Baseball, more than any other major team sport, has the reputation for having the least athletic athletes. Jose Molina is obligated to, at times, sprint. Jorge de la Rosa must swing a baseball bat. David Ortiz sometimes has to play in the field. Having skills like catcher defense, pitching, and hitting with power will earn you playing time, and many players have such elite strengths that it’s worth it just to deal with those weaknesses. So many of baseball’s skills are unrelated that players have to spend a lot of time doing things they aren’t good at, at least relative to other MLB talent. A good way to make anyone look unathletic is to make them perform a long list of skills that have little to do with one another and compare them to the best in the world at those tasks.

I wanted to assemble a list of players who experienced something like this phenomenon the most frequently. Essentially, I wanted to see what players’ strengths and weaknesses were the farthest apart. To determine those players whose skills varied the most between themselves, I gathered what I consider to be the six stats that best describe what a player’s strengths and weaknesses are. BABIP and K% for contact, BB% for discipline, ISO for power, and Fielding and Baserunning values. I then gathered stats from 2011-2014 to better control for less reliable fielding metrics, assigned each player’s stats a percentile rank, and calculated the standard deviation of those six stats for each player.

For instance, Mike Trout’s attributes look like this:

His strikeout rate has been higher than MLB average, but he is otherwise an exceptionally well rounded player, as we know.

The most evenly talented player in baseball has been Kyle Seager, who is almost in the middle third at every stat.

Many players have much more severe strengths and weaknesses. Here are the 10 players whose stats show the greatest variation from one another.

10. Dexter Fowler

9. Ichiro Suzuki

8. Jose Altuve

7. Curtis Granderson

6. Mark Reynolds

5. Giancarlo Stanton

4. Miguel Cabrera

3. Darwin Barney

2. Adam Dunn

1. Ben Revere

The whole list is fun to look through and play around with, so feel free to click here and look through all the qualifying players.

In 2013, Baseball Prospectus chronicled the Rays’ “changeup revolution,” explaining how the Rays’ pitching development has succeeded in part because they teach pretty much everyone to offer a plus changeup in unusual situations. But while successful small market teams have thrived off using analytics to find market dislocations on players, the Rays’ changeup prowess has actually allowed them to create them.

Recently, the Rays were ridiculed for giving up David Price for a package whose most proven player was Tigers’ 5^th starter Drew Smyly. At the time of the trade, Smyly had a 3.93 ERA and 4.08 FIP. In other words, he was an average starter. But looking closely, one can see that he pitches to a drastic LHH/RHH split, with opposing wOBA’s of .196/.355, respectively. The reason behind his inability to get righties out could very well be the lack of a good secondary pitch to use on them. For his career, his most effective pitch has been his slider, with hitters putting up a meager .226 wOBA against it. His worst pitch was none other than his changeup, which has been crushed to the tune of a .488 wOBA.

Knowing that his organization specializes in teaching the changeup, I don’t believe for a second that Rays GM Andrew Friedman gave up their ace without thinking that Smyly was essentially a good changeup away from being a potent starter. A free agent in 2019 at the earliest, Smyly should easily provide more long-term value than Price will over the next 1.5 seasons. (Obviously, the Tigers will try to extend Price, but the Rays did not have that option.)

The key takeaway here is that to most teams, Drew Smyly was probably viewed as a league-average pitcher without a secondary pitch that could put righties away. But to a team like the Rays, who have proven to be adept at implementing a changeup, Smyly’s ceiling can appear to be much more feasible. So far with Tampa (small sample size warning), Smyly has thrown 36 innings in 5 starts with an ERA/FIP/WHIP of 1.50/2.82/0.69. He will certainly come back down to earth, but a valuable lesson can be derived from this trade that appeared to be a blatant ripoff. By having an organization’s pitching development specialize so much, the Rays actually manufacture their own list of “buy low” pitchers, many of whom may have plateaued in the minds of other teams.

When they traded Matt Garza, they got current front-end starter Chris Archer in return. From Prospect Instinct’s 2011 scouting report:

The Rays got a haul for Matt Garza from the Cubs and Archer was considered the Cubs top pitching prospect. He has a plus fastball and above average slider, but he still has a lot of work to do before he becomes MLB ready. His changeup is lacking and his command has been erratic. But with enough time he does have #3+ upside.

With a Tampa Bay Rays changeup in his arsenal (.198 wOBA against it in 2014), Archer has done very well for a 3 starter, with a 3.15 ERA and 3.49 FIP over 286 innings since 2013.

Many have noted that Yankees’ starting pitcher Masahiro Tanaka has experienced so much success because he is one of the few pitchers who regularly throws a splitter in the MLB. Perhaps an organization can do what the Rays have done with the changeup and make the splitter a cornerstone of their pitching development. Obviously, such a plan comes with inherent risk. Making the splitter a more commonly offered pitch could take away some of its unfamiliarity-related effectiveness. Also, the splitter is believed to be very taxing on the elbow, a definite red flag given the recent wave of Tommy John surgeries. However, doing what the Rays did with the splitter could make it so that pitchers who are one additional plus pitch away from reaching their ceilings are safer to bet on.

The 2014 season has been weird for the San Francisco Giants. They began the year an MLB-best 42-21 (.667) and have gone 27-41 (.397) since. They led the N.L. West by 9.5 games on Jun. 8, but currently trail the first-place Los Angeles Dodgers by five games.

At 69-62 (.527), San Francisco leads the second wild card by one game over the Pittsburgh Pirates.

Marty Lurie, a host on the Giants’ flagship radio station, KNBR 680, says that a baseball season is like a mosaic: you can’t judge it by its individual parts, its moments, games, and plate appearances. Only when you step back and look at the big picture do things come into focus and make sense.

So, now that we’re about to enter the season’s final month (can you believe it’s September already?), it’s appropriate to look back on the season that has been and see how all the moments add up. That’s what baseball is all about.

