Mike Matheny’s Rookie Mistake
First, let’s start off with caveats. The Cardinals offense didn’t score a single run last night, so there was probably nothing Mike Matheny could have done to change the final result. The Giants rally from the point of contention consisted of a broken bat base hit that no one had ever really seen before, an infield single, and a pair of ground balls to shortstop that weren’t handled properly. From a retrospective point of view, it’s impossible to look back at the third inning and think that the decisions made by the Cardinal manager were the cause of the team’s downfall.
At the same time, however, the way Matheny handled the third inning is a mistake he needs to learn from.
Through the first two innings of the game, Kyle Lohse had thrown 38 pitches, and he wasn’t exactly sharp. The Giants had already collected four hits — including a line drive by pitcher Matt Cain — and put two runs on the board, and now the middle of the order was due up again. With Scutaro, Sandoval, and Posey set to hit, the possibility of Lohse getting chased early was very real, and in a Game Seven situation, you have to start warming two relievers just in case. This isn’t a situation where you care about getting a guy loose and then not using him – odds are pretty good you’re emptying the bullpen today no matter what, so anyone who gets loose now is going to pitch, and probably pretty soon.
Instead, Matheny started the inning with no one warming behind Lohse. Two pitches into the inning, Marco Scutaro was on third and Pablo Sandoval was on second, and now it was scramble time. Down 2-0 with two men in scoring position and nobody out, the Cardinals Win Expectancy was already down to 12%, and if you account for the fact that they were on the road and facing Matt Cain, it was probably a little bit lower than that. If you can somehow get out of this situation only allowing one run, you’re still in the game, but there was also the distinct possibility that this rally could blow the game open and decide the game right then and there.
You already have runners at second and third with nobody out, and you’re pretty sure that Buster Posey is going to draw a walk with first base open, even though it’s not officially an intentional walk. That will bring Hunter Pence to the plate with the bases loaded, who will then be followed by Brandon Belt (LHB), Gregor Blanco (LHB), and Brandon Crawford (LHB), then the pitcher’s spot. It’s a pickle, but it’s a pickle you can get out, especially if you can strike Pence out, because then a double play gets you out of the inning without a run scoring. A double play from Pence is a worthy trade-off too, but that one costs you a run, so your strong preference should be for a strikeout of Pence first.
This sets up two pretty decent options:
1. You can take advantage of the R-L-L-L bats coming up, by first bringing in a high K right-hander to face Pence, then bringing in Marc Rzepczynski to go after the three lefties. The best option for a one batter righty specialist when you want a strikeout is probably Edward Mujica, as his career K% against RHBs is 21% and his BB% is just 4%. Mujica’s a strke-thrower who misses bats, which is exactly what you need in a bases-loaded situation. Rzepczynski has struggled lately, but has historically been extremely tough on left-handers, and has a career 63% GB% against LHBs. Even in a down year, he ran a 2.18 K/BB and 4.18 GB/FB against LHBs this season. Using Mujica against Pence and Rzepczynski against Belt/Blanco/Crawford maximizes your chance of the strikeout/double play combination that keeps the score 2-0 and gets you out of the jam.
2. Bring in Trevor Rosenthal. The Cardinals fireballing youngster has been dominant out of the bullpen, striking out 11 of the 22 batters he’d faced during the postseason before last night, and his velocity/curve combination makes him highly effective against hitters from both sides of the plate — he was actually slightly better against lefties in the minors, and the same has been true in a very small sample during his big league career. During the regular season, Rosenthal faced 89 batters, and 54 of them (61%) either struck out or hit the ball on the ground. Rosenthal is the strikeout and groundball package in one pitcher, capable of escaping that jam all by himself, and allowing you to save the match-up relievers for later in the game if necessary.
Either option had upside. This is where the Cardinals bullpen depth — combined with having Adam Wainwright available out of the bullpen — was supposed to pay off.
Instead, Matheny got Joe Kelly up. Kelly isn’t a bad pitcher, but he’s essentially a right-handed groundball specialist. He’s mostly a sinker/slider pitcher who throws from a low 3/4 arm slot, which makes him vulnerable against left-handed hitters. In the minors, he posted huge platoon splits, with LHBs posting a .780 OPS against him, compared to .616 for right-handers. While his Major League data is still a small sample, the trend has continued – .635 OPS against right-handers, .908 against left-handers. He only struck out 22 of the 200 LHBs he faced in the Majors this year, and lefties hit .313/.389/.528 against him. There’s no way Kelly should be asked to run through three left-handed batters even if he gets Pence out, so Rzepczynski needs to be ready to face Belt if you’re going to use Kelly as the righty-one-out-guy.
