Games: New York Yankees, May 4: Judging the Royals with Lee Judge

Games » New York Yankees

May4

Derek Jeter is really good

The next time Derek Jeter comes to the plate, look at Alex Gordon. The Royals left fielder will be swung around toward left center and you’ll see a big gap between Alex and the left field line. The whole Royals outfield will be playing Derek Jeter to hit the ball the opposite way. The plan is to jam Jeter with fastballs on his hands. Jeter hits that pitch to right field and the Royals know this. If they get the pitch where they want it, Jeter won’t be able to extend his arms and he’ll hit weak grounders and fly balls to the right side — normally. Right now, the plan isn’t working and here’s why:

Derek Jeter is really good.

Not only is Derek Jeter really good, right now he’s a really good player who is also on fire. The Royals are making the pitches they want to make and Jeter’s still smoking the ball. Think about this: they’re throwing fastballs up and in, Jeter’s hitting the inside half of the ball and he’s hitting those balls hard.

I was told if Jeter pulled the ball that would probably be an off-speed pitch in a bad location. (Apparently this strategy is not a big secret, everybody pitches Jeter the same way and he knows it.) The double down the left field line against Danny Duffy Thursday night was a curveball. The home run against Bruce Chen last night was a curveball. You might wonder why they’re throwing hittable curves to Jeter if the game plan is to make him hit a fastball to the right side. Two reasons: 1.) They’re not supposed to be hittable, but if a Royals pitcher makes a mistake, Jeter isn’t missing it; 2.) They’ve thrown everything but the kitchen sink at him — and if they did he’d probably hit that for a double. Jeter’s getting hits on fastballs, change-ups and curves. Make a great pitch and he fights it off, make a mistake and he makes you pay.

Derek Jeter is really good.

The seventh inning

I asked Ned Yost if leaving Bruce Chen in for the 7th inning had to do with saving the bullpen. Royals starters have not been going deep in games and the pen is getting called in early almost every night. Ned said that was part of it — they’re always looking for quality innings out of a starter, but the main reason he left Bruce in the game was that Bruce was still pitching really well.

Chen gave up a home run in the first inning to Mark Teixeira — a cutter he wanted in on Teixeira’s hands that didn’t get in far enough. After that, Bruce settled down and matched CC Sabathia for the next six innings. Chen got through the 6th on six pitches, retiring Curtis Granderson, Teixeira and Alex Rodriguez with ease.

In the 7th Chen gave up a single to Cano, got two quick outs and then quickly gave up hits to Eduardo Nunez, Chris Stewart and Jeter. Ned said he thought Bruce was still rolling, wasn’t tired, but then gave up three runs in the space of eight pitches to a lineup that can do a lot of damage in a hurry.

Stealing third base

Jeff Francoeur’s attempted steal of third base Thursday night generated a lot of comments on the site, so Friday afternoon I asked some questions. I wanted to know what the Royals look for and when they think a steal of third is a good bet. I’m only dealing with a couple of the factors here, but the factors I’m dealing with are pretty interesting:

The pitcher is largely responsible for preventing a runner from stealing second base. The pitcher does this by throwing over to first, shortening the lead, varying the time he holds the ball in the set position and getting the ball to home plate in 1.4 seconds or less. As I’ve explained before, base stealers go from first to second in 3.4 seconds, catchers get the ball down to second base in 2.0 seconds, so the pitcher is the variable. (All these times are major league averages.)

Once a runner reaches second base, the middle infielders take most of the responsibility for preventing a steal of third. With a runner in scoring position, the pitcher needs to shift his focus almost completely to the plate — a hit might mean a run. As a result, pitchers tend to slow down their delivery — they’re not giving as much thought to the runner.

The middle infielders control the runner in two ways: shortening his lead and making sure his feet are not moving (at least in the direction of third base). If they think the runner’s lead is too big, they’ll run to the bag and force the runner back. Holding an open glove out to the side is the signal for the pitcher to throw the ball to second base.

If the middle infielders feel they’ve controlled the runner (short lead, feet stopped), they’ll break back to their position. That signals the pitcher that it’s safe to deliver the ball to home plate. When the infielders break, he throws home. But that delivery home often slows down. I’ve been timing pitchers lately and some deliver the ball to home plate in a glacial 1.8 seconds with a runner on second base.

Here are some more major league averages: a catcher takes 1.7 seconds to deliver a ball to third base, a runner with a slightly larger lead at second base takes 3.3 to get to third, so the pitcher needs to be 1.6 or over for an attempted steal of third base to makes sense. The best combination for a steal of third is one out, a right-handed hitter at the plate and a lefty with a big leg kick on the mound.

