By In Stuff

The Three Toughest Outs

Does Robertson have the guts to get the last three outs? (AP) 

The other day, I heard an announcer call the ninth inning “The toughest three outs to get in baseball.” I will not name the announcer for a variety of reasons, the main one being that I could probably turn on the television tonight, scan the DirecTV baseball package and within a few minutes hear another announcer say the same thing, almost word for word. This “toughest three outs to get are the last three outs” is pretty well engrained into the grand baseball conversation.

Before I get into this, though — and, yes, we are going to go pretty deep into this — I would like to say a few words about how baseball’s immenseness and complications seem to lead people to just say stuff that sounds right in their minds. I thought I might have mentioned this a couple of weeks ago, but maybe I didn’t: I was watching a game on television and Yankees second baseman Robinson Cano struck out looking. Tim McCarver was doing the game, and he immediately said something like: “We just saw an extremely rare thing. Robinson Cano almost never strikes out looking.”

I thought: How could he possibly know that? Did McCarver happen to have a list of batters who rarely strike out looking right in front of him? Did one of his fine statisticians happen to scan that list the second it happened and whisper it in his ear? Did he recall the statistic from some previous conversation? Not impossible, I suppose. But not likely. I think it’s much more likely that McCarver simply deduced that Cano probably doesn’t strike out looking much. He probably could not remember seeing him strike out looking. And, hey, it does SOUND right. Cano doesn’t strike out all that much looking OR swinging (he has never whiffed 100 times in a year) and he’s a hacker, so it just figures that he probably wouldn’t strike out looking very often.
But the point is not whether he was right or wrong.* The point is, I’m guessing that McCarver had absolutely no idea if it was true beyond what his gut was telling him. I’m sure it just SOUNDED right. So he said it.

*It took a while, but I did look it up: Cano does not strike out looking very often, but it’s not exactly a Royal wedding kind of occurrence. Last year, he struck out looking 2.2% of the time, which is a fairly low percentage, but wasn’t in the Top 25 among hitters with 400-plus at-bats. The year before that it was 1.3% of the time, which did put him in the Top 10, but still well behind guys like Miguel Tejada and Jason Kendall. The year before that it was 1.6%, which put him about 20th. Like I say, Cano doesn’t strike out looking a lot, but it’s hardly noteworthy. Vlad Guerrero, now THERE’S a guy who hardly ever strikes out looking.

This is no great crime, of course. I think this tendency for people to say what just sort of sounds right to them — without actually checking to see if it is true — is an immutable part of the human psyche. I would guess that we all fall for it. I would also guess that in the last three days at least one person, and probably more than one — from an array of friends, relatives, neighbors, co-workers — has told you something directly in your field of expertise or line of interest that is undeniably and unquestionably wrong. But it SOUNDS right, at least to them, so they say it. And say it. And say it. And, often, they say it with conviction. I can only imagine how many times someone went up to Albert Einstein to tell him some right-sounding point about the theory of relativity.

This tendency is especially strong in baseball. I think it’s something about the game. The marvelous thing about baseball is that because the season is so long, you will see just about anything you want to see over 162 games. You will see hitters get a bunch of hits at one time (“The ball looks like a BEACH BALL to him right now”) and also go hitless for a long time (“He’s hit the ball hard a few times but the hits are just not falling for him right now”). You will see great plays and dreadful blunders, You will see games won and lost in every imaginable way.

And you can get carried away by any of those things. If you see a team steal seven bases, drive the pitcher and defense mad by taking extra bases, and score a couple of runs on errors, you can came away convinced that speed is how you win baseball games. If you see a team make seven diving catches and an outfielder take away a home run, you can come away convinced that defense is the only sure path to victory. If you see a plodding and patient team take 11 walks, knock the starting pitcher out in the fourth inning because of pitch count, foul off so many pitches that the other team begins to lose its mind, you might buy into plate discipline being the secret. Baseball’s ubiquity fills the mind with certainties that aren’t certainties at all.

The ninth inning idea — that the ninth inning features the toughest three outs in baseball — is like that too. We’ve all seen enough good and bad things happen in the ninth inning to come away thinking that it’s just DIFFERENT from all the other innings. This is why teams pay huge amounts of money for closers whose jobs (almost entirely) is to protect one-, two- and three-runs leads in the ninth inning. The latest talk in New York, as the city shakily tries to recover from the Mariano Rivera injury, is about David Robertson and whether he has the stuff, the guts, the pluck, the spirit, the moxie, the cojones, the courage, the bravery, the backbone, the grit, the fire, the fearlessness, the toughness, the determination, the audacity, the boldness, the spine, the mettle, the heroism and the nerve to actually get batters out in the ninth inning.

You know. The toughest three outs in all of baseball.