Instead, Matheny used Kelly there because, in his own words:
“Right there we needed a ground ball, exactly what Joe Kelly got us. A pitcher would have been up fifth, which we would have risked losing Trevor Rosenthal for a partial inning instead of being able to stretch him out a bit.”
He didn’t want to use Rosenthal to get three critical outs in the third inning because there was a chance that the Cardinals would get two hits in their next at-bat, forcing a pinch-hitting situation that would chase Rosenthal from the game. Instead, he was willing to use an inferior reliever in the most important part of the game so that he could save Rosenthal for two innings in a less critical situation later. And, of course, there was a pretty good chance that Rosenthal wouldn’t hit in the next inning anyway, as getting two guys on in four plate appearances against Matt Cain isn’t something you should count on.
Kelly ended up giving up the broken bat “double” to Pence that broke the game open, then went on to face all three left-handed hitters, the pitcher, a switch-hitter batting from the left side, and finally one more right-hander before being lifted so that Mujica could retire Pablo Sandoval to end the inning.
All told, Kelly faced seven batters. Four of them hit left-handed, and all four lefties put the ball in play. One of the three right-handed batters he faced was the pitcher.
You could maybe make a case for using Kelly against Pence before bringing Rzepczynski if you wanted to save Mujica for a full inning, since he doesn’t have huge platoon splits. While Kelly isn’t the pitcher you really want on the mound in the highest leverage situation of the game, he wasn’t a terrible match-up against Pence, and he basically won the match-up, only to get screwed over because of the physics of a thrice-hit baseball. But, the plan was bad from the beginning. Even if Kelly had gotten Pence out, Matheny was still asking him to get Brandon Belt in another critical situation. Kelly is literally the worst option in the St. Louis bullpen versus left-handed hitters. If you lose this game, your season is over. Having him face Belt (and then Blanco, and then Crawford, and then Pagan) just should have never been an option, especially with Rzepczynski and Rosenthal sitting in the bullpen watching the entire thing.
The way he handled Jason Motte this year showed that Matheny is willing to be flexible in his bullpen usage when it benefits the team. Last night, however, he let the inning dictate which reliever he used in an elimination game. Even for the staunchest believer in the value of planned roles for each pitcher, that’s just overly fanatical to the idea. Matheny had good options in the third inning, and given how Rosenthal has been pitching in the playoffs, it’s not that hard to see him getting out of that situation with the score still 2-0. Instead, the inning ended with the Giants up 7-0, and the rest of the game was basically academic from that point on.
Rosenthal, Mujica, Boggs, and Motte still took the mound in Game Seven. They just pitched when it was too late to actually matter.
Dave is the Managing Editor of FanGraphs.
Go back to the NLDS Reds/Giants and you will see Crusty Baker made the same rookie mistake in leaving in a not-so-sharp-that-day Latos to give up 2 runs, then load the bases then give up a GS. All with not even 1 pitcher warming up in a do-or-die game. In this case it likely make a difference because the Reds scored 4 in an attempt to come back, in addition to having a rested bullpen. Dusty is a great guy and I wish him well, but please please let someone else manage the games, especially post-season. His track record in post season speaks for itself. Matheny making a mistake his first time around is more than acceptable.
Baker has had issues with bullpen management dating back to his days in San Francisco. Ask any Giants fan if they’ve forgotten the Game 6 debacle in Anaheim in 2002. I like Dusty, but his bullpen management is, and has always been subpar.
I can’t get over Game 7 of the 2002 WS, when Dusty let get Livan get battered around when it was clear he had nothing that day
Yeah, Big Fish and Ted. I was (and still am) a Giants fan in 2002 and can never forgive Baker for losing that series.
I live in Cincinnati now and would love to root for Cincinnati as my 3rd or 4th team, but I can’t do that because I hate Baker so much.
I’ve been saying all year that Matheny is Dusty-lite. His players seem to like him and play hard, but he is just helpless on game situations and line-ups. Not only that but like Baker he just refuses to learn after he gets burned.
He did learn not to have Rzep face Tyler Moore rather than having Boggs face Chad Tracy after Game 1 of the NLDS. He is capable of learning.
I would say that the holder of a very unique and extremely desirable job that draws a seven figure salary to make decisions that aren’t very difficult should not be making mistakes like this let alone have people find it “more than acceptable”. By far and away the easiest part of a baseball manager’s job is in game decision making. It astounds me that any in game mistakes are ever made that aren’t razor thin at the margins.
