Archive for Three True Outcomes

Last season was the year of the three true outcomes specialist.  Aaron Judge’s dominant three true outcomes season was the most prominent example of this: he ranked second in home runs (52) and walks (127) and first in strikeouts (208).  In total, 57% of his plate appearances resulted in one of the three true outcomes.  He was the American League Rookie of the Year and in the running for the 2017 American League Most Valuable Player award, finishing second.  His performance helped the Yankees reach the American League Division Series.

We know that the three true outcomes rate has been increasing.  In part, this is due to the average player increasing his rate of home runs, strikeouts and walks.  But there is also the unusual player in the mold of Judge who takes an extreme approach at the plate resulting in dominant three true outcomes seasons.  The number of these hitters has been increasing over time.

Figure 1. Three True Outcomes Specialists per Season, 1960-2017

View post on imgur.com

Figure 1 shows the number of dominant three true outcomes player seasons over time.  To get here I examined all players since 1913 with at least 170 plate appearances in a season.  I considered a dominant season one with a three true outcomes rate of at least 49%.  There have been 132 player seasons with a three true outcomes rate of at least 49%.  All of them have taken place after 1960.

The graph shows that the number of dominant seasons has been increasing over time.  Since Dave Nicholson first did it in 1962, most years have had at least one player cross the threshold.  Since 1994, every season has had at least one.  From 2001 to 2010 there were four seasons with five three outcomes hitters.  There was six in 2012 and eight in 2014.  The trend is currently peaking with 13 in 2016 and 16 in 2017.  The trend is a bit more extreme but similar to the average increases in three outcomes rates over time.  It seems that more players pursue (and teams tolerate) an approach to hitting that includes extreme rates of the three outcomes.

It is worth pointing out that those 16 players in 2017 account for about 4% of all players with at least 170 at-bats.  Three true outcomes specialists are more common but still rare.  Who are those 16 players?  Table 1 lists them including the home run, walk and strikeout rates, and the combined three true outcomes rate for the year.

Table 1. Three True Outcomes Specialists, 2017

The list includes many of the unique player stories of the year.  Aaron Judge’s rookie year was historic.  Joey Gallo made waves, particularly for his extreme three true outcomes rates.  Miguel Sano was an All-Star who helped lead the Twins to a bounce back year and a wildcard spot.  Eric Thames was a surprise story of the year, returning from a year in Japan and sparking the Brewers to an early lead in the National League Central.

Notable about this list is the young cohort of hitters who have consistently taken the all or nothing approach of the three true outcomes specialist.  Judge, Olson, and Blash all made their MLB debut in 2017.  Gallo still qualified as a rookie despite making his debut in 2016.  Keon Broxton, Ryan Schimpf, and Kyle Schwarber are in their second year.  Sano has been a specialist for three years running.  Sure, there are old hands like Napoli and Carter, and Davis who take the all or nothing approach, but the record number of specialists the last couple years have been due to this young cohort of three true outcomes specialists.  A new record will come down to 2018 rookies who practice this all or nothing approach heading into their major league debuts, and the number of teams willing to tolerate the strikeouts that come with this approach.

Jim Thome was elected to the Hall of Fame on January 24th.  Given my recent obsession with the three true outcomes, I immediately recognized the significance of this event.  I believe Jim Thome is the first, and likely the last three true outcomes Hall of Famer.

Table 1 shows Thome’s home run, walk, and strikeout rates along with his three true outcomes rate for each season.  The final column is the MLB average three true outcomes rate for the season.  Thome was a three true outcomes machine from 1996 until his retirement in 2012.

Table 1. Jim Thome, Three Outcomes Hall of Famer

Thome was part of a small group of specialists with multiple dominant three true outcomes seasons.  Table 2 provides a list of players with 4 or more of these dominant seasons.  I consider a season with at least 170 plate appearances and a 49% three true outcome rate as a dominant season.  The casual three true outcomes observer will recognize the players on this list as notable specialists.  Rob Deer, of course, is the iconic three true outcomes hitter.  I used Deer’s career three true outcomes rate of 49% and 4 dominant season to construct the table.

