitches have pretty similar locations. Even at the 30-foot mark, it’s probably pretty difficult to distinguish between five of his six pitches.

If we compare it to the 2017 data, we see a considerable difference.

It’s pretty clear, right? At 55 feet, the release points aren’t “pretty similar,” to use my own wording, they’re practically identical. And the trajectories remain extremely close to one another until about the 20-foot mark, when they break. 20 feet at 93 miles per hour (an all-time low velocity for Siegrist) gives the batter about a tenth of a second to decide what to do.

There is no denying that Kevin Siegrist has a velocity problem that he would do well to fix. And if the first half of 2017 is any indication, it needs to happen fast. It is unfortunate that he has not been able to reap the benefits of an improved delivery. The consistency in release points that Siegrist has shown during an abysmal 2017 is encouraging and should provide a source of hope going into the second half of the season.

Throughout the baseball season, I like to estimate teams wins, but I don’t do it in the traditional way. Some time ago, I discovered that I could use innings pitched to get a close estimate. Here’s what I do:

1) Take team games played and divide by 2;

2) Take the team’s innings pitched and subtract the team opponents’ innings pitched;

3) Add 1 and 2.

For example, the Washington Nationals, as of the All-Star break, have played 88 games. They have 789.33 IP, and their opponents have 781.33 IP. So I take 88 divided by 2, which gives me 44. Then I take 789.33 minus 781.33, which gives me 8. Then 44 plus 8 gives me an estimate of 52 team wins. Checking the standings, I see that Washington indeed has 52 wins.

How does my method compare with the traditional Pythagorean? (The Pythagorean method, of course, takes runs scored squared and divides by runs scored squared plus runs allowed squared.) I’ve set up some charts to demonstrate. First, let me present the relevant statistics for all teams as of the All-Star break (all statistics courtesy CBS Sportsline):

Now let me present a chart showing how many teams wins are predicted by my method and the Pythagorean method (for the Pythagorean method, I’m using 1.82 as my exponent, as shown by MLB on their Standings page):

My method appears in the second column, and the Pythagorean method appears in the third column, with actual team wins in the last column. My method, as shown above, gives estimated wins directly. The Pythagorean method actually computes winning percentage. To get the estimated wins for the Pythagorean method, I multiplied the team’s estimated winning percentage by the team’s games played.

The methods are pretty close! On a couple of teams, though, the methods miss by a wide margin. I’m way off on the Angels, for example, while Pythagoras is off on the Giants. But which of these methods is closer overall? I did an r-squared between each of the estimated win columns and the actual wins and got these results:

Mine’s a little higher, but let’s use mean squared error (MSE) as a cross-check. Here are my numbers:

I’m not a numbers person, so if I’ve made made errors in my calculations, please let me know, and I will never, ever trouble you fine readers again with another post. But I’ve published previous studies of both methods (in other places, under other names) and have found each time that my method edges out the Pythagorean in both r-squared and MSE.

If my method works at all, it’s because better teams typically have to get more outs to finish off their opponents. If the Dodgers, say, are at home against the Phillies, chances are they’re already winning when they go to the bottom of the ninth, and so the Dodgers don’t have to come to bat. That means the Dodgers had to get 27 outs and the Phillies had to get only 24. Conversely, on the road, if the Dodgers are leading the Phillies, the Phillies have to come to bat in the bottom of the ninth, and the Dodgers have to get the full 27 outs to end the game.

One caveat: my method tends to be more descriptive than predictive, so it’s a better measure of how a team has performed, not a good predictor of how a team will perform in the future. The Pythagorean method is much better as a predictive tool.

So there it is! My estimated team wins method. I hope you find it useful.

Many of the talking heads on radio and TV have commented on how playing in the WBC and skipping part of spring training negatively affects player performance during the regular season. As a Texas Rangers fan who has wondered the same thing, I decided to do a quick and dirty analysis.

The Ground Rules

• WBC rosters were pulled from Wikipedia 2017 World Baseball Classic rosters.
• Player WAR data was pulled from FanGraphs on July 10, 2017.
• Only MLB players were included.
• Only players with MLB statistics in both 2016 & 2017 were included.
• A WAR differential is defined as the difference of the 2017 WAR and 2016 WAR (2017 WAR – 2016 WAR)

The Results

Here’s the RAW data as I compiled it from the above sources.

The last column in the spreadsheet is the difference of the 2017 WAR and 2016 WAR and has a mean of -1.1 for all the players in the list.

The histogram below shows how the data is skewed to the negative, which is easily seen in the list just scanning visually.

Another interesting chart depicts the correlation between 2016 and 2017 WAR. The slope of that trend line is 0.59.

Here are the top (bottom!) 20 players, and two of my Rangers are in the list. Rougned Odor is 36th on the list with a -1.8 WAR differential.

There could be many other reasons for the decline in WAR and it very well could have nothing to do with the WBC.  It was an interesting exercise and the numbers make me wonder if MLB has really looked at the WBC and how it affects the MLB players that participate.

Being an avid lover of both baseball and video games, I naturally like to participate in both from time to time, at the same time. In fact, San Diego Studio’s MLB THE SHOW 17 is quite possibly my favorite game at the moment considering how many hours I put into it. Anyways, the reason I bring this up is that the topic of this post (the under-the-radar talent that is Travis Shaw) was brought to my attention while watching a live-stream of my favorite MLB THE SHOW YouTuber. After hearing of the inevitable rise to power that Shaw should see within the next few weeks, I decided to look more into his stats and see just how plausible this claim was.

I assume that unless you are a Brewers fan, Shaw’s ability and stats could possibly be low on your radar, especially since he didn’t crack the National League’s All-Star lineup for 2017. But after taking a close look at his stats, maybe he should have. At the time of writing this article, Shaw is hitting .296 with 18 dingers and 61 RBI. This is impressive when you compare his stats to the rest of the N.L. All-Star starting lineup that collectively averaged a .320 average, 16 home runs (2 fewer than Shaw) and 55 RBI (5 fewer than Shaw). Then, we can take it a step further and compare him directly to the lineup’s starting third baseman (Shaw’s position), Nolan Arenado, who is hitting .298 with 15 homers and 63 RBI.

At first sight, it seems as if these two are on par with one another, with a slight advantage given to Arenado in the average and RBI department. This, however, is not the case when taking into consideration the advanced stats. Shaw pulls away from Arenado in ISO, weighted On Base Average (wOBA), and weighted Runs Created Plus (wRC+), averaging .268, .386, and 135 in each stat, respectively. These stats are known to tell more of the “story” of the player, giving more details as to what is going on. Shaw is hitting for more power, creating more runs, and overall is a bigger asset to his team than many other players in their respective situations that were graced with All-Star status.

I, of course, am not saying that Arenado or any other player should not have been awarded All-Star status because they are all amazing ball players with enormous talent. Really, the only point that I am trying to get across is that, based on stats, Shaw should have most definitely been a part of the current National League All-Star group. And as for the rest of the season, the future is very bright for Shaw, especially considering that he is now a sleeper candidate for National League’s Most Valuable Player, according to ESPN.

*Side note* This is my first post in the FanGraphs community! And while I am very excited, I at the same time want to be sure to improve with each and every post and write about things that people want to hear. If you, the readers, do not have anything to say about the content of the articles but do have some constructive criticisms please feel free to leave a comment! Have a good one!