Dusty is loved dearly by the players. Dusty is clueless at game management, because ne wants and needs to be loved by the players … especially the veterans. The Reds have a team over the last 3 years that could have made a lot of postseason noise. Instead we are writing posts like this.
A question: Do win probabilities change in playoff games? As I understand, WP’s are based on the results of many thousands of regular-season games. But a game late in a playoff series isn’t like a regular-season game. Managers are going to try different strategies because they have more resources and because more is at stake. Possibly those factors cancel each other out.
That is an interesting point and probably impossible to determine since the postseason is such a small sample. Going back more than 30 years and to some extent even more than 15 bullpen usage was very different so comparing the situations would be tough.
Win probabilities are based on average players coming up to pitch and bat, and thus are only rough approximations anyway. I doubt that the difference made by the additional strategies available in playoffs would effect them significantly.
Excellent, insightful article. Thanks, Mr. Cameron.
Would you have pinch hit for Lohse in the top of the second? I would have.
Can I safely assume that you’re joking?
No, you cannot. I’m not as convinced that it’s what he should have done as I am that he should have pinch hit for Lynn in the bottom of the 2nd of game 5, but between the bullpen depth, the leverage of having multiple baserunners on (in both cases), and the difficulty of scoring against the Giants, I’d say you take your chances when they come.
I probably would have let him him hit with a quick hook if he allowed any more runners. Wainright would have been the replacement with a possible burn of a standard reliever to get out of a jam.
What the Giants did right is to capitalize on Lohse’s tendency to throw cheese for a first ball strike. Second time through the order and the Giants were waiting in the bushes. Matheny should have been anticipating that as it was painfully obvious to most people watching.
Both managers were horrendous last night.
Bochy managed a 9 run lead perfectly.
This is a tremendous piece, as usual, Dave. My only possible quibble is with the title, which implies that Matheny’s mishandling of the 3rd inning is a byproduct of his inexperience. Yet I’ve seen nothing to indicate that Matheny has learned from his mistakes, or will in the future. If anything, the Cardinals’ success this year seems to be encouraging his very worst instincts, so we can probably expect more of the same mismanagement from him going forward.
I agree, but I hope that we’re both wrong.
This was a great piece but it’s worthy of mention that Rzepczinski, in all likelihood, doesn’t inspire a lot of game 7 confidence from Mike Matheny. He’s been awful a good chunk of the year and probably barely made the postseason roster. In fact, if it weren’t for the other horrible left handed options out of the pen, he wouldn’t have. Rosenthal was clearly the best choice in that situation.
Even during his relatively poor 2012 season, he had a 63% GB% against LHBs. If you want a double play ball, Scrabble vs a lefty is still a good match-up, even if you don’t trust him right now.
Still writing about the Redbirds anywhere, Mr. Gunn? You had the best stuff on the internets several years back, sir.
I’m a left-coaster now, so don’t see many Cardinal games. I’m very curious: to what consistent Matheny “mismanagement” do you refer?
Since Matheny didn’t have the benefit of knowing when the highest leverage situation would be during the game, suppose he had done exactly what you would have done. Suppose also, that a much higher leverage situation had then occurred later in the game and he was then forced you use Kelly against suboptimal matchups. Would you still lecture him for needing to learn from his mistakes?
The manager has to go with the information he has at the time. At the time, the team was down 2-0 with and the Giants had the bases loaded with no one out. It was at least as likely that this plate appearance would be the last important plate appearance of the game as it was that there would be another one of equal importance later on. Moreover, Matheny would’ve likely been inclined to use someone like Mujica, Boggs, or Motte later on — the last 2 of which weren’t considerations for the bottom of the 3rd inning.
And if this situation happens in the 12th and all Matheny has left is Joe Kelly, then Matheny would have to use him as Kelly would be the only option. Clearly that wasn’t the case here. He can hardly be faulted for using one of his best relievers in a game-changing situation and then not having him around for another one later on.
It’s clear, though, that — given Kelly’s profound splits — he absolutely shouldn’t have been the guy used in that situation unless he was just going to use him to pitch to Pence. Whether that’s a rookie mistake, inexperience, stubbornness or stupidity hardly matters now. Hopefully for the Cards, it’s just a rookie mistake that he’ll learn from but, given the way in which he’s managed the bullpen many times during the season, there’s reason for Cards’ fans to be concerned.