Table 2. Dominant Three True Outcomes Specialists

Thome’s 10 dominant seasons are more than any other player.  He is also the only Hall of Famer on the list.

Maybe Mark McGwire should be in the Hall of Fame (depending on your PED era position).  Already past eligibility to be inducted by the Baseball Writers Association of America (BBWAA), perhaps he will have a chance in the future with the Veterans Committee.

Adam Dunn will be on the 2020 ballot.  He was a consistent three true outcomes specialist, but we will see if the BBWAA consider him a dominant player over the course of his career.

Russel Branyan and Jack Cust are interesting players to see on this list.  Branyan makes the list because of my 170 plate appearance requirement.  Cust was a dominant three true outcomes hitter for five straight years, 2007-2011.  Neither are on the Hall of Fame ballot.

Carter and Avila do not have contracts for 2018, but could land somewhere.  Davis is signed with Baltimore through 2022.  Joey Gallo and Aaron Judge are two young hitters in the three true outcomes mold not yet on the list.  So maybe it is too soon to make a judgement on the Hall of Fame potential of three true outcomes hitters in the future?

But I am going out on a limb to say that despite the trend towards three true outcomes baseball, we have seen our first and last three true outcomes Hall of Famer in Jim Thome.

During this quiet winter of baseball, I’ve entertained myself with a mild obsession with the three true outcomes (not outfits).  Dave Cameron took note of the three outcomes trend in early April 2017, only one week into the new season.

 “…while samples are still tiny for players and teams, things tend to stabilize pretty quickly at the league level. And, not surprisingly, the first week of the season was filled with the two things MLB games are becoming known for; strikeouts and home runs.”

Cameron goes on to predict a record year for the three true outcomes in 2017.  He was right.

A steady slog towards three outcomes baseball

Figure 1 is an update of Bill Petti’s analysis back in 2012.  It looks at the average rate of the three outcomes by player for each season since 1913.  The top blue line shows the proportion of plate appearances that have resulted in either a home run, strikeout or walk across seasons.  I got here pretty easily: Like Bill, for each player, I added their home runs, strikeouts and walks in a season and divided that by the number of plate appearances.  That provides the proportion of three outcomes plate appearances for each player.  Unlike Bill, I used at least 170 plate appearances in a season as my cut off (rather than 500).  Then, for each season, I found the average proportion of three outcomes plate appearances for eligible players.  I followed this procedure for home runs, strikeouts and walks separately.

The trend of the blue line is clear: players’ average rates of three outcomes have increased starting around 1920, dropping off a bit in the 70s, and continuing on through 2017.  There has been a spike in recent years, and the 2017 average rate of 33% three outcomes is clearly the highest since 1913.

It is interesting to look at each outcome separately as well.  The tremendous growth in strikeout rates is clearly a big part of this story.  Average walk rates have consistently hovered around 8 or 9 % of plate appearances after peaking just over 10% in 1948.  Home run rates are increasing, but home runs are still rare compared to the other outcomes.  Through the 1940s average home run rates per player averaged around 1% of plate appearances.  They jumped to around 2% in the 1950s, and stayed pretty consistent until the mid-1990s.  Since then they have been increasing and peaked in 2017 with an average of 3% of plate appearances.

What does this trend towards three outcomes mean for the game?  Perhaps Cameron subtly showed his feelings about it when he titled his essay, “The League’s Continuing March Towards Three Outcomes Baseball.”  A march; a steady march.  Figure 1 suggests an uphill march.  Cameron could have called it the race towards three outcomes.  Or the growing excitement of three outcomes baseball.  Even the uphill battle of three outcomes baseball sounds more engaging than a march.  Is three outcomes baseball more of a slog than it is an exciting new dynamic in the game?

First things first, I would like to credit my friend Elling Hofland for coming up with the main idea of this piece. He’s the one who provided me with his thoughts and theories that allowed me to expand on this topic in the first place. Give him a follow on Twitter for sports and stats-related banter; his handle is @ellinghofland.