It’s interesting (and fun) to look at a team’s overall numbers in some key areas, then find individual players whose career or single season statistics are comparable. Let’s get right to it:

2014 San Francisco Giants wRC+: 98

Notable hitters with a career 98 wRC+:

Rich Aurilia: .275/.328/.433, 7.2 BB%, 13.7 K%, .158 ISO, 23 SB, 6,278 PA

Delmon Young: .283/.317/.425, 4.2 BB%, 18.0 K%, .141 ISO, 35 SB, 4,143 PA

2014 San Francisco Giants starting pitcher FIP: 3.66

Notable starting pitcher(s) with a career 3.66 FIP:

Ben Sheets: 3.78 ERA, 7.47 K/9, 2.08 BB/9, 1.04 HR/9, .295 BABIP

Mike Krukow: 3.90 ERA, 6.07 K/9, 3.15 BB/9, 0.81 HR/9, .288 BABIP

Notable starting pitcher(s) with ~ 3.66 FIP in 2014:

Ryan Vogelsong: 3.68 FIP, 3.78 ERA 7.26 K/9, 2.58 BB/9, 0.78 HR/9, .299 BABIP

2014 San Francisco Giants relief pitcher FIP: 3.24

Notable relief pitcher(s) with a career 3.24 FIP:

John Smoltz: 7.99 K/9, 2.62 BB/9, 0.75 HR/9, .283 BABIP

2014 San Francisco Giants UZR/150: 0.0

Notable player(s) with ~ 0.0 UZR/150 in career:

Matt Holliday (0.0 UZR/150 spanning ~ 13K innings in LF)

Edgar Renteria: (0.2 UZR/150 spanning ~ 11K innings at SS)

As you can see, the Giants’ lineup this season (including the pitcher’s spot) has essentially been nine Rich Aurilias or Delmon Youngs, or any combination of the two. Having nine Delmon Youngs in your lineup (disregarding defense) is not the worst thing in the world, but it’s also far from the best. The potential for damage is there, but he’s going to let you down more often than not. If this sounds just about right for the Giants, that’s because the comps are accurate.

Next, the Giants’ starting rotation has been five Mike Krukows or Ben Sheets, or any combination of the two. Or it’s been five 2014 Ryan Vogelsongs. This means that Vogelsong is the typical Giants starter this year—he’s right in the middle of an up-and-down rotation.

The bullpen has been good. John Smoltz (in his career) is a pretty good comp to have for your bullpen as a whole in a season.

Lastly, the Giants defense as a whole in 2014 has been equivalent to how Matt Holliday plays left field or how Edgar Renteria plays shortstop. It’s possible to do worse, but it’s also possible to do a whole lot better.

Delving deeper into the Giants’ defensive issues, Michael Morse has an atrocious (and I mean atrocious) -24.6 UZR/150 in 577 innings in LF this season. His deplorable defense almost completely offsets his terrific 135 wRC+, as he’s been worth just 1.0 WAR this season.

Let’s take the comps a step further by looking at two elite teams in the N.L.:

The Dodgers’ 105 wRC+ this season means they’ve essentially had nine Ray Durhams in the lineup every night.

Durham’s career stats: 105 wRC+, .277/.352/.436, 9.7 BB%, 14.3 K%, .158 ISO, 273 SB, 8,423 PA

And the Dodgers’ 3.50 team FIP in 2014 means that their entire pitching staff has been Garrett Richards.

Richards’ career stats: 3.66 ERA, 3.50 FIP, 7.25 K/9, 3.07 BB/9, 0.63 HR/9, .288 BABIP

Even scarier, the Nationals’ 3.23 team FIP this season means they have been a staff of Curt Schillings.

Schilling’s career stats: 3.46 ERA, 3.23 FIP, 8.60 K/9, 1.96 BB/9, 0.96 HR/9, .293 BABIP

And Washington’s 1.5 UZR/150 team defense means they’ve collectively played as well as Justin Upton plays right field and Erick Aybar plays shortstop.

In summation, the Giants are a decent/pretty good MLB team, but they are clearly not as good as some other teams in the N.L. (and the A.L. for that matter) in some key categories.

On any given day, Ryan Vogelsong might pitch a shutout; Curt Schilling sometimes got rocked. Every now and then, Delmon Young goes 4 for 4 or hits a home run and a double; Ray Durham surely took his share of 0 for 5s. These things happen sometimes. That’s baseball.

But when you step back and look at the big picture, Schilling dealt, Durham outplayed Delmon, and Justin Upton made a fine running catch and throw while Matt Holliday just couldn’t quite get there in time.

In a recent article on the Community Research pages Andrew Patrick looks at how players and teams fare when they take more or fewer pitches per plate appearance (P/PA), the idea being that you will benefit if you tire the opposing pitchers, see more of their repertoire, etc.  While there is a small correlation there, you’d be hard-pressed to tell a batter not to swing at a hanging breaking ball just because it’s the first pitch.  Take, for example, Madison Bumgarner’s start on June 21, 2011.  The Twins only took 2.5 P/PA, but those PAs resulted in one strikeout and nine hits, chasing Madison after 0.1 innings.  Maintaining a low P/PA didn’t help Bum on that day because he failed to convert those PAs into outs. To contrast with another Giants pitching performance, Tim Lincecum’s 148 pitch no-hitter on July 13 of last year saw more than 4.6 P/PA, an extraordinary number.  But only 4 of those PAs (all walks) failed to make an out.  Clearly not making an out trumps taking pitches.

All of this leads me to a question.  If P/PA correlates weakly with performance, perhaps that’s because we would be better served by looking at pitches seen per out made (P/Out)?  I went ahead and ran those numbers for 2014 to compare the results with Andrew’s P/PA graphs.  For “outs made” I’m considering only outs made at the plate without sac bunts (i.e. outs considered in OBP).  P/PA is in red (left) and P/Out is in blue (right).

