Managing with Analytics in Mind


For a long time now, I’ve wanted to write an article about what I would want in a major league baseball manager if I were hiring one. Of course, I want a manager to be all of a student, teacher, and leader of the game of baseball. I want a manager who will make bonds with his players. I want someone who will take charge when needed (for example, making a player take a day off when he needs one even though the player wants to be in there every day, etc.). I want someone who will keep the clubhouse positive in order for players to ideally thrive off of that positive energy. I want someone who will address what the team does poorly and try to put energy towards fixing those flaws - or at least masking them to the best of his ability. I want someone who trusts his other coaches and adequately delegates responsibility when needed.


I also want someone who will push players to become better at their craft through any way possible. One of the latest frontiers of doing so is looking at both game management and teaching/refining of important concepts with a much more analytical focus in mind. The format of the rest of this post, with analytics in game management and teaching in mind, will be as follows...


What this post is:

  • A snippet of things my manager for next year’s Cardinals would know about statistical data and how to not only make use of that data but convey it to players. (didn’t understand the “what it is trying to show” part. What what?)

  • A comparison of “old school” to “new school.” There is much more to compare than to contrast if you know the lingo. It’s not a whole new game they’re trying to teach.

What this post is NOT:

  • A post about what the game has become, bemoaning OR celebrating the new times compared to old.

  • An attempt to drag any current or former managers through the mud.

Offensive Concepts that “my manager” needs to know:


Lineup Construction - Gone are the days of Whiteyball. The whole point of the game in today’s day and age is to avoid outs. Outs are the currency of a baseball game. You only get 27 of them. There is no clock. The offense doesn’t possess the ball. You have to simply keep the ball rolling. The new book on lineup construction says that your best five hitters should hit in the top five in the order, no matter what. The three spot - in the past seen as the spot in the order you put your best overall hitter - is actually mathematically the LEAST important of the top 5 spots. It comes up the most often with nobody on base and two outs. That means the least impact potential, mathematically. Your best on base guy (with speed if possible) should hit #1, unless he’s your best hitter. Your best hitter hits #2. Your next best hitter with power should hit #4. Then #5. Then #3. In the National League, it is quite possible that batting the pitcher eighth and moving your worst hitter to ninth is the best option. That worst hitter is still better than the pitcher and has a better chance to be on base for the hitters at the top of the order.


For example, let’s say you have the following eight players (numbers listed are the league average batting line at each position) to put in the lineup, this is how I would arrange them:

  • Right Fielder - .257/.333/.435/.768 - 60 xbh, 10:4 SB:CS

  • Third Baseman - .259/.332/.439/.771 - 61 xbh, 6:3 SB:CS

  • Shortstop - .260/.320/.418/.738 - 57 xbh, 14:5 SB:CS

  • First Baseman - .251/.332/.438/.770 - 60 xbh, 4:2 SB:CS

  • Left Fielder - .259/.330/.431/.761 - 58 xbh, 10:4 SB:CS

  • Second Baseman - .255/.318/.401/.719 - 53 xbh, 11:4 SB:CS

  • Catcher - .234/.306/.379/.685 - 47 xbh, 2:2 SB:CS

  • Pitcher

  • Center Fielder - .252/.317/.403/.720 - 52 xbh, 19:6 SB:CS

The old guard would have probably gone CF, SS, 3B, LF, RF, 1B, 2B, C, P instead. That’s okay. It worked when you had an Astro-turf field and the team had little to no power to help create runs in other ways. That's just a difference in today's game, I believe. It's not to say that this lineup would have been better in the 80s in old Busch Stadium.


Also, “my manager” would know when sample sizes tend to stabilize. That way, the manager would know when to sit a slumping established player for a younger player who is currently accomplishing more either at the plate, in the field, or both.


Launch Angle and Exit Velocity - Did your coaches ever tell you growing up to “hit the ball hard” or to “try to hit line drives” or to “get some loft on your swing” or to “shoot the gaps?"


Launch angle isn’t just hitting the ball in the air higher and higher and swinging for homers. It’s not. I promise you. Here’s the very lame (by comparison) definition of Launch Angle, per Baseball Savant: “Launch Angle represents the vertical angle at which the ball leaves a player's bat after being struck. Average Launch Angle (aLA) is calculated by dividing the sum of all Launch Angles by all Batted Ball Events.”


Every single ball hit has a launch angle. Even balls hit at -20 degrees, down off the bat. Exit Velocity is defined, again by Baseball Savant, as “the speed of the baseball as it comes off the bat, immediately after a batter makes contact. Average Exit Velocity (aEV) is calculated by dividing the sum of all Exit Velocities by all Batted Ball Events.” The best type of “batted ball event” is defined as a “barrel.”

