Is there ever a must-win game in April? One of the standard arguments in the MVP discussion each season is that a particular player’s performance was better in September, “when it counts,” than it was at other times. The stathead counter to that is fairly simple: all games count exactly the same in the standings. That we know more about the arc of the season in September, or that a team has more time to make up a deficit in April, doesn’t weight the games any differently.
The Braves and Phillies began their three-game series tied for first in the NL East. The Braves have taken the first two games, holding the Phillies to just one run total. The Braves exploited a key Phillies weakness in the first game, and have beaten up both Phillies starting pitchers. If they can win today, not only will they complete a sweep, but they’ll go three games up on the Phillies. That reads almost like snark, but it’s not; today’s game will cause a two-game swing in the standings, standings that already favor the Braves. If it were September 8, this afternoon’s contest would be the center of the baseball world. Maybe it shouldn’t be, and I’m just missing something, but it’s a very important game in determining who will eventually win this division. A two-game swing in the standings is massive, and if any team should know that, it’s the Phillies, who won the East by a single game two years ago and by three-clinching on the next-to-last day-last season.
We miss the importance of these games early in the season because we’re just not trained to look for them. But peek back a year, to April 20, 2008, and consider how important this sweep-avoiding win by the Phillies was. Think about how the Mets‘ bullpen failing to protect a tie, and the Phillies’ bullpen throwing four shutout innings, was something of an introduction to the theme of the race. Think about how differently the last weeks of the season play out if the Mets have an extra two-game cushion in the standings.
Just because we don’t know if a two-game swing will be important doesn’t mean that we can’t behave as if it will. Sure, these teams are going to play another 15 games, and there’s plenty of time for story lines to develop. Math is math, though, and two games in the standings are enough to change an awful lot of baseball history. The Mets blew a 6-2 lead to the Brewers last April 13, eventually losing 9-7. The Mets finished one game in back of the Brewers for the NL Wild Card. You could argue that they lost that race on April 13.
The entire season matters. Each game counts exactly the same as every other, and working toward the elimination of the “when it counts” fallacy would be a very good next project for those of us in the advocacy wing of the performance analysis house.
It’s not just the Phillies/Braves series that highlights this phenomenon. The Royals may be the least of five AL Central teams, but it’s a division with so little separation that intradivisional games will be huge. Losing to the White Sox in a fairly ridiculous manner costs them ground that could be valuable. Blow a game with your fourth-best reliever to break a tie for first place in September, and they’ll have your head; it’s just as silly of a mistake in April.
- The Astros‘ decision to sign Ivan Rodriguez was a low-cost, moderate-upside play. Rodriguez is mostly done, and there’s not that much difference between him and Humberto Quintero-an underrated defensive catcher-at this point.
If you’re going to make it work, though, you can’t invite Rodriguez to be a disaster for the offense. Rodriguez was, at one time, an interesting option as a #2 hitter. He hit for a very high average that gave him a good OBP, and his ability to get hits made him attractive for batting behind a player who could run. The tradeoff was that Pudge, even in his heyday, hit a lot of balls on the ground from the right-hand batters’ box, so when he wasn’t hitting doubles, he was often hitting into double plays.
The 2008 version of Rodriguez comes with all the negatives of that package and none of the positives. He is, in fact, the anti-#2 batter: a low-OBP, low-power bat who hits a million balls on the ground (at an increasing rate-his three highest GB/FB ratios came in 2006, 2007, and 2008, peaking last year) and has lost most of his speed. Last year he grounded into a double play in more than one in five opportunities, after doing so at an 18 percent rate in ’07.
He brings absolutely nothing to the table as a second hitter. He makes Michael Bourn, who had a .288 OBP last season, look like Dustin Pedroia. Bourn will probably have a higher OBP than Rodriguez will, and he won’t wipe out the runner in front of him as often as Rodriguez does. Do you know how bad a player has to be to make me advocate for Michael Bourn as a #2 hitter?
I’ve taken you the long way around to this point: Cecil Cooper doesn’t know what he’s doing.
-
Then we have the Tigers. On January 24, the Tigers signed Brandon Lyon to a one-year deal. At the time, they made it quite clear that he wasn’t promised the closer’s job, just a chance at it, but by the time camp opened, he was considered the favorite for the job. A month later, Jim Leyland made it clear that he wanted to have one closer, though he wasn’t specific about which candidate he favored. Lyon was so bad in March, however, that by the middle of the month, Leyland backtracked on that, saying, “I don’t know that we’ll have one closer.” By the time camp broke, though, Leyland had come full circle, choosing Fernando Rodney, rather than Lyon, as his closer.
(Thanks, Rotowire.)
Take a look at that. The Tigers committed $4 million to Lyon in January, coming off of a season in which the pitcher had been a closer for most of the year, and where the Tigers had a vacancy in the role. They went into camp wanting a one-closer solution, with Lyon the frontrunner. Lyon, who even while collecting saves was no better than a middling reliever in the weaker league, was so bad that he pitched himself out of any share of the job.