The winner (in P/Out) is Mike Napoli, and he has it in a walk, so to speak.  You can see that the correlation gets a bit stronger once you consider outs rather than PAs.  This makes good sense: the spread in P/PA averages is only about 25%…not that big.  By not making an out you do more to increase a pitcher’s count than by making a longer-than-average out.  In fact, let’s just ignore pitches for a moment and look at PAs per out:

The winners here are Stanton and McCutchen; a little more in line with what we expect when we think of good hitters.  While they came in 6th and 5th in P/PA respectively, their OBPs of  0.412 and 0.409 drive them to the top of the PA/Out list.  Overall we see an even better correlation and higher slope.  In turns out the more we focus on outs the more fidelity we get to batting outcomes.  This isn’t to say that seeing pitches isn’t important, but a great way to elevate a pitch count is not to get out.

But let’s change gears for a moment.  Let’s hypothetically stick Mike Napoli on a team with a bunch of free swingers.  The fact that he’s doing all he can to elevate pitch counts won’t really matter if he’s the only one on his team.  He may not see a wOBA benefit from his hard work.  But if the entire team is trying to wear out the pitcher we might see a synergistic effect that drives everyone’s wOBA up.  Here is the same data in the first graph, but for teams instead of players.  The wOBA are also park adjusted.

Impressively there is no correlation between a team’s P/PA and wOBA.  The Red Sox, despite sporting the highest P/PA in the league, have a dismal park-adjusted wOBA while the Brewers’ league-low P/PA leads to a league-average park adjusted wOBA.  You do get a small correlation once you consider P/Out, again demonstrating the supremacy of not making outs.  I won’t put up the graph, but if you look at PA/Out on a team basis you get yet a stronger correlation, as you might expect.

The takeaway from this is that you should take pitches if it helps you be a better batter, but that taking pitches in an of itself does not appear to do that.  It certainly doesn’t add up to anything at the team level.  If you think about it, this makes sense — in order to win a war of attrition it’s not enough to run up the starter’s pitch count.  He’ll still start again in 4-5 days regardless.  What you really want is to chase the starter early and run up the bullpen’s pitch count.  That way in subsequent games they will have fewer options to close out a game or back up a starter who’s having a bad day.  Of course, you may not actually reap the benefits of your hard work if you are near the end of a series.  You might actually expect that if there was any effect to the Red Sox’s patience it would be to help the team their opponents face in the subsequent series.

The ability to quantify the value of catcher framing has been one of the biggest sabermetric breakthroughs of the last decade. By parsing through PITCHf/x data, analysts like Mike Fast, Max Marchi, Dan Brooks, and Harry Pavlidis have managed to shed light on which catchers are adept at turning balls into strikes, uncovering hidden value in otherwise unremarkable players, including Rene Rivera, Chris Stewart, and of course, Jose Molina.

MLB front offices have taken notice. Several teams, including the Yankees, Rays, Red Sox, Pirates, Padres, and Brewers have begun hoarding good-framing catchers over the past few years. But one team that’s missing from this list are the Oakland Athletics, who have historically been among the first adapters of sabermetric principles. One would think that the A’s would be all over the Jose Molina‘s and Chris Stewart’s of the world, yet Billy Beane and co. seem to have missed the memo on acquiring good framers. In fact, they’ve made a habit of employing poor ones. According to Baseball Prospectus‘ model, A’s catchers rank fourth from last in framing runs saved this season. This isn’t a one year anomaly, either. Here’s a look at all of the catchers the A’s have used since 2010, along with their career framing numbers.

That right there is a pretty sorry group of framers. There’s not a single catcher in the group who’s even above average. So what gives? Why has Billy Beane — who’s nearly synonymous with the term “market inefficiency” — been so reluctant to exploit the latest market inefficiency?

As far as I can tell, there are two possible explanations, and the real answer is probably some combination of the two:

1) The A’s have chosen to employ catchers who excel in areas other than pitch-framing.

2) The A’s aren’t completely buying into all of this pitch-framing stuff.

Let’s start with the first explanation. Since 2010, A’s catchers have accumulated 12.1 fWAR (which doesn’t account for framing), putting them 15th out of 30 MLB organizations. But since 2012, the year after Mike Fast’s research first brought the value of pitch framing to the public’s eye, the A’s rank 10th. The average wRC+ from a catcher is 93, but the A’s have done much better than that of late by employing guys like John Jaso (136 wRC+) and Derek Norris (110 wRC+). Even if you were to dock the Oakland’s catchers for their poor framing skills, they’d still fall somewhere in the middle of the pack in terms of total value. Basically, the A’s have managed to find good, cheap catchers, who generate value in ways other than framing pitches. Plus, for all we know, the A’s might have reason to believe these guys excel in other overlooked areas. They could be superb game callers, for example.

But that can’t be all that’s going on. Sure, the A’s have done a decent enough job of finding catching talent without prioritizing framing, but it’s not like they’ve had Mike Piazza or Johnny Bench behind the plate. Jaso and Norris are fine players, but aren’t exactly superstars. Plus, it should tell us something that they haven’t even brought in any bottom-of-the-barrel framing specialists. Eric Kratz or Chris Stewart were both traded for warm bodies last winter, but the A’s instead chose to roll with Vogt as their primary catching depth.

Perhaps the A’s have reason to believe that publicly available framing models overstate the value-add of a framed pitch? As Dave Cameron recently pointed out, its not entirely clear if the full value of a framed pitch should be attributed to the catcher, with none of the credit going to the pitcher. Current models don’t account for how a pitcher might change his approach based on the framing abilities of his catcher, and research shows that pitchers do in fact change their approach based on who’s catching, throwing a few more pitches outside of the strike zone:

Oakland’s brain trust is about as progressive as they come, and have a proven penchant for unearthing value from unlikely places. When a team like that zigs while others zag, it probably makes sense to ask why. This isn’t to say that the publicly-available framing data is useless, as having a good framer undeniably adds some value, even if it’s only a few runs. But the fact that the A’s have yet to employ a single plus framer should lead us to wonder if there’s a piece of the puzzle we might be missing.