I don’t know how deep the Royals get into run expectancy, Francoeur’s base running history or any of the other numbers that have been discussed since Thursday night But if the score, number of outs, man at the plate and man on deck say attempt a steal of third — the Royals will go if the pitcher is delivering the ball to home plate in 1.6 seconds or more.

It’s pretty much based on what’s happening right now, tonight. So, if the pitcher was getting the ball to home plate in 1.5 seconds or less, it wasn’t a smart base running move on Francoeur’s part.

Get out your stopwatches.

Comments

Brendan Woodbury
1 year ago

Lee -

Thanks for following up on the steal of third. It’s nice to have a channel into the team’s thoughts on these things. I’d like to better understand their approach.

Also, given that Francoeur doesn’t have a stop watch out there, how is he supposed to tell the difference between 1.5 and 1.7 seconds? Relying on a runner whose heart is racing in a pressure situation to estimate tenths of a second seems strange. Are the Royals timing in the dugout and relaying the information out?
Matthew LaMar
1 year ago

Two things:

I know hindsight is 20/20, but I would think leaving a suddenly struggling Chen to face Derek Jeter with a man on base is a bad idea, especially when you have a bullpen full of good arms that get you the platoon split. I know Yost had his reasoning, but his bullpen use (or lack of use) is mystifying at times.

Also, the Royals are blindly running into outs on the basepaths, with is infuriating. They have been caught stealing 10 times, tied for most in the AL and third worst in the MLB and have been caught stealing third 4 times, which places them dead last in caught stealing third out of 30 teams.
Donald Wilson
1 year ago

I see that Yost is leaning on keeping Mendosa in the bullpin because he wants 2 long relievers. That might be a good idea (given their starters can only go 5 innings) if it was a descent pitcher. But keeping Mendosa (because he is out of options) over Hottovy, Coleman, and Herrera is a terrible decision. Truth is - they are keeping Mendosa because nobody else wants him.
Lee Judge
1 year ago

Brendan: They have the pitchers timed out before the game ever starts. They must do it off video or get them from the advance scout. (Someone who watches the upcoming opponent to give them an idea of what they’re doing right now as opposed to historically.)

Since I’ve started timing pitchers they seem to be pretty close on each delivery, but if a guy is 1.4 one time and 1.7 another, they probably go over that, too. Of course, it’s always a bit of a gamble: we can tell you what he normally does, but will he do something different THIS time?

Doug Sisson has a white card he keeps in his back pocket with all the times and keys written on it. When a new pitcher comes in you might see him pull it out and remind himself of what this guy does. Every time a runner reached first, Doug leans and reminds the runner of those times also.

I thought it was interesting to hear how they think about it as well and it does help you understand their decision-making process.
Lee Judge
1 year ago

Donald: Mendoza was the Pacific Coast League pitcher of the year, a PCL All-Star and a PCL playoff MVP last season.

His ERA was 2.18, got a September callup, went 2-0 with a 1.23 ERA and continued to pitch well in spring training.

Of course, he hasn’t pitched nearly that well this season—although he did have one pretty decent start as I recall—but it wouldn’t surprise me if someone wanted to take a chance on him.
Lee Judge
1 year ago

People ask if the “eyeball” guys and the stat guys can get along—sure, if we’re both polite, present our case in a reasonable way and aren’t dogmatic about our opinions.

One of the site’s readers, Curtis Ruder, likes to view the game froma statistical standpoint, but doesn’t believe he has a monopoly on truth. (Neither do I, by the way—I can tell you what I’ve experienced and observed and what ballplayers think, but that doesn’t make us right. It’s just one point of view.)

Anyway, Curtis like to run numbers for fun—apparently he hasn’t heard of “beer” yet—and he sometimes sends me the results. He ran some numbers on Jeff Francoeur’s steal of third and sent them to me. His email included charts that broke up when I copied and pasted them. I’m too lazy to try to put them back together, but I left the raw numbers just in case you can figure them out, but if not, just read what he has to say about the numbers and you’ll get the gist.