“The Barrel classification is assigned to batted-ball events whose comparable hit types (in terms of exit velocity and launch angle) have led to a minimum .500 batting average and 1.500 slugging percentage since Statcast was implemented Major League wide in 2015.”
http://m.mlb.com/glossary/statcast/barrel

I’d be remiss to fail to mention to you to listen to the 7-9 minute mark of this conversation with Birds On The Black minors guru Kyle Reis interviewing Cardinals prospect Andy Young in regards to Launch Angle. Young was a player who in college had 74 extra base hits in 740 career plate appearances. He literally had 1 XBH per 10 PA. In pro ball, he dropped to 1 every 16.1 PA in his first season at rookie ball and low A. According to what was said in Kyle’s interview, they retooled before last season and since then it has jumped to 1 every 12.4 PA.


“My manager” would believe that Launch Angle should be taught in a different way to every player because not every player is capable of the same things offensively. A player like Dexter Fowler, who has a career exit velocity of 86.9 mph against a league average of 87.3 mph, cannot do the same things as a guy like Matt Carpenter, who in 2018 has an average exit velocity of 90.4 mph. A 3.5 mph difference off of the bat may not seem like a lot, but consider a (randomly chosen) 10 degree launch angle. The difference between an 87 mph exit velocity and 90 mph exit velocity is:

  • 39 points of batting average

  • 88 points of wOBA

  • a 50% rise in the probability of hitting a double.

“My manager” would teach Fowler that:

  • A 10-19 degree launch angle on an 87 mph batted ball produces wOBAs of over .500 - very elite level.

  • A 12-17 degree launch angle on an 87 mph batted ball produces over a .750 wOBA.

  • A 14-16 degree launch angle on an 87 mph batted ball produces over a .900 wOBA.

“My manager” would teach Carpenter that:

  • An 8-17 degree launch angle on a 90 mph batted ball produces wOBAs of over .500 - very elite level.

  • An 11-16 degree launch angle on a 90 mph batted ball produces over a .750 wOBA.

  • A 13-14 degree launch angle on a 90 mph batted ball produces over a .900 wOBA.

So, my manager would teach Carpenter to try to hit it at a slightly lower trajectory than Fowler, if at all possible. For reference, Carpenter has averaged a 19 degree launch angle this year, down from 21.0 and 21.4 averages the past two years (reverse chronological). Fowler’s average launch angle in 2018 has been 12.5 degrees. He should actually try to raise his slightly, whereas Carpenter should try to be lowering his by a little bit.


Another example that is possibly easier to see because the exit velocity numbers are much closer is that Marcell Ozuna actually has hit the ball harder on average than Carpenter, 91.3 mph to 90.4 mph. However, Ozuna’s average launch angle of 10.1 degrees is much different than Carpenter’s 20.9 degrees. Their current difference in extra base hits has Carpenter walloping Ozuna with a 66 to 28 edge (at time of this writing). It’s not because of exit velocity.


Stolen Bases and Hit and Runs - Old school managers were all about the hit and run and stealing bases. I am not against that at all, but I’d like to put in a caveat. “My manager” would not be as worried about stolen base totals as he would be worried about taking advantage of worthwhile base running opportunities - net stolen bases (SB-CS), taking good lead offs, getting even better secondary leads to avoid double plays, and advancing on as many bases as possible. Simply put, we are risk averse in that we would avoid outs like the plague, but still be as aggressive as possible in the process.


Fangraphs, back many moons ago (2011), put up a post about break even points for stealing bases. The general consensus is that no matter how many outs there are, you generally have to steal 2nd base successfully between 70-75% of the time in order for it to be worth it. Stealing third base was a different story, with it ranging from 69-88% in different out-states. Stealing home was more drastic. With double steals (2nd and 3rd), the rates fell between 60-76% in their analysis. In 2014, another study by Beyond the Box Score showed that while 70% is commonly used as the break-even point overall, 80% of stolen bases (in their study) were not necessary in run creation. This leads me to believe that stolen bases are not nearly as important as simply being good runners on the bases - taking the extra base as often as possible and making as few outs on the bases as possible.


Bunting - An old school manager wanted a guy in the 2nd hole who could “handle the bat” because he was banking on his leadoff man getting on base and being able to run - not to mention that the old school manager was more worried about speed than OBP potential in the leadoff spot, which was quite counter-intuitive when players had trouble getting on base. You can’t run if you can’t get on base.


He would then use the 2nd hitter to bunt the guy to 2nd or, if the guy was able to steal second, bunt him to 3rd. Bunting is not inherently a bad thing. Giving up outs usually is. Please allow me to explain. Tom Tango created a Run Expectancy Matrix, based upon quite literally decades of seasons worth of data. Those tables can be hard to understand the first time around. Right now, let’s focus on the 2010-2015 table in the top set of tables, shown below:

Average # of runs scored, from that base/out state, to the end of that inning.