Statistics courtesy of FanGraphs and Baseball Prospectus.

Teams have both success and failure in quite a lot of ways, so I am playing with a way of showing what teams look the most alike.  To do this I have created a percent similar score as follows:

First I pulled team level WAR data split into what I am calling HWAR (position players/hitting) and PWAR (pitching) for all teams from 1947 to 2013.  I then converted each of those numbers into a percent above or below league average for that particular season.  For instance, the 2013 Rangers had 21.5 HWAR/19 HWAR league average minus one to convert to percentage, so they have an HWAR% of 13.1 or 13.1% better than average by cumulative war (actual HWARs above are not rounded in the data so it doesn’t round to 13.2% like it does in the example).  I did that for each team and also a PWAR% for each team in the same manner.

Next I compared each team to each other team with a giant 1610 by 1610 matrix, or a little over 2.5 million team pairs, to see how similar the teams were to each other.  The formula for this was 1/((1+ABS(HWAR%i – HWAR%j))*((1+ABS(PWAR%i-PWAR%j)), which gives a percent similarity based on nominal absolute deviation for each team from each other team multiplied together.  That way the deviations can’t cancel each other out and we are bounded between 0 and 1, and each team compared to itself will yield a similarity score of 100% as you would expect.

From this we can find some fun historic pairs, but also I will add 2014 YTD data and see who the best matches are for current teams and their results.  The two most similar teams out of the 2.5 million+ pairs were the 1999 Cardinals and the 2005 Nationals with a similarity score of 99.9%.  Both were slightly below-average teams.  The Cardinals were 15.5% below average by PWAR% and 9.6% below by HWAR%, and the Nats were 15.6 below and 9.5 below respectively.  That St. Louis team ended up going 75-86 on the season as we would expect from these numbers, but Washington managed to scrape by at an even .500 at 81-81.

On the other end of the spectrum, the least similar teams were the 1998 Braves and the 1979 Athletics.  That was a fantastic Braves team with PWAR 80.7% above league average and HWAR 97.5% above.  Meanwhile, the 1979 A’s were awful at 65% below average in PWAR and 151% below in HWAR, yes they had a negative HWAR as a team which is impressive if you like train wrecks.  These two teams had a similarity score of 11.7%, and their records show it.  That Braves team won 106 games and that A’s team lost 108 games, that is about as far apart as two teams can get.

There are some legitimately useful things I am planning on doing with these scores down the road, but for today I also thought it might be fun to see who is most like the 2014 contenders and how their respective seasons turned out.

The teams in the best probability for the playoffs have the best comps as you would expect with the exception of the Nationals who drew a very mediocre 83 – 79 team as most similar.  Baltimore had the only 100-game winner , but there are plenty of good teams in the mix like the Dodgers comp of a 95-win Expos team.  The different eras prevent us from seeing a ton of playoff outcomes, but none of the comparable teams made it to the World Series.  This year’s lack of any dominant teams might make that an expected outcome, even Buster Olney on the Baseball Tonight podcast today was discussing this very topic.  Of course everyone expected this year’s Detroit team to look like last year’s Royals.

Anyway, this could be a good way to create groups of historical comparisons for teams and the methodology could be broken out more if you want to separate defense, base running, bullpen vs. starters, which could all be done.  How you multiply them together to get appropriate weighting would be the sticky part with that.  It is a simple way to look at teams that had similar outcomes, and WAR allows us to control for ballpark factors and such.  I welcome any comments on other things you think could make it work better.

It’s the bottom of the fourth. No outs. Your beloved Milwaukee Brewers are up to bat trailing the Dodgers 1-0, with Clayton Kershaw on the mound. They’ve picked up two scattered hits and drawn a walk over four innings, but the sentiment in the dugout and the stands seems to read if they haven’t scored yet, chances don’t look so good.

Consider the same situation, now, with one small change. Your Brewers are still down by a run. It’s still the bottom of the fourth. Kershaw is still dealing. But it’s 2-1 Los Angeles this time. Milwaukee has still only gotten two hits and drawn a single walk, but the timing has worked out such that a run scored. By the numbers, things are almost exactly the same. No question about it. The sentiment, though, is certainly different. We’ve broken through once already, think the players, manager, and fans. We can do it again. Well, of course the Brewers can do it again. But, statistically speaking, will they? That is: when trailing by one run as they enter a half-inning, is a team more likely to come back in a non-shutout than in a game in which they haven’t yet scored?

The answer is “yes,” although only by what initially appears to be a small margin. In 2013, 5705 half-innings began with the batting team trailing by a run. 11.4% (651) of those half-innings ended with the batting team tied or in the lead. The same year, 2915 half-innings began with the batting team trailing specifically by the score of 1 to 0. 11.1% (324) of those ended in a lead change or tie.

At first glance, a 0.3% difference between odds of scoring when down by a run versus the specific case of being down 1-0 seems minor. And it is, really. For years with complete-season data available since 1871, the percent of half-innings started where it’s a one-run game and the losing team up to bat which resulted in a lead change or tie (let’s call this %ORLC) averages out to 11.5% ± 1.3% (1 σ). The subset of these in which the batting team was being shutout (let’s call this %ORSLC) has an average of 10.6% ± 1.1% (1 σ). Middle-school statistics will tell you that while, yes, %ORSLC is on average nearly a percent lower than %ORLC, they fall within a standard deviation of each other and, thus, their difference is not statistically significant.

That’s true. But baseball isn’t middle-school statistics and two subsets whose error ranges overlap are not for all practical purposes equal. Quite remarkably, %ORLC has exceeded %ORSLC for each consecutive season of Major League Baseball since 1977 (when %ORSLC was 0.2% higher) and every year since 1871 except for five seasons (out of the 111 years of complete-season data that were available).

That is: in 106 out of the last 111 seasons for which box scores have been logged every game, a batting team behind in a one-run ballgame has successfully erased the deficit more often when not trailing 1-0. The margin isn’t huge, of course, but the trend is meaningful.