Then, in the first situation of the season where the Tigers need a good right-handed reliever, they go to Brandon Lyon, solely because it was the eighth inning and not the ninth. The game was on the line, runners on base, a win in the balance. Jim Leyland had definitively chosen Fernando Rodney as his guy… and yet he let Lyon, who had pitched his way out of an important role with the team for six weeks, pitch in the first high-leverage situation of the season.
On top of the Kyle Farnsworth/Trey Hillman rant yesterday, you may be thinking I’m becoming a bit shrill on this subject. Maybe so, but it’s time for this nonsense to end. There shouldn’t be “eighth-inning” and “ninth-inning” relievers. Partitioning relievers by how many outs are left in the game was stupid when managers started doing it, and it’s even moreso now, as we find every bullpen in the game set up this way.
To all 30 managers, I issue this directive: Figure out who your best pitchers are, or more accurately, who your best pitchers are for various situations. For when you need a complete inning against the middle of the lineup; for when you need multiple innings; for when you need a ground ball; for when you need a strikeout; for when you need to get Jim Thome out. Then use them accordingly regardless of what time it is. Stop relying on the crutch of which inning you’re in to make these decisions for you. Your pitchers want roles? Their role is to get guys out.
These are not difficult concepts. Facing the middle of the lineup in the eighth is harder than facing the bottom of the order in the ninth, no matter how many ex-players who are invested in the myth of “closer” say otherwise. Stop using your better pitchers in lower-leverage spots. Getting four outs instead of three isn’t going to break anything that wasn’t going to break anyway, so stop losing games without getting your best pitcher into them.
Bullpen management is horribly broken in today’s game, and the first manager to fix it-Joe Maddon, I’m looking at you-is going to the Hall of Fame.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.
Subscribe now
Maybe that's exactly what Leyland and Hillman were doing.
There were at least 3 other pitchers who would have a better chance of getting Thome out, or at least certainly keeping him in the park, than Farnsworth (Mahay, Soria, Cruz).
There is NO argument that Farnsworth was more likely to get Thome out than these other 3 guys.
But that's really the point, isn't it-- it's what really makes September baseball more exciting.
It's true that rhetoric like "win when it counts" is misleading and mostly just wrong, because all games count.And yes, all games have the same value when it comes to overall standings.
But the reason that September games have more importance in most people's eyes is that the time frame for "making it up" grows ever shorter. A 2 game swing in April is much easier to erase as the season moves on than a 2 game swing in September. The math backs that up as well-- a 2 game deficit with 10 more to play impacts your playoff probabilities much more than a 2 game swing with 160 games left to play. Thus each game means more in September than it does now.
That's what most people mean, I think, when they talk about September baseball "counting for more".
Which kind of gives your argument a different flavor and somewhat negates what I wrote. So never mind.
It's why you'd see CC Sabathia pitch on three days rest in September but never in April: the games in September represent a big enough swing in playoff odds that they merit the increased risks of injury/fatigue that come with expanding the playing time of your best players.
I'm not saying that this justifies Hillman's specific choice of Farnsworth over the other options, but the baseball season is a marathon. A manager can't start tactically sprinting in April.
Basically, I'm saying that Joe's point about closer usage is valid, but not because of the time of year.
You can't have everybody loose and ready in the bullpen for whatever situation might present itself, can you? Maybe my question should be this: Realistically, how long should it take a major league pitcher to warm up and be ready to come into the game? 5 minutes?
Any longer than that and I don't think you can predict what situation you might face to ensure that you have the right guy warmed up and ready in the pen, can you?
Another example is Mike Scioscia. Monday night, with a 3-0 lead, he uses Arredondo for 1 out in the 7th, Shields for 1 inning in the 8th, and Fuentes for 1 inning in the 9th. Why not just let Arredondo or Shields pitch 1.1 inning and save the other for the next day (or extra innings)?
In fact, the next day, the Angels and A's were tied 3-3 in the 7th facing the top of Oakland's lineup. This is about as high-leverage as it gets, and who does Scioscia call on? Kevin Jepsen, who promptly surrenders 2 runs.
Utter foolishness.
Cecil Cooper is just (Dusty Baker - Barry Bonds). And I thought Trey Hillman had more sense.
>.>
the late knockout punch
You never see football coaches call defensive timeouts like you do in basketball. But instead, football coaches always save their TOs for offensive possessions. Such bull.
You can make the argument that using an inferior pitcher in a high-leverage situation could be a legitimate probability play that panned out poorly, rather than simple bad judgment. I'm probably giving Hillman too much credit, but I do think there are valid excuses for not having your best pitcher on the pound with the game on the line. Of course, I stil can't come up with an excuse for using Farnsworth in the 8th inning to begin with.
In a close game, in the 7th inning or later, imagine yourself on the other team. What pitcher on the opposing team (that would be your Royals, Trey) would you least prefer to face right now?
If the answer is Joakim Soria, then maybe he's the pitcher you should have in the game right now.