BABIP, or batting average on balls in play, is an incredibly useful stat. It does a fantastic job at using both luck and quality of contact to give a better grasp as to how a player actually performs during batted-ball events. These batted-ball events only take up a certain percentage of a player’s plate appearances. BABIP rate focuses on how many plate appearances a player has relative to the number of batted-ball events they have. To calculate BABIP rate, you take at bats minus strikeouts and home runs, plus sacrifice flies, and divide that by plate appearances. For example, if a player has 600 PA during a single season along with a 300 batted-ball events, they have a BABIP rate of .500.

Now, if you look at the three variables taken out of that equation, you’re left with walks, strikeouts, and home runs, otherwise known as the “three true outcomes.” These are called true outcomes due to the fact that none of them (for the most part) involve defense on the field. A shortstop can’t screw up a strikeout, walk, or a home run. You can take these three true outcomes and turn them into a rate as well. If you add up a player’s strikeouts, walks, and home runs and then divide them by plate appearances, you get TTO rate.

Let’s look at Mike Trout. In 2017, Trout’s BABIP currently sits at .369. However, he has a BABIP rate of .550 along with a TTO rate of .435, meaning that 55% of his at bats end with a ball in play, while 43.5% of his plate appearances result in a strikeout, walk, or home run. Both BABIP rate and TTO rate are useful stats, as they essentially show how well and how often a player makes contact. While BABIP itself is useful, it can be hard to tell how luck is involved in a batted-ball event when it isn’t hit over a fence for a homer. BABIP rate attempts to bridge the gap between BABIP and the three true outcomes.

Miguel Sano is a well-known slugger. In his three seasons in the majors, he’s smashed the ball when he’s hit it, boasting exit velocities of 94.0 in 2015, 92.3 in 2016, and 93.1 in 2017. Despite these consistent EVs, his BABIP has fluctuated from 2015 to 2017, with marks of .396, .329, and .385, respectively. If we look at his BABIP rate from 2015-2017, they look like this: .429, .478, and .473. Despite the difference in his BABIP from 2016 to 2017, his BABIP rate has stayed nearly the same, meaning that he’s still making the same amount of contact with the ball despite fewer balls falling for hit in 2016. Looking solely at BABIP, it could be argued that 2016 was his “regression” to where he should be after sporting an incredibly high BABIP in 2015. In 2017, one could say his high BABIP is a cause for concern, as he may just be getting lucky. However, his BABIP rate shows that isn’t the case.

Let’s look at another player, Brandon Phillips. Phillips’ BABIP has been incredibly consistent during his past three years, sitting at .315 in 2015, .312 in 2016, and .305 in 2017. Additionally, his BABIP rates have been .820, .816 and .802. Phillips puts the ball in play nearly 80% of the time on a regular basis.

So, as you can imagine, there is a real link between BABIP rate and TTO rate. The more contact a player makes, less they tend to walk or strikeout. Thus, a high BABIP rate equals a low TTO rate. This is exactly what we see if we attempt to correlate these two stats. Below is a snapshot of a graph that shows TTO rate vs. BABIP rate.

Players names aren’t included because, A) it clutters the graph, and B) they aren’t necessary at this point. Accompanying this graph is a trend line with an R squared value, otherwise known as a correlation coefficient. Essentially, an R squared value measures how well your model fits your data, or in this case, how closely correlated  TTO and BABIP rate are to each other. It turns out that the R-squared value is .991, which means that the relationship between BABIP rate and TTO rate fit very well together: in fact, you’ll find that TTO rate and BABIP rate are almost the exact opposites of each other. The players with the top 10 lowest BABIP rates in the MLB all have TTO rates of .437 or higher, meaning that their at bats result in an outcome of a walk, home run or strikeout 43.7% of the time. Inversely, players with the lowest BABIP rates all have TTO rates of .225 or lower.

We can also derive more information from these numbers using this correlation. Players who have a low BABIP rate have a very high OPS. Remember, these players also have high TTO rates. The top 10 players, Judge, Sano, K. Davis, Souza Jr., Reynolds, Morrison, J. Upton, C. Santana, Lamb, and Stanton all have an OPS of .841 or higher. The players with the highest BABIP rates (or lowest TTO rates) have an OPS of .798 or lower.