Just play it out – it’s easy to see how far-fetched this scenario actually is.
Put in Rosenthal – he pitches the 3rd inning. Say the team actually gets two baserunners against Cain in the fourth, and you have to pinch-hit for him, so you burn him on “just” getting three huge outs to keep the game close. For the fourth inning, you’re looking at the top of the order again, so you probably want to go with Mujica or Boggs to keep another rally from developing. Depending on how the fourth inning goes, you can either stay with current reliever against the bottom of the order or go to Rzepczynski to run through the lefties. Now it’s the 6th inning, and you still have one of Mujica/Boggs, plus Wainwright and Motte, and you only need to cover four innings. You probably want Wainwright to have enough time to get warmed up since he’s used to starting, so you hand the ball to Boggs/Mujica for the 6th while having Wainwright warming up to pitch the 7th. He can then go two innings and hand it over to Motte for the 9th, who can stay in for the 10th if need be.
If it goes to the 11th, you might need Joe Kelly, and you might end up with him facing some left-handers in a tight game. But making your team less likely to get an out in the 3rd inning to improve your bullpen depth for the 11th is just silly.
Well, that would be great if it worked out like that, Dave. But your hypothetical involves exactly zero additional trouble and involves everyone performing perfectly. Suppose they get out of the third and the fourth begins equally bad with Mujica or Boggs… Should a manager not plan for trouble?
I’m not arguing that Matheny was correct. I’m just pointing out that the strategy of blowing your load early is no way to go through life (and not necessarily the best way to go through a baseball game). These things are not necessarily as obvious as you indicate, because a manager has imperfect information during a game, and the earlier in the game, the greater his uncertainty.
The trouble you are in right now is always the most important trouble of the game. Hoarding bullets for the prospect of, maybe, another troublesome situation later in the game is pointless if you don’t survive the first onslaught.
nope, you definitely never want to blow your load early.
of course, if you do survive, you are that much better off. …suppose the weird broken bat triple hit had not been a hit….
First, Matheny could have put Ryan Jackson on the roster instead of relying solely on Pete Kozma. The “Kozmanaut” came back to earth for the NLCS and became the player he was previously. However, the cards continued to run him out there in games 6 and 7 even though he had become the replacement level player he was meant to be. At a minimum they could have played Schumaker at 2nd and Descalso at SS to get some actual offense. Matheny also put Holliday in despite the fact that Matt Carpenter is a far better option against Cain. Carpenter has owned Cain in his brief career.
Completely agree on the rest. Kelly was a particularly poor choice.
You could also take note of the fact that Lohse has been a bad postseason pitcher yet he was set up as our ace throughout the postseason. Its like Matheny didn’t even notice how ineffective Lohse was all last postseason.
stan, regarding your last paragraph, Lohse was terrific in 2012, and he pitched well in his first three post season starts this year (4 runs in 18+ innings).
Lohse was equally terrific in the 2011 regular season. Lohse had a nice game in the LDS but he was allowing a lot of baserunners in his other starts.
Replacement-level shortstops either hit .227/.320/.273 or they play defense like a monkey f*cking a football, not both. Kozma was and is a sub-replacement level player.
Heh. I thought I was being just a little harsh on Kozma already, but you may be right.
Frankly I’dbe surprised if we ever see Kozma in a major league uniform again-except that Matheny is his manager and Matheny gave replacement-level Descalso 374 at bats this year and ran him out there every day in the playoffs too.
I don’t get all great article comments – this was yet another “what an idiot” articles; while there were #’s thrown in here and there, there was no actual analysis
What’s the run expectancy difference in the 2 different situations? Large? Small? If you are going to project 2-0 in one scenario, how about projecting the expected outcome of of Matheny’s actual move (If you are critiquing the decision, shouldn’t it be base on the expected outcome of the decision?)
And what does that do to the OVERALL win probability. a 2-0 game vs the two different expected outcomes… did that change the win probabilty a lot? A little?
Here’s some of the nuggets the article gave us:
“If you get out of the inning, you are still in the game”
“It’s not hard to see him getting out of that inning 2-0, instead it was 7-0” (
so a projection based on “it’s not hard to see it” vs the hindsighth of what did happen – instead of looking at the expected outcome given the moves Matheny made)
It’s easy to write yet another article throwing stones at managers, but this is Fangraphs – how about a statistical and analytical approach to quantifying what the potential magnitude of the mistake was? And not using hindsight against a thumb in the air “it could have been 2-0 if he was smart”
Ok. I’ll bite. Defend the decision to use Kelly as good process.