“Here is a useful numerical example:

Base Runners

1993-2010

1969-1992

1950-1968 1B 2B 3B

0 outs 1 outs 2 outs

0 outs 1 outs 2 outs

0 outs 1 outs 2 outs _ _ __

0.544 0.291 0.112

0.477 0.252 0.094

0.476 0.256 0.098 1B _ _

0.941 0.562 0.245

0.853 0.504 0.216

0.837 0.507 0.216 _ 2B _

1.170 0.721 0.348

1.102 0.678 0.325

1.094 0.680 0.330 1B 2B __

1.556 0.963 0.471

1.476 0.902 0.435

1.472 0.927 0.441 _ _ 3B

1.433 0.989 0.385

1.340 0.943 0.373

1.342 0.926 0.378 1B __ 3B

1.853 1.211 0.530

1.715 1.149 0.484

1.696 1.151 0.504 __ 2B 3B

2.050 1.447 0.626

1.967 1.380 0.594

1.977 1.385 0.620 1B 2B 3B

2.390 1.631 0.814

2.343 1.545 0.752

2.315 1.540 0.747

That is the run expectancy matrices for the last 60 years.

I think I have run through these with you before. But the key numbers from this position are runner on second, one out (0.721), runner on third, one out (0.989) and none on, two out (0.112).

Let’s run through the numbers first assuming a normal situation, with an average runner, average hitter, average pitcher, and average catcher. Another quick assumption is that the more favorable outcomes – Martin airmails the throw into left field and the run scores directly, for example – and the unfavorable outcomes – say Quintero lines out to an infielder and the runner is thrown out – roughly offset.

If Francouer steals and is successful, we move from 0.721 runs in the inning to 0.989 runs in the inning, which is a gain of 0.268 runs. If Francouer steals and is unsuccessful, we move from 0.721 runs in the inning to .112 runs in the inning, a loss of .609.

So if we assume the probability that Francouer is successful is X, then the probability he is unsuccessful is 1-X. What we want to know is where is the break even point, and that leads us to the following equation:

(0.268)X - (.609)(1-X) = 0. And the solution to that equation is roughly 69%. If Jeff can make the steal successfully 69% of the time or more, than it is a good gamble. If he cannot, then it is a poor gamble. (Quick aside – the number of runs in a game is down significantly in 2011 and 2012 and closer to the numbers from the earlier eras, so I re-ran the numbers using the middle matrix from 1969 to 1992, and it dramatically changed the results – the success rate would need to be at least 68.8%, not 69.4%)

So that gives us our baseline – the next question is, does the particular situation of the game call for being more aggressive than usual or less aggressive than usual. I think the answer is very stark – this is a situation that calls for aggressiveness. The marginal value of one run in the bottom of the eighth inning which extends a one run lead to two runs is huge. Furthermore, we are at the bottom of the order, and the probability of getting a base hit that can drive in the runner from second is lower from Quintero, Getz, and Escobar than it would be if Gordon, Butler, and Hosmer were due up.

Base Runners

1993-2010

1969-1992

1950-1968 1B 2B 3B

0 outs 1 outs 2 outs

0 outs 1 outs 2 outs

0 outs 1 outs 2 outs _ _ __

0.293 0.172 0.075

0.267 0.153 0.064

0.263 0.154 0.066 1B _ _

0.441 0.284 0.135

0.426 0.269 0.125

0.410 0.264 0.122 _ 2B _

0.637 0.418 0.230

0.623 0.411 0.224

0.615 0.410 0.227 1B 2B __

0.643 0.429 0.237

0.632 0.421 0.230

0.623 0.425 0.232 _ _ 3B

0.853 0.674 0.270

0.840 0.664 0.274

0.818 0.650 0.278 1B __ 3B

0.868 0.652 0.288

0.855 0.647 0.280

0.849 0.648 0.287 __ 2B 3B

0.866 0.698 0.280

0.855 0.678 0.275

0.839 0.664 0.285 1B 2B 3B

0.877 0.679 0.334

0.874 0.668 0.320

0.849 0.652 0.316

This shows the probability matrix for scoring at least one run. Basically, this matrix chops off the tail of the statistical chart that represents big innings. I think this is a situation where we are not so concerned with a huge inning, but much more concerned with adding an insurance run, and so if we re-run the equation, we get a success rate of 57% being the break-even point.

Usually, the break even point for a situation is roughly 75-77% using the above calculations. Which, in my mind, means the true break even point in most situations is actually in the 70% range, as I do think these cold calculations do not quite capture the benefit of putting pressure on the defense. But regardless, if the usual number is 75%, and we have a situation where the actual break even is 57%, and then we have the particulars of the situation calling for even more aggressiveness, then this is one of the most productive running situations we will ever encounter.

Next time, he just needs to not slip on his third step. (He was out by so far I knew something had to have happened, and so I went back to the Yankees broadcast and sure enough, you can see him lose his stride just a bit. If it had been the first or second step, he probably could have called off the steal and just gone back to second. But he had too much momentum and was too far gone.)