View post on imgur.com

Above: Percentage of one-run game situations and specific 1-0 game situations (%ORLC and %ORSLC, respectively) in which the team losing scores to tie or take the lead

After all, baseball is a game of small but meaningful margins. The 111-year average relative difference between these two metrics (10.6% vs 11.5%) is proportional to a .277 batting average versus .300, or 89 wins in a 162-game season instead of 97. The latter is perhaps a more relevant comparison, since it is gaining (and maintaining) a lead that is crucial to winning games.

Among teams in 2013, however, these differences aren’t so marginal. In %ORLC (percentage of half-innings in which a team trailing by a run ties it up or takes the lead) the Royals finished first at 16.7% and the Cubs finished last at 6.5%. In %ORSLC (same stat but for the score 1-0), the Rays finished first at 16.7% (same number, coincidentally) and the Red Sox finished last at 4.9%. Considering the Royals didn’t make the playoffs in 2013 and the Red Sox won the World Series, I wouldn’t use %ORLC and %ORSLC as indicators of a team’s ultimate success unless you’re looking to lose a lot of money in Vegas.

While one could theorize for hours on the meaning and utility of each made-up statistic, it sure doesn’t seem like %ORLC and %ORSLC are indicative of much on a team-by-team basis. But that doesn’t mean they’re useless. Let’s go back to the long-term trend of %ORLC and %ORSLC, where the former was higher than the latter 106 out of 111 times.

Some underlying process, it would seem, must be responsible for this impressive stat. If we are to believe that teams truly underperform, ever so slightly, when they’re losing 1-0 due only to the fact that they’re being shut out, shouldn’t we able to see the effect of psychology on performance somewhere else?

As it turns out, you don’t have to look far. Let’s consider the general situation of a team coming up to bat down by a run (not only the specifically 1-0 case), which is colloquially termed a “one-run game.” We’ll abbreviate any instance of this (a trailing team coming to bat in any half-inning) as OR. Now this situation could happen at any point in a game. A visiting team leads off with a run in the top of the 1st, the home team comes up to bat – that’s an OR. It’s all tied-up in the top of the 13th, the third baseman slugs a solo shot to left, three outs are recorded, the home team steps up the plate with one chance to stay alive – that’s an OR. So, in what inning on average does an OR occur?

In 2013, the answer was the 4.95th inning. In 2012 and also for the last 111 years of available records, the 4.91st inning. Baseball amazes us once again with its year-to-year consistency in obscure statistics. But this obscure stat isn’t all that meaningful on its own. Okay, so most one-run situations occur near the 5th inning – so what?

Well, let’s take a look now at the average inning in which a team scored in an OR to tie or take the lead. We’ll call this a one-run game situation where the lead changes, or ORLC. In 2013, of all the instances of ORLCs, the average time they occurred was the 5.18th inning. In 2012, the 5.10th inning. And for the same 111 seasons of recorded game data, the 5.20th inning. Once again, we see a marginal but nonetheless compelling deviation from the average, just as we saw with %ORSLC. Teams score in one-run situations about a third of an inning later than the one-run situations tend to occur themselves. That may not seem like a whole lot, but consider that in our 111-season dataset only two years – 1902 and 1912 – saw earlier ORLCs than ORs on average. Just two years in one-hundred eleven.

View post on imgur.com

Above: Average innings of occurrence for one-run game situations (OR) and one-run game situations in which the trailing team scores to tie or take the lead (ORLC)

So what’s going on? I like to think of average ORLC minus average OR as a league-wide statistic for urgency. Consider the following: if the inning number had no effect on the performance of a trailing team in a one-run situation, then we would see roughly the same average inning of occurrence for both OR and ORLC. Out of 111 years, we’d expect to see about 55 years in which OR occurred earlier on average than ORLC and around 55 in which it didn’t. But we don’t see this at all, which strongly suggests that inning number has an effect on how a team does at the plate when down by a run. This is the urgency statistic. It describes a trend that has rung true for the past 101 consecutive seasons of Major League Baseball – when time is running out and the 9th inning is rapidly approaching, teams in close games get their acts together and produce runs. Not every time, of course, but we’re speaking in averages of massive sample sizes here.

So, while your Brewers are likely to fare worse trailing Kershaw and the Dodgers 1-0 than 2-1, take solace in the fact that it’s the fourth inning. Statistically speaking, they’ll have a better chance breaking through as the game goes on and their need for a run becomes more urgent. The effect of team psychology has left its imprint on the records of baseball games since the sport’s earliest days.

As I noted here on August 15, major league batters are being hit by pitches at rates not seen in over a century (measured by HBP/game). I offered data illustrating this but was at a loss explaining it. Fortunately, I spent the following weekend at the Saber Seminar in Boston, surrounded by a bunch of really smart people, so I posed the question to them.

To be clear, everyone was surprised by the conclusion. Through Sunday’s games, there were 67.4 hit batters per 100 games in 2014. That’s the highest rate since 2001-2008 which, in turn, is the highest since the two leagues were formed in 1901. (Note that these numbers are different from the ones in my original post. When I downloaded league totals from FanGraphs, I hadn’t realized that Games referred to player games, not team games. So I was using a denominator that was too large. The conclusions still hold. I’ve updated the figures in a comment to the August 15 post.) If you didn’t notice this spike in hit batters, join the club. This appears to be an entirely under-the-radar trend.

Asking around, I got several possible explanations. Dave Cameron, FanGraphs managing editor, suggested that since PITCHf/x has clearly demonstrated that left-handed hitters suffer from an elongated strike zone on the outside part of the plate, lefties might be setting up closer to home in order to reach those outside “strikes.” That would make them more likely to be hit by inside pitches.