Also, you can’t really use a “hindsight” argument when Cameron and others disagreed with the move as it was happening. (During the fangraphs chat last night, I remember someone lamenting the use of the Cards nth best reliever – a comment chosen and ‘published by Dave, presumably because he thought it was insightful.)
Finally, the inning played out pretty much as the numbers would suggest. Granted the Pence double was lucky…but the lefties, who process suggests would succeed, succeeded against Kelly. It was a bad decision. The results proved it. And it wold have been a bad choice (and lucky result) even if Kelly had managed to shut things down. This article is not nearly the “bad things happened thus Matheny must have done something wrong” argument you’re proposing it is.
The hindsight I’m talking about it using the actual outcome against some hypothetical outcome. If you are going to critique the decision – use an expected outcome for both scenarios.
If you were critiquing a pinch hitting decision would you use the actual outcome of one AB against the expected outcome of another? Or would you logically use an expected outcome for both?
I’m not defending Matheny’s decision or saying it was a good process, I just think the magnitude of the mistake is not nearly as substantial and is being spun; like Dave tends to do. The same good process should apply to the analysis as well; instead a picture of 7-0 vs 2-0 is being painted.
If you think the likely outcome was 5 runs allowed, then I guess the analysis is spot on.
What Tom said. You have magnified the costliness of this decision greatly. People on fangraphs always bag on “mainstream media” for inappropriately using stats, yet when the same is done and said about some fangraphs articles they are seen as ignorant and know nothing about baseball.
What you suggest would have been impossible, as WP’s are not calculated for specific players. Dave correctly stated the WP at the time to help indicate that the game would get out of hand if there were any more scoring and then correctly used stats of various pitchers (including K%, GB%, platoon splits) to analyze whom Matheny should have used.
It is a perfect instance of what a FanGraphs article should be.
Spot on, Dave. Except, I would argue that Matheny’s inflexibility with his lineup costs the Cardinals games and in particular games in this series. Holliday has been in the three hole all year long despite having an average with RISP hovering around .225 all year. This is 80 points lower than last year. Many managers would have moved a guy down in the order that has been as unclutch at Holliday. Heck, Bochy even moved Pence down for one game and that worked out wonderfully. Further, Holliday was hurt this week, and there’s a hot replacement in Matt Carpenter able to take his Game 7 spot. Instead, Carpenter rides the pine all game and Adron Chambers and Shane Robinson pinch hit ahead of him. Also, Beltran has been the hottest post-season player on the planet this century. Why not move him down to the 3 hole? Molina is Mr. Clutch all year with a .304 average with RISP and 2 outs and .309 average late in close. He ended up going 4 for 4 in game seven. Why not move him behind Beltran? Except Matheny doesn’t do that. He’s had the same lineup since opening day no matter rain or shine and I have to think, it cost the Cards big this year.
Past clutch is not predictive of future clutch, so you should keep your best batters in the prime lineup spots regardless of how clutch they’ve been.
FWIW, his wRC+ with RISP was 108 this year, so he wasn’t really a disaster in those situations.
I love FanGraphs
“it cost the Cards big this year.”
they lose one of the Top 3 greatest players in their great history — baseball’s best player of the past decade — move on after one the sport’s long-time best managers (as well as one of the best-ever pitching coaches), avoid the modern-day-common hangover and defend a WS championship to the point of Game 7 for the NL pennant?? i’d say that’s a big cost any team would pay
Have you watched Pujols at all in the last three years? He was not much of a loss and Cardinals fans were well aware of it. I don’t think anyone in St. Louis would rather have had Pujols this post-season than Beltran, especially because they would have been extremely right handed and even worse defensively with Craig in right field.
that said, Matheny absolutely — no probaby numbers needed — made the wrong decision with Kelly. i was scratching my head to remember if Kelly had even thrown in the series
probability numbers
The playoffs and world series are full of improbable outcomes and head scratching decisions. That’s why we watch it with so much anticipation, hope against hope that the improbable decision will payoff and our team will win.
Me go the wrong way
So no easy double play
Manager gets blamed
I like what Hurtlockertwo just said.
After Cain’s line drive (which almost hit Lohse in the head), Matheny needed to get Wainwright up in the top of the 3rd and get Lohse out of there after the leadoff hit to Scutaro. The real blunder was letting him face the Panda for a second time. We can argue if Kelly was the right choice, but by then the animals are out of the barn.