It is not a bad decision every single time a runner gets thrown out. Sometimes, you have to tip your cap. If I could log-in from here and post this as a response. Feel free to use as much or as little as you would like.”

I didn’t post this as an attack on other people into metrics. I don’t know if Curtis is right since he lost me right after “Here is a useful numerical example.”

I posted it because I think it demonstrates that there is more than one way to look at this stuff—even from a statistical point of view. Change the variables, change the outcome.

And I’ll guarantee you nobody on the field went through the situation in the same manner Curtis did. They were thinking, “If the pitcher’s under 1.6, don’t go.”

From their point of view, if the pitcher was 1.5 and Frenchy still went, then it was bad gamble.

It’s good for all of us—especially me—to remember nobody handed us the truth on a stone tablet. There’s almost always another point of view.

And thanks to Curtis Ruder for providing one.
Jim Wilson
1 year ago

Frenchy has a SB% of 60 and the run expectancy table Curtis is using is too low, so I’m not sure that this presented quite as good of an opportunity as Curtis makes it out to be. But the effort is awesome and very interestesting.
Lee Judge
1 year ago

Jim: You may be right, but it is very interesting and another way to look at the same situation.
Mark Harkins
1 year ago

Lee, that’s one HUGE post. Joe Pozanski would be proud… :)
Lee Judge
1 year ago

Mark: Always happy to take credit for work I didn’t do.
Luke Healy
1 year ago

I just want to express again my extreme gratitude for a site like this. Lee, the work you do is so great, especially for a fan base like the royals have. One commenter said it was nice to have a channel into the team’s thoughts—it’s more than nice, it’s really truly remarkable. It really makes me appreciate the team a lot more and see so much more than a normal fan would, which hopefully makes us better fans and more supportive. So thank you again, and thank you to your bosses at the star or whoever makes this possible.
Brendan Woodbury
1 year ago

Lee -

Curtis is good to dig up those numbers, but by using the 1969-92 data, he’s going with percentages that are too low. The Royals were a 4.5 runs/game team last year, and since they’re hitting about the same this year, they’ll probably end up around there again. The 1969-92 era is closer to 4.0 runs/game.

Using a lower run environment makes the chances of scoring lower (lower runs resulted from runners being less likely to score) which makes stealing a base a little better proposition (which is one reason teams stole so much more in the 70s and 80s).

Curtis’ source (Tom Tango) has a chart that has run expectancy for all run environments:

https://spreadsheets.google.com/pub?key=pkimQBCeCjbjIgxJ9vKvJSQ

If we use this chart to look up RE1 (the chances of scoring exactly one run), we get numbers a bit higher than Curtis had, which makes the breakeven success rate for stealing third a bit higher than he calculated.
Chuck Smith
1 year ago

Jim W.

You can’t just use Frenchy’s stats for his entire career for SB% (I mean you can, but any stat head would know that is wrong). If you look at just his SB% since becoming a Royal, and thus just the time he has been following the guidelines that the Royals implement for base stealing.

Looking at a last year (and yes, I know you can’t take one year in a vaccum, but you always start a time period analysis with the start of the change) his SB% was 68.8%. This is a whole 0.8% higher than Gordon. It’s also 12.7% higher than his career averages before then. Add in he is a whole 33 days older than Gordon I don’t think we’re looking at Frenchy having “lost a step” as someone posted. And finally, it’s not an overal stat matrix that determines when they run, it’s a situational matrix.
Chuck Smith
1 year ago

:) not sure how i made that Jim W. become larger and bolder. Wasn’t trying to yell at anyone. Kinda funny.
Jim Wilson
1 year ago

Chuck, you got my attention! Using the Royals’ current guidelines (this year), Frenchy’s SB% is zero. And I do believe that career numbers are relevant here. In a more conservative environment and at a youger age (but with plenty of experience), Frenchy SB% was still low. Coincidentally I’m sure, Frenchy’s SB% with POs included during his time with the Royals is 57.9%. But I’ll give him the benefit of the doubt and continue to use 60%. :)

I don’t understand the relevance of the information you cite about Gordon to whether Frenchy should be attempting to steal third.

The situational matrix needs to be evaluated regularly to determine whether it’s working. It hasn’t been.
Chuck Smith
1 year ago

Jim, I went and pulled stats for this year, and since I like to believe I’m open minded, I tried to keep an unbiased examination while trying to prove my point.