To test this, I looked at data from 2010-13, when there were 63.8 hit batters per 100 games, and compared them to 1980-83, when there were just 32.6. Switching from HBP/100 games to HBP/1000 plate appearances (since games contain a combination of left- and right-handed batters), the hit batter rate went from 4.3 in 1980-83 to 8.3 in 2010-2013. Right handed hitters got hit at a rate of 4.6 per 1000 plate appearances in 1980-83 and 9.0 in 2010-13, an increase of 95%. For lefties, the HBP rate went from 3.8 to 7.3, and increase of…95%. Exactly the same. Handedness hasn’t been an issue.

Former major league pitcher Brian Bannister suggested that I correlate HBP rates to measures of power. He noted that he didn’t like to come inside on sluggers, for fear that they’d pull the pitch down the line and into the bleachers. I thought this was a sharp, counter-intuitive insight: A rash of longballs makes pitchers work away rather than come inside. With offense in retreat in recent years, pitchers are more willing to pitch inside, and when they miss, the batter gets hit.

I looked at three measures of power: home runs per plate appearance, slugging percentage, and isolated power. I correlated these metrics to hit batters per game for the period 1980-2013. If Brian’s hypothesis is correct, there should be a negative correlation–as power increases, hit batters decrease. However, the opposite was true: 0.84 correlation coefficient between hit batters per game and homers per plate appearance, 0.55 for slugging percentage, 0.65 for isolated power. Maybe my endpoints were wrong? I checked 1970-2013 and got pretty much the same results: Correlation coefficients of 0.84 for HR/PA, 0.66 for SLG, 0.73 for ISO. I was ready to think that maybe hit batters are a result of more power, not less, but then I picked 1990-2013. At least during those 14 years the correlations were weaker, coming in at 0.82 for HR/PA but 0.25 for SLG and 0.43 for ISO. That’s still consistent with the observation that hit batters have remained high in the post-PED era. I don’t see a strong case for fingering the long ball as a cause for hit batsmen, one way or the other.

SiriusXM radio host Mike Ferrin thought we may be seeing a cultural shift of sorts. In college ball, he pointed out, batters view getting hit as an on-base weapon. Might an influx of college players be driving up HBP rates?

Unfortunately, neither the FanGraphs Leaders board nor the Baseball Reference Play Index have college vs. non-college splits, so I looked at the most-plunked batters in 2013 and 1983. In 1983, players with four or more HBP represented the top 53 overall and collectively comprised 274 of 717 HBP that year, or 38%. Of those 53, 30 (56%) attended a secondary school in the US. (I am going to use “attended college” instead of “attended secondary school” going forward, but I mean players who went on for any schooling, including junior college, following high school in the US.) They were hit 10.2 times per 1,000 plate appearances. Players who didn’t go to college were hit 10.4 times per 1,000 plate appearances. That’s our baseline: No evidence of college kids leaning into pitches the year “Every Breath You Take” and “Billie Jean” were top hits.

Now, 2013. There were 15% more teams than in 1983. As it happens, there were 61 hitters with seven or more HBPs in 2013, and 61/53 = 1.15, so 61 is the appropriate sample size for consistency. Those 61 batters were hit 587 times, 38% of all HBP, just like our sample for 1983. Here are the relevant metrics:

• Percentage of most-hit players who attended college: 57% (30 of 53) in 1983, 49% (30 of 61) in 2013
• Percentage of HBPs by most-hit players that were players who attended college: 56% (167 of 298) in 1983, 47% (275 of 587) in 2013
• HBP per 1,000 plate appearances, all most-hit players: 10.4 in 1983, 18.2 in 2013
• HBP per 1,000 plate appearances, most-hit players who attended college: 10.2 in 1983, 18.5 in 2013
• HBP per 1,000 plate appearances, most-hit players who didn’t attend college: 10.8 in 1983, 17.9 in 2013

Mike has a point. College players appear to be getting hit more, relative to other hitters, than they were in the past. The rate of HBP per 1,000 plate appearances increased 82% over 30 years for batters who went to college. However, the HBP rate for batters who didn’t go to college was up 65%, which is also pretty dramatic. And the list of HBP leaders has more players who didn’t go to college than in 1983. So while college kids may be bringing a lean-into-the-pitch ethic to the plate, there is still strong evidence that players who didn’t attend college are getting hit more, and the limited data I used don’t indicate that college kids are comprising a growing percentage of plate appearances.

Some of the commenters on my post from the 15th suggested that maybe HBPs are up because pitchers are throwing harder, giving batters less of an opportunity to get out of the way of an errant delivery. Per FanGraph’s PITCHf/x data, average fastball velocity has climbed from around 91 mph in 2007-9 to 92 mph in 2013-14. That’s a pretty tiny difference from a hitter’s perspective (about five milliseconds, or 0.005 seconds, over 60.5 feet), but it’s something. I’m not ruling it out.

Last Wednesday on the Effectively Wild podcast, Baseball Prospectus’s Sam Miller noted that the rate of batters reaching base via catcher’s interference is near all-time highs. (And you thought hit batsmen per 1000 plate appearances was obscure…) He hypothesized that one of the reasons is that batters are standing further back in the batter’s box in order to get extra time — maybe like five milliseconds? — to identify and swing at an incoming pitch. By setting up deeper in the box, batters increase the possibility that their bat may hit the catcher’s glove at the end of their swing, drawing the catcher’s interference call. If that’s correct, wouldn’t moving back also give pitches that break horizontally — two-seam fastballs, sliders, cut fastballs, some changeups — more time to drift into the hitter? It makes sense!

Unfortunately, the numbers don’t back this up. The correlation coefficient between catcher’s interference and hit by pitches is 0.00 since 1962 (expansion in both leagues), -0.10 since 1969 (divisional play), and 0.07 since 1994 (three divisions per league). That doesn’t necessarily mean that the increase in hit batsmen isn’t caused by batters positioning themselves toward the back of the batter’s box, but it does say that whatever’s driving catcher’s interference isn’t the same thing that’s driving hit batters.