I'm not really defending Hillman - he had better options available, and failed to use any of them. But you DO have to find out how a guy is going to handle a situation when it comes up - both the manager AND the pitcher. And occasionally, it is going to cost you. Point being, if you can see that it is costing you more times than not, then you have a far sounder basis to make the call to eat a contract or cut bait (as I expect the Royals are going to have to do eventually, although I'd love to be wrong).
[boggle]
On the importance of April, one fun thing to do would be to write up a summary of the past season, played in reverse. Where the "important" games that "change the races" and have "clutch performances" are all the April games.
I'd like to see more closers get up in the 8th inning at the sign of trouble, and I'd like to see more of these guys get 4 or 5 outs. But you do have to take preparation time into account, as Davey Johnson learned at the WBC (sigh).
Not sure where you get this from. Does Putz need less time now that he's a "setup man" than when he was a "closer"?
The thing that drives me more nuts than what Joe is saying is when Leyland will pull a pitcher who's doing really well and replace him with another in order to satisfy a "righty-lefty" or "lefty-righty" match-up. Seay (a lefty I believe) was pitching really well against his *one* batter, then Leyland pulled him because a left-handed batter was coming up. He then put in Lyon, and you could feel through the TV that Lyon was going to blow it. (Please correct me if I have the "left/right" thing screwed up; I get these confused.)
A job of a pitcher, as has been said, is to throw strikes and get people out. If the pitcher is smokin', he should stay in regardless of the batter he's facing.
The one thing Joe doesn't mention, which matters moreso for the 7th inning than the 8th, is that if you give up runs in the 7th, you have at least two more chances to get them back, whereas if you give them up in the 9th, you often don't. That doesn't mean all closers should be saved for the 9th, it just means you have to consider that the marginal value of giving up a run in the 9th is higher than one in the 7th or 8th.
Hypothetically, let's assume Trey Hillman started using Joakim Soria only in the most high leverage situations. The first time that Juan Cruz or Kyle Farnsworth blow a 3 run lead in the 9th inning against the bottom of the other team's order, Hillman would be hammered by the conventional baseball for not using his closer in a save situation. To the contrary, when Trey Hillman suboptimally uses Kyle Farnsworth in the 8th inning to pitch to Jim Thome, the story is about Kyle Farnsworth's failure, not Trey Hillman's.
To change bullpen usage patterns, BP's target audience should be the conventional baseball media (which is slowly coming around on things like OBP, for example). Bullpen usage patterns won't change until the Bill Plaschke's of the world start asking the right questions, and not giving Trey Hillman a free pass for decisions like yesterday's debacle.
The evolution of the relief pitcher is very interesting. Use of relievers has changed throughout baseball history and I think it will again. The media tends to be the followers here and not the leaders. Once a manager makes a new use of relievers successful, the media follows along.
Today's Braves game was a perfect example of a great time to bring in Gonzalez (in the 7th). He could have faced the lefties and chances are he would have gotten out of the inning with the lead. And they still had Soriano to close. Cox ain't going to get fired anytime soon most likely. And the Braves really have two closer type guys anyway.
(The Braves' weaknesses were really exposed in this game. Other than fragile armed Gonzalez and Soriano, their bullpen is pretty shaky. And Chipper and Anderson are hurt already).
FWIW, the Red Sox did try something along the line of this model in 2003, it didn't really work, and they abandoned it, without it costing Grady Little his job. He took care of that all on his own later in the season. They got hammered for it in the Boston media, but Epstein's ownership of the idea meant that it was harder to direct criticism at the manager.
If a player fails in a situation he shouldn't be in (e.g. Farnsworth v. Thome), it's the player's fault, because he didn't perform his "role." It's not the manager's fault for putting him in that situtation.
Knowledgable fans see through this farce, but knowledgable fans are a distinct minority.
I always thought the home team should play "Knock Three Times ", or, " Tiptoe thru the Tulips " or some other nonsense when the visting team's closer came in looking all mean.
What I would be interested in knowing is how many games are lost due to the " closer " NOT coming in earlier ?
However, the " closer ", in a lot of cases, is not the team's best relief pitcher, so maybe NOT coming in early has actually " saved " ( new stat ) some games, and it all evens out.
Again, great article.
Again, this is not to say HIllman couldn't have used Soria to get an extra out the other day. I just think that Joe is oversimplifying the April vs. September argument, and weakening his overall point in the process.
The Nationals' highest leverage games are in April, and the Phillies' highest leverage games are in September.
The standard deviation of any team's performance relative to its abilities is about 6 games over the last 159 games of a season. Over the last 20 games in a season, the standard deviation is a just 2 games. There's a gigantic difference between a two-game swing when the vast majority of all teams will be within 3 wins of each over the last 20 games, and when most pairs of good teams will NOT be within 3 games of each other over the last 159 games.
What I really want to see is an expansion of his willingness to pull Percival if he gets into trouble in the 9th. Last year he did that, but only under extreme duress and with the excuse that Troy was hurting. If he can incorporate that into normal practice he will have moved very far from the current irrational bullpen usage.