Except …

* * *

Hitting by inning since 1970 (batting average/on-base percentage/slugging percentage):

1st inning: .270/.342/.413
2nd inning: .254/.318/.389
3rd inning: .263/.327/.402
4th inning: .267/.330/.419
5th inning: .265/.328/.408
6th inning: .267/.333/.418
7th inning: .261/.331/.399
8th inning: .256/.327/.389
9th inning: .247/.316/.372

* * *

Now, before we get into this, let’s acknowledge that ninth inning statistics are wildly skewed. The home team does not bat in the ninth when it has the lead, so ninth inning statistics lean heavily toward road teams — ninth-inning stats are about 2/3 road teams. And road teams do not hit as well as home teams. Also, the ninth inning has been handled so many different ways over the years — there was an era where starters were expected to finish their games, an era of multi-inning firemen, an era of the closer, an era of the high-priced, one-inning closer — that you can’t really directly compare the ninth inning to any other.

That said, I didn’t make up the cliche. The cliche is that the ninth inning has the three toughest outs in baseball.

And when you look at the numbers you see it: The cliche is absolutely, undeniably, absurdly wrong. They are the easiest three outs.

Well, you can see it in almost any number breakdown. Since 1970 — this is through many different kinds of relief strategies — batters in the ninth inning have the lowest batting average, the lowest on-base percentage and the lowest slugging percentage for any inning. Pitchers have their lowest ERAs in the ninth inning. Fewer runs are scored per ninth inning than any other inning, and this goes back decades:

Average runs per inning (1-8) .47
Ninth inning: .43

Average runs per inning: .49
Ninth inning: .42

Average runs per inning: .53
Ninth inning: .45

Average runs per inning: .53
Ninth inning: .45

By the way, I don’t believe hitters are WORSE in the ninth inning than any other inning. I tend to think ninth-inning hitting and scoring is largely suppressed by those factors mentioned earlier — that winning home teams don’t bat in the ninth and that good pitchers/closers are used for one inning to finish games. My point is not that hitters are worse in the ninth, but they are certainly not any BETTER in the ninth.

But what interests me more than why those hitting numbers are so low is why this absurdly inaccurate cliche began in the first place. The toughest outs? Really? I mean, if you think about it: Why would those outs in the ninth inning be harder to get than outs in any other inning? What could have inspired such spurious thinking in the first place? I suspect the reasons why people ever believed this is because:

1. Late-inning pressure changes the dynamic for the pitchers (a.k.a., only certain pitchers can handle the tension of the ninth inning).
2. The hitters’ latent ability to hit in the clutch emerges at the end, with the game on the line, when it’s for all the marbles, when backs are against the wall, when it’s gut-check time and so on.
3. Managers will dig into their voodoo bag of pinch-hitters and base-running sorcery to score runs in the ninth inning, and such go-for-broke maneuvers make it harder to get those three outs.

But even if you believe these kind of nonsensical things, does it make sense even on the gut level? Wouldn’t pressure change the dynamic for both sides equally? Wouldn’t the hitter’s clutchability be matched by the pitcher’s? Wouldn’t the other manager be able to dig into HIS voodoo bag?

All of it seems kind of like pointless thrashing about to me. We have a remarkable ability to believe what feels plausible. And every one of us have seen our teams, many times, struggle against ninth-inning comebacks or put together valiant comebacks of their own. And such things stand out in the mind much more than valiant sixth-inning comebacks or busted rallies in the third inning. The ninth inning feels particularly in focus because the game is almost over, and all the other outs have already been exhausted, and this is the last chance. So when someone says, “You know, the ninth inning are the three toughest outs in the game,” it’s tempting to just nod because they are the only outs left. The cliche perfectly fits the mental picture. And then someone else says it. And someone else. And before long you find yourself saying it too.

But no matter how many times you say it or hear it, well, it just ain’t so.

ERA by inning since 2000:

1st inning: 4.81
2nd inning: 3.93
3rd inning: 4.44
4th inning: 4.54
5th inning: 4.49
6th inning: 4.66
7th inning: 4.27
8th inning: 4.14
9th inning: 3.71

* * *

You know what the three hardest outs are in baseball? Yep: The first three outs. This too has been consistent going back 50 years. I think it’s probably obvious why: It’s the one time in the game when the manager gets to determine EXACTLY who will come to the plate. Every other inning is a mishmash of lineup-orders … only in the first inning are you guaranteed to have the 1-2-3 hitters bat. Since 1960, about 10,000 more runs have been scored in the first inning than in any other inning.

I’ve heard from a couple of radical thinkers inside the game that teams might be better off to have official game STARTERS rather than CLOSERS, that is to say, one- or two-innings specialists who start two or three games a week and whose job is to get you off to a good start. Obviously stuff like this might work better in theory than in practice*, but in theory this is a very interesting concept.