BABIP rate can tell us a lot of about a player. Just by glancing at a player’s BABIP rate, you can have an instant idea of how often the player walks, strikes out, or hits dingers. Not only that, but it you can tell you a lot about their offensive production. High TTO rates usually mean high hard-hit rates along with high exit velocities. BABIP rate also helps understand BABIP itself better and teaches that you can’t judge a player by BABIP all the time. In most cases, players with an over-inflated BABIP (relative to past performances), just tend to mash the absolute heck out of the ball, as told by their low BABIP rates and high TTO rates. On the opposite end, players with a steady BABIP will have very high BABIP rates and tend to be contact hitters that put the ball in play and don’t hit for power. BABIP rate, along with its correlation to TTO rate, has the potential to be a powerful, tell all offensive stat.

INTRODUCTION & EXPLORING THE QUESTION

So there’s been a lot of attention paid to Three True Outcome guys recently. The subject was touched upon in a recent article by Craig Edwards, as well as in this community blog by Brian Reiff. These articles brought attention to guys who are notable for putting 7 of the 9 defensive players to sleep. However, what caught my attention the most was a comment on Craig’s piece by “steex” who proposed a hypothesis about these sluggers:

I think this makes selecting TTO players strictly by the numbers difficult. For me, the spirit of TTO is a player that does enough good (HR+BB) to balance out for a lot of bad (K). Harper and Votto don’t really fit that definition in the intended way, but rather show up on the list because they do SO much good (HR+BB) that their total HR+BB+K makes the cut despite having not as much of the bad (K).

I wonder if a better list of players comparable to one another would be obtained by first sorting by TTO%, then subdividing that by the percentage that Ks represent from the TTO events (i.e., K/[HR+BB+K]). That provides a lot of separation between guys like Harper, Votto, and Goldschmidt who have strikeouts represent less than 50% of their TTO events and guys like Carter, LaRoche, and Belt who have strikeouts as more than 65% of their TTO events.

This was also supported by follow up comments speaking about how they differentiate the players into two groups, those who strike out at a higher clip and those who have BB% and HR% compensate for a reduced K%. My goal was to figure out whether the quality aspect of the Three True Outcomes was different between these high-K% players and the low-K% players, beyond the walks and strikeouts.

THE PROCESS

First, let me define how I picked out my sample, and how I classified the players into two groups, and then I’ll begin to discuss the details of the study. I pulled all the data from 2010-2014 for player seasons who qualified for the batting title (minimum 502 Plate Appearances). This gave me a sample of 723 player-seasons (where a single player may be listed as a qualifier separately for up to five seasons). Of these 723 player-seasons, I set the Three True Outcome bar at 40%. Why 40%? Well the simple average (weighted to PA) was 29% Three True Outcome (I’ll abbreviate to TTO from now on), with a standard deviation of approximately 8%. So that would make 40% TTO somewhere around 1.5 standard deviations above the mean, which seemed like a reasonable line to draw in the sand.

There is now a sample of 52 player-seasons (7.2% of the qualifiers). From here, I had to draw a new line, and I wanted to go by “steex”’s suggestion of using the proportion of strikeouts to TTO% as the barrier. The key was getting a decent number of player-seasons on either side. I started off with 50% (using the formula K/[HR+BB+K]), but that would have left me with only two player-seasons (2011 Bautista, 2013 Votto, for those who are curious). I bumped it up continually until I reached a 60% ratio, which seemed to be reasonable. That placed 11 player-seasons in the low-K TTO group (which will be referred to as TTO-L) and 41 player-seasons in the high-K TTO group (which will be shown as TTO-H).

The whole TTO population is now divided into two groups, TTO-L (with 11 player-seasons) and TTO-H (with 41 player-seasons). Now what? I was truly curious about how these two groups differed in their hitting abilities. It seems fairly obvious that those who have lower K% and higher BB% will have higher (better) wOBAs, wRC+s, and the like (just due to trading strikeouts for walks). As Craig showed in his article, the average TTO player is an above average hitter due to a typically lumbering stature and a penchant for not being great at defense. Those who aren’t above average hitters and are bad at defense usually find themselves riding minor league buses around the country. But I’m not trying to compare TTO hitters to non-TTO hitters, rather comparing the two halves based on TTO quality.