That is how many runs score at that base-out state in the game. So the green box is how the inning starts, nobody on and nobody out. In those six years worth of data, the league averaged 0.481 runs per inning. If the first player singles, you go down to 0 outs, runner on first. In that case, the league averaged 0.859 runs per inning. If you bunt him over to make it runner on second with 1 out, the number of runs per inning a team averaged at that point goes down to 0.664 runs per inning. Basically, you’ve given away an average of nearly 0.2 runs if you do that. Early in a game, I advocate almost never bunting - unless Jon Lester or the 2006 Detroit Tigers are on the mound to throw the ball down the right field line. Here’s where it gets interesting and the reasoning behind me saying that “bunting is not an inherently bad thing.” Here is the second set of charts.

Chance a run will score at some point in the inning, from each base/out state.

This second set of charts is the decimal version of the percentage chance (so just multiply by 100 to get the percent) that at least ONE run will score in the inning. Let’s again look at 2010-15 and say that there are zero outs with runners at first and second. The percent chance one of those guys will score is 61%. If you bunt them over to 2nd and 3rd with 1 out, the percentage chance goes UP to a 67.6% chance of one of those runners scoring. Late in a game, with a very finite and shrinking number of outs remaining to attempt to score runs, attempting to ensure that you get at least one run across the plate can be the difference between one in the W column and one in the L column. “My manager” has to know this.


Contact rates/high strikeouts - This kind of goes along with the launch angle/exit velocity conversation. An old school manager might just say to put the ball in play - strikeouts are bad. Well, just like bunting, strikeouts aren’t necessarily a bad thing. They’re statistically much better than some outs (double plays, triple plays, some force outs, some fielder’s choices) but they also are very rarely productive outs either. “My manager” would very much preach damage with the bat - each player learning their preferred launch angle based on how hard they hit the ball. Hopefully, they get better at hitting the ball harder more often and thus will need to see how their preferred launch angle would need to change. As that process is occurring, you should be learning how to best make contact with the ball, thus lowering the strikeout totals. It doesn’t always work this way, obviously, but that’s the theory.


Your best chance to do damage is when you’re ahead in the count, so that is when you’re looking for your pitch to drive, still. That is just like the “old days” of baseball. When you are behind in the count or you have two strikes, then conventional baseball wisdom and analytics need to merge cohesively. The situation needs to be considered. If you’re down one, late in the game with nobody on base and you’re a slow, power hitter...when you get two strikes, you’re still probably swinging for the fences on a close pitch or driving the ball. If you’re a speed-not-power guy, you’re making contact to try to just get on base and pass the buck for the next player up.


Defense and Pitching Concepts “My manager” needs to know:


Bullpen Management - The game is not always on the line in the 9th inning. Andrew Miller of the Cleveland Indians was famously used in the 2016 World Series consistently as a “fireman” - hearkening back to the 70s term of reliever usage. He was brought into the game when the game was on the line. The Cardinals have discussed using relievers like that. I’d love to see “My manager” use that type of role.


Therefore, saves would become a thing of the past in terms of paying for players in my bullpen. Holds (especially difficult holds), Win probability added (WPA), and leverage index would be terms “My manager” would be very comfortable with. He would also be able to comprehend Win Expectancy and use his bullpen accordingly. Using a replacement level reliever in the 6th inning with the game on the line because you want to save your best relievers for the 8th and 9th innings of games that are well in hand is not the best way to maximize wins. Know the stats behind it.


This would take a shift in reliever mindset; I get that. I believe this would hearken back to more of the 1970s "fireman" role. I believe Mike Maddux is equipped to do this.


Another thing is that starters, as I’ve noted before, typically are not nearly as good the third (or any subsequent) time through the order. As discussed in the previous link, using eight of your staff’s pitchers to throw to 18 batters a game 40 times a year (we’re talking ~160-200 innings still) is a huge plus for “My manager.” That’s how he would use them - and my guess is it would be quite effective.


Agents would hate it. I get that.


Shifts - “My manager” would have the confidence to tell pitchers that we’re going to shift and we’re going to do it using collections of data. “My manager” would have the wherewithal to talk down pitchers who give up a hit in a shift situation that goes to the uncovered portion of the field. From the beginning of 2016 to present, without a shift on, Cardinals pitchers have allowed a .297/.298/.377/.675 line. On traditional shifts, it’s a .291/.291/.367/.658 line. The OPS drops 17 points. The Cardinals have not fared well on “non-traditional” shifts, however. I’d keep the shifts traditional to start, but just have more of them. Even more subtle shifts, perhaps.


These are many of the things that are obvious to me and that have stuck out like a sore thumb for me for quite a while now that I’d love to have confidence in my manager being completely on board with and comfortable with.

What is really interesting to me right now is that Mike Shildt may actually be this person. I really have no clue if he is. MANY of the changes that he has made in the clubhouse seem to be reported on as what I am looking for. MANY of the in-game decisions seem to be fairly consistent with my thoughts above, but not always, and that’s fine. No one bats 1.000. Especially managers. . If the organization truly thinks that he can be this manager, then I truly hope that “interim” is soon lost from his title.