So basically I’m back to where I was going into the Saber Seminar. We’re seeing batters hit by pitches at rates not seen in a century. This change has not been widely reported, and I haven’t identified an obvious underlying cause. After talking to people at the Seminar, I still don’t have a great explanation. It could have a little bit to do with fastball velocity, or batter positioning, or players who went to college being willing to get plunked. But I haven’t identified a clear reason thus far. Even with smart guys helping me.

I think I might be the biggest fan of the world of the Ten-Pitch Walk.  I don’t know why, but I get overly excited when I see a player really battle for a long time, against everything the pitcher has, only to win the battle through patience.  Perhaps it’s because it’s so contrary to the spirit of what’s actually exciting about baseball; seeing players run around and field a batted ball.  It’s wholly a battle of attrition.  It’s the baseball equivalent of watching somebody run a marathon; you may not think the act itself is exciting, but it’s certainly an impressive feat in a vacuum.

So this has also lead to a fascination with pitches seen per plate appearance.  I’ve long wondered if certain teams place an emphasis on teaching their players to see more pitches per plate appearance.  It seems fairly self-evident that seeing more pitches is, in a microcosm, better than seeing fewer pitches.  You tire the pitcher out quicker, you see more data for your next at-bat to work with, and you give your team a chance to see what the pitcher has, and how he’ll react in different situations.  I hypothesized, purely based on colloquial wisdom, that the A’s would be good at this and the Blue Jays would be bad at this.  That’s not to say that one approach is better than the other, but just that some teams seem more patient than others.

Fortunately, FanGraphs has data available per hitter as to how many pitches they see.  I pulled that data out and found out each player’s average pitch per at bat since the year 2003 (the earliest we have this data, from what I can tell) and restricted the findings to active players only.  Then I ran some regressions to see if there was any correlation between pitches per at bat and useful batting stats.  Here’s what I found:

We see a slightly positive correlation between P/PA and wOBA.  It’s not really anything to write home about, but it’s more than negative.  It doesn’t seem immediately that seeing more pitches relates heavily to overall performance at the plate.  What about on base percentage?

Slightly better here, but still not great.  Seeing more pitches does have a little more correlation to getting on base, but there are plenty of aggressive swingers that don’t follow that model, so it means the correlation is loose at best.  What if we talk just about taking walks?

Here we have a real correlation.  .59 is a fairly strong correlation, and that makes sense.  The more pitches you see, the more likely you are to take a walk.  If you can successfully foul off anything in the strike zone, you will eventually walk (or the pitcher will die of exhaustion, either way, you win).  This is reasonably useful.  If you’re trying to find a way to make your team walk more, maybe you can invest in some players that see more pitches per plate appearance than normal.  This strong of a correlation makes me think about strikeout percentage too, though, because every pitch you foul off makes you closer (or just one whiff away) from striking out.

There is a positive correlation here, but not nearly as strong as between BB% and P/PA.  It’s stronger than the other useful stats like wOBA, but it’s interesting to know that seeing more pitches relates much more strongly to taking a walk than it is to striking out, at least on a grand scale.  There is some research to be done here to see what the odds are of a plate appearance as the pitch count increases, but I’ll leave that for another day.  My next thought was to see if there are, in fact, any teams that are better at this than other teams.  Here’s what we’ve got on a team level:

Well, my original hypotheses were not great ones.  The A’s and the Blue Jays, at 11 and 13, are both decidedly middle of the road teams.  I find it most fun in times like this to look at the extremes; in this case, the Red Sox and the Brewers.  The difference in pitches seen per plate appearance between these two teams is 0.38.  That may seem small, but it adds up.  If we assume the average pitcher faces 4 batters per inning, that’s an additional 1.5 pitches per inning, and 9 pitches by the end of the sixth, just purely by the nature of the hitters.  In a tightly contested contest, that may mean the difference between getting to the bullpen in the 7th rather than the 8th, or even the 7th rather than the 6th.

It should be noted that I limited this data set to 2014 (in contrast to the earlier data which was 2003 onwards) just so we could get a realistic look at roster construction, and to see if any teams are, right now, putting any particular emphasis in this area. The BoSox are carried by the very patient eye of Mike Napoli (4.51 P/PA), but hurt by the rather hacky eye of AJ Pierzynski (3.42 P/PA). Even on one team, that’s more than a pitch per plate appearance, which is pretty startling. The Brewers don’t have nearly the same difference; their best is Mark Reynolds with 4.04 P/PA and their worst is Jean Segura with 3.42 P/PA. As an aside, Chone Figgins is by far the best in this with a whopping 4.99 P/PA, though it was in just 76 PA. Kevin Frandsen brings up the rear with 3.16 P/PA in 189 PA. A lineup of all Mike Napoli’s would see 24.3 more pitches than a lineup of Kevin Frandsens before the leadoff Napoli even comes up a third time. I would feel bad for that pitcher.

The talk about teams possibly emphasizing this data made me wonder if I could make a huge difference if I compiled a team solely to do this; just make sure the pitchers throw a ton of pitches.  With that, I present to you the 2014 All-Stars and Not-So-All-Stars in this area, with a PA minimum thrown in to eliminate Figgins-like outliers:

Despite the fact that there isn’t a strong correlation between wOBA and P/PA directly, it’s worth noting that the P/PA All-Stars are significantly better than the Not-So-All-Stars. Their difference in wOBA is .328 as compared to .314. The Not-So-All-Stars certainly present a fine lineup though; the All-Stars just have the benefit of having Mike Trout in their lineup. It’s nice to know that this is one other area that Mike Trout simply is amazing at, confirming the obvious. The All-Stars have a collective P/PA of 4.26, while their counterparts sit down at 3.43. That’s .83 pitches per plate appearance, which over the course of two turns through the lineup is 14.94 pitches; that’s definitely something notable.