These are some of the things that stood out. * The Royals are in the top 3rd in steal attempts. * The Royals are in the bottom 3rd for SB success. If you break it down to AL stats only: * KC is 4th in SB Attempts * KC is 10th in SB Success rate

Since last year, KC is down 9.5% over last year in Successful SB%. At the same time they are attempting almost 13% more steals per game.

Add in a spring training post from Lee about more teams refocusing on preventing the SB and I’d have to aggree with you that the matrix might need to be reviseted along with their whole SB thought pattern.

People are aware that KC will run on them now, that alone may be enough to get the positive effects of distracting the pitcher that is wanted.
Jim Fetterolf
1 year ago

“The situational matrix needs to be evaluated regularly to determine whether it’s working. It hasn’t been.”

Yesterday’s blog looked at that and it has been evaluated and worked on and the team is 9 of the last 11 of steals.

“I asked Ned Yost if leaving Bruce Chen in for the 7th inning had to do with saving the bullpen.”

Ned Yost gets hammered for leaving Master Chen in too long, just days after getting questioned about pulling Sanchez and Duffy too early, even though both are historically about five inning pitchers and Chen is the closest thing we have to a horse at the moment. Some days you get the bear, some days the bear gets you.
Jim Wilson
1 year ago

I think it’a matter of having a lot of different opinions. You can never make everyone happy. I would have pulled Chen. I agreed with pulling Sanchez and Duffy for the reasons you cite. Sanchez needs to become more efficient before he can go beyond five or six innings.
Jim Fetterolf
1 year ago

“I would have pulled Chen.”

Why?

“Sanchez needs to become more efficient before he can go beyond five or six innings.”

Sanchez was at 83 pitches, as I recall, 16 or 17 pitches/inning. Duffy was in the same range. For both of them they were surprisingly efficient. Perhaps Yost saw something beyond the numbers?
Lee Judge
1 year ago

Luke H.: Thanks, a lot of people deserve credit for this site. The Star gave me the chance and the support to make it happen, the players along the way that taught me to see the game in a different way, but most all the Kansas City Royals.

They’ve been extraordinary about granting access and information. I’ve got to be careful to not reveal anything that would help another team, but other than that, everyone seems willing to share what they know.

It makes watching baseball a richer experience for all of us.
Scott Boden
1 year ago

I’ve made few posts but one of them was frustration at the metric guys for constantly attacking Lee.

I was soooo happy to see the comments this time. This is what I like to see. Some great metric dialogue with no attacks. Excellent! Very good arguements and very interesting from all sides. Thanks guys.
Antonio Cutolo-Ring
1 year ago

I agree with Scott. The posts this time are interesting without being rude. I really appreciate Lee’s giving us so many specifics about what goes into playing at this level, and what the players and coaches are thinking. It’s why I keep coming back to this site (and you do a good job of writing and explaining it, Lee) Consequently, I’ve been considering a lot more aspects of the game.

I also am interested in hearing those with statistical points of view, when it’s in the form of respectful comments. But it’s not the main reason I read this blog.
Tracy Anderson May
1 year ago

Ditto Scott and Antonio!
Curtis Ruder
1 year ago

The reason I used the 1969-1992 numbers is that they are closest to the current scoring environment, but the bottom line is that using any one of the three sets of numbers changed the break even expectation by less than 2%. The main point that I was trying to convey is that the situation of runner on second with one out late in a close game requires a much lower success percentage to be a good gamble than almost any other time. Compare the 57% calculated here with the 75%-80% you get in almost every other situation. If the current run scoring environment is different, it would affect both rates in the same direction, which is why I didn’t worry about getting the best possible matrix.
Jim Fetterolf
1 year ago

Nice work, Curtis. Still think you should do your own blogs. You have a nice combination of journo cred and statistical comfort that may be unique in our market.
Jim Wilson
1 year ago

Jim F., 16 or 17 pitches an inning is suprisingly efficient when all he (Sanchez)had given up was a double and two walks in five innings? He was struggling to throw strikes but it had not bitten him yet. Sanchez is averaging 21.2 pitches an inning, so 16 or 17 is an improvement, but “surprisingly effecient”? I would have pulled Chen because I felt like Chen and the Royals needed a good start from him after his prior disaster. The bullpen has been pitching well and we’ve got enough bodies on the roster and in AAA, so I would have pulled him after he made it through the line-up the second time. But that’s just me and a gut feel. Not really criticizing Yost. And I agreed with his decisions on Sanchez and Duffy.

Judging the Royals