BATTED BALL DISTRIBUTIONS

I decided to compare them using  statistics that might glean differences between good and bad hitters. I looked at batted ball distributions to start. I compiled the GB%, LD%, FB% and IFFB%, as well as the PULL%, CENTER% and OPPO% from the leaderboards (plus HR/FB for good measure), and computed the mean, standard deviation, and p-value based on a two-tailed T-Test. The results are in TABLE ONE:

Interestingly enough, the batted ball distributions are very similar between the two groups. The groups are pretty much interchangeable, with the only thing close to being statistically significant is the percentage of balls hit to center field. However, when looking at that in the bigger picture of pull/center/opposite, the numbers are nearly identical. So far, the two groups are relatively indistinguishable from one another.

BATTED BALL AUTHORITY

At this point, my mind went in another direction: do TTO-L player strike the ball better than their TTO-H counterparts? If you’ve got a good eye and can take a walk more easily, then you’re probably able to see the ball better, and therefore are able to drive the ball harder. So, even though it may not have manifested itself in the GB/LD/FB numbers, perhaps these “elite” players in the low-K group have better pop. To evaluate this, I pulled the HARD%, MED%, and SOFT% of balls by each group, along with BABIP for good measure, summarized in TABLE TWO:

Again, a little surprising to me. There’s no statistically significant difference between these low-K guys and high-K guys in terms of batted ball authority. Each group hits roughly the same, with the low-K guys trading a few medium hit balls for some hard hit ones (albeit not enough to differentiate the groups). BABIP would manifest itself in these guys striking the ball harder, and it comes out roughly even. One note that BABIP would control itself here more than in most hitter studies because the subset of TTO players typically have similar builds and are not artificially increasing BABIP by beating out infield hits (neither group would have a distinct advantage).

BATTING SELECTIVITY & CONTACT RATES

So where do these two groups separate? Something has to cause the disparity between the groups and show a differential in ability. And that something is at the plate in their selectivity – which only makes sense. Players who draw walks are those who lay off bad pitches out of the zone, and those who strike out typically struggle to identify strikes from balls, or lack the ability to contact balls when they swing (usually not both, or else they wouldn’t be in the majors). The data is summarized below in TABLE THREE:

Here’s all that red you’ve been waiting for. Starting with the first three rows, there’s a statistically significant difference (p < 0.05) between the two groups in swinging at balls (O-Swing%), which goes to show the selectivity of the TTO-L group is better than the TTO-H group. In rows four to six, we see that for swings on pitches both in and out of the zone, the TTO-L group makes contact more often, with in-the-zone contact being statistically significant at the p < 0.01 level. To summarize this table, the TTO-L hitters don’t swing as often, but when they do they are better at making contact with the pitch as compared to the TTO-H batters.

GROUP SUMMARY

The final table, TABLE FOUR, summarizes the groups for anybody who was curious. 

Obviously above you see that the K-rates and BB-rates are statistically significant, which only makes sense because that’s how we divided the groups, so that was artificially implanted. And, of course, you’ll always have a better wOBA if you walk more and strike out less, because walks count for approximately 0.7 runs based on linear weights each.

SUMMARIZING THE FINDINGS

Of the 723 player seasons between 2010 and 2014, inclusive, 52 were deemed to be Three True Outcome seasons (with 40% of the plate appearances ending in BB, K or HR). From there, the group was subdivided into two by the relative amount of K’s compared to total TTO% (with [K%/TTO%]>60% as TTO-H, and [K%/TTO%]<=60% as TTO-L.

The groups were compared against one another on Batted Ball Distributions, Batted Ball Authority, and Batting Selectivity & Contact. The vast majority of the statistically significant differences between the groups appeared in the third table, with the TTO-L group displaying a better eye for strikes, while also contacting the ball better when they decided to swing. Perhaps the most interesting finding of the study was that this increased contact did not manage to create better authority when hitting the ball, nor did it change the batted ball distribution significantly. Just because the TTO-L group made contact more often on their swings did not mean they were able to drive the ball better than the TTO-H players.

Just a quick thank you to end this, to the FG community comments that inspire people to write things like this and make my last college summer a little more (less?) exciting.