So, it appears this is a demonstrable skill with some value, though not a ton. We can see that some teams are better at this than others, and we see some positive benefit from this, most notably in walk rate. While we see plenty of players on both sides of the scale who are excellent ballplayers, the data does seem to suggest that seeing more pitches is better than not doing so, though only marginally on a league wide scale. When we isolate leaders in this area vs. those more aggressive, we can see some startling differences though, suggesting that perhaps there is an advantage to be gained here.

I started writing this post prior to the trade deadline; viewed through this lens, the Lester-and-Gomes-for-Cespedes trade makes even more sense for the A’s than it already did, especially considering their parallel acquisition of Sam Fuld from the Twins.

The A’s are ahead of the curve again.  This time it’s not just about better overall player evaluation (concentrating on certain metrics that other teams undervalue), but about building a roster that maximizes each player’s skill set to get the most out of the talent on your roster.  I wrote a while back that WAR is not the be-all, end-all of player evaluation, emphasizing that there is more to Wins than WAR.

WAR is great at certain things; it’s useful to remove factors that are outside a player’s immediate control: park factors, sequencing, etc.; it’s useful for comparing players across eras by controlling for run scoring environments and translating Runs to Wins; it’s also useful because it encompasses multiple aspects of a player’s skill set (hitting, defense, baserunning), and uses the same units (Runs and Wins) to combine these into a single number.  It’s a nice package.

But if I’m a GM, I don’t evaluate each player in a vacuum.  I want to know how he fits into my system, my lineup, my park, etc.  A given player will bring different value to different teams.  Some examples:

• Certain players might be more tailored to certain parks based on their offensive profile.  A contact hitter who hits a lot of infield singles and steals a lot of bases isn’t worth as much (compared to a team playing in league-average conditions) to a team filled with roided-up sluggers playing in pre-humidor Coors field – the value of those stolen bases and infield singles just isn’t as high.  WAR does normalize for park factors, but it assumes all players are affected by a given park equally, which on its face isn’t true.
• A player’s contribution varies based on how his team uses him.  If a team platoons a player so that he often has the advantage, his offensive contribution (per plate appearance) will be increased, whereas if he faces a more standard distribution of pitchers, his contribution would be lower.  Likewise if he plays a position he’s not as used to for the good of the team, his own contribution (as measured by WAR) will be less than if he plays his primary position.
• Likewise, defensive versatility has value to a team.  A player who can play multiple positions allows his team more flexibility in roster construction and in-game management; setting the daily lineup, platooning, and late-game substitutions (matchups when pinch-hitting, or defensive replacements – especially double-switches in the NL).

As a GM, you don’t just add up each player’s projected WAR (and add in the replacement-level constant) and say that’s how many you project to win that year.  There are all kinds of interrelated variables at play that will determine how your team performs.

The A’s are the epitome of this philosophy and appear to be better at this optimization of roster construction.  They’ve loaded up on defensively versatile players with outsized platoon splits and are the king of the platoon.  They’ve started doing this in the last few years, and this year even more so.  Take a look at MLB averages for platoon splits as compared to the A’s:

League:

A’s:

We notice two things: first, the A’s splits are a bit wider than the league splits: their righties hit better against lefties by about the same split as righties league-wide, but their lefties really hammer righties: a .171 OPS split for A’s lefties, as compared to a 0.068 split for lefties league-wide.  They’ve made a conscious effort to go after this style of player.  Second is the distribution of plate appearances: the average team gets 55% of its plate appearances with the platoon advantage.  The A’s get 70% of their plate appearances with the platoon advantage.  They’ve constructed their roster in such a way that they can alter their day-to-day lineup as much as possible to maximize the platoon advantage.

What allows the A’s to do this?  Defensive versatility (and the DH).  They’ve got guys like Brandon Moss playing LF/RF/1B/DH; Craig Gentry playing all OF spots; Stephen Vogt playing RF/C/1B; Alberto Callaspo, John Jaso, Josh Donaldson, and Bud Norris dividing time between DH/1B/3B/C; and so on.  All this versatility allows them to mix and match their lineup to get as many plate appearances with the platoon advantage as possible.  And, by not being pulled down by having any full-time DH, they get additional flexibility.

Cespedes didn’t really fit in with this philosophy.  Nearly all his appearances came in LF – 343 PAs, compared to 17 as a CF and 69 as a DH.  With the exception of Donaldson (423 PAs as 3B), he had the highest concentration of PAs at a single position.  The next-highest was Crisp, a switch-hitter, with 306 PAs as a CF.  Everyone else is playing all over the field.

Cespedes has a bit of a platoon split (0.844 OPS vs. 0.765), but not as much as other A’s like Reddick (0.843 vs. 0.398 this year), Donaldson (1.098 vs. 0.704), or Norris (1.031 vs. 0.771).  Gomes’ platoon split: 0.875 vs. 0.722.

So maybe the A’s think there isn’t that big a difference between Gomes and Cespedes, especially considering that Cespedes’ defense can be partially replaced by Fuld’s, his performance against lefties can be replaced by Gomes, and his performance against righties can be replaced by the left-handed Vogt, who stands to get more appearances in LF now.  If they play their cards right, Fuld/Gomes/Vogt is a better player than Cespedes.

The A’s appear to have a leg up on the competition.  Rather than evaluating players in a vacuum and estimating “How many wins we will get if we add player X and remove player Y?”, they’re looking at “What does our lineup look like with player X?”  “How will his presence affect the number of plate appearances players A, B, C, and D get (with platooning taken into account)?”  “How will our various defensive alignments look?”  “How does his presence affect the availability of late-game pinch-hit and defensive replacement options?”  And for each of those questions, they boil it down to the impact on expected runs, expected runs allowed, and expected wins.  They’re all-in for this year, and they’re pulling out all the stops to optimize their lineup.

Next up, I want to look at whether there are any signs of the A’s trying to get a similar edge based on:

• Park factors – targeting players who fit in with their park
• Clutch hitting ability; the A’s lead the league in the split between hitting with runners on base vs. with the bases empty; why?  Is it just luck, or have they found a way to get players who are better at hitting with runners on base?