*I’ve got a post coming up revolving around the poll in which Brilliant Readers voted on whether teams would win more, lose more or stay about the same if managers NEVER sac bunted or used the intentional walk. Obviously, this is also a skewed sample, but only 12% of you thought that a team might lose more games.

Beyond these sorts of theories, it’s helpful to think a little bit about how baseball works. Runs, as has been written before, are scored by a chain of events. Sometimes it’s a short chain — Raul Ibanez hits a home run. But more often, it’s a longer chain of events — a guy walks, the next guy singles, there’s a double steal, a sacrifice fly — and your best shot for a longer chain good of events is to have a series of your best hitters come to the plate. You would rather not have a pitcher hit in the middle of the chain, or a light-hitting middle infielder. Obviously, if a manager gets it right, the best hitters come up in the first inning, three of them at least. That gives teams their best shot at starting the chain. The second inning has always been a very low-scoring inning because (I suspect) the lower part of the lineup is likely to come up.

Whatever the reasons for this phenomenon — and I’m sure I’m missing some reasons — I’d love to create a cliche. I’d love for this cliche to be: The toughest three outs in baseball are the first three outs. It may not sound true. And it doesn’t really fill our irresistible need for clarity and finality — after all, getting the first three outs without giving up a run doesn’t mean you’re going to win, or even that you’re likely to win. You could get the first three outs without giving up a run and still lose the game 48-1.

Still, based on the numbers, it’s a generalization but also something close to a fact. The toughest three outs are the first three outs. Sometimes the true story isn’t as much fun or as emotional as the myth. That doesn’t make it any less true.

Print Friendly, PDF & Email

94 Responses to The Three Toughest Outs

  1. Gurm says:

    Excellent post, Joe. Could you get rid of the 9th inning skew by just looking at the road teams’ batting averages, and home teams’ ERAs over the period?

  2. aznguyen316 says:

    Also sometimes in the ninth you just don’t need that many runs to win. You see games end after scoring a run b/c that’s all you need rather than making a rally scoring 4-5 runs in the 1st inning or any other inning before the 9th. So the average run per inning vs the 9th may be skewed. Managers play to win the game in the 9th. Moving runners over at a cost of an out just to get the tying or winning run in.

    • drunyon says:

      Very good point. The rules of the game serve to suppress scoring for home teams in the bottom of the 9th. Case in point:

      Bases loaded, tie ball game, batter hits a line drive that the LF dives for and misses.

      In any other inning, that’s a triple; 3 runs score.

      In the bottom of the 9th inning? It’s a single; 1 run scores, game over. The other 3 runners on base don’t even get an opportunity to try and score.

      The bottom of the 9th inning is also the only inning where it can end before the team records 3 outs (since the game ends as soon as the home team takes the lead). I’d guess that on average, due to this reason, far fewer batters come to the plate in the bottom of the 9th compared to any other half inning, which would suppress scoring.


    • Moe Giblack says:

      These are excellent points that Joe’s analysis should have picked up.

    • Rob says:

      But give Joe credit. His blog “sounded” right.

    • migdal says:

      That may be true, but if that were the main factor suppressing 9th-inning run-scoring, we would expect to see the 9th inning have less scoring but *not* a lower AVG/OBP/SLG than other innings.

    • no base says:

      very interesting point! but that only count if you are tied and only need 1 run to win, isn’t that right? what if it’s 5-4 or behind by 1 run and scores 3 runs! is it the still won by 1 run, in this case 5-6 or do you get 5-7?

  3. aznguyen316 says:

    Also sometimes in the ninth you just don’t need that many runs to win. You see games end after scoring a run b/c that’s all you need rather than making a rally scoring 4-5 runs in the 1st inning or any other inning before the 9th. So the average run per inning vs the 9th may be skewed. Managers play to win the game in the 9th. Moving runners over at a cost of an out just to get the tying or winning run in.

  4. One more possibility to consider: teams (probably visiting teams) down by 3+ runs in the 9th and just wanting to hit the showers and get out of town. Any chance of breaking down these stats by inning and by runs up/down?

  5. Scoops says:

    While I agree with the idea that the last three outs aren’t necessarily the hardest, even in a close game, I’m not sure these are the numbers that prove it.

    Given that the 1970s to now have been largely dominated by the specialist closer, who comes in fresh, throws hard, and probably is the best pitcher in the bullpen, I would expect the ninth to have the lowest slash lines and ERA. I’d say that the numbers provided tend to offer proof that really good, strong, fresh pitchers tend to pitch in the ninth. Given the choice between Corey Wade in the sixth and Mariano Rivera in the ninth, which inning would you expect to have the lower slash line and ERA?

    I’m not sure how easy it is to look up, but I would think the best way to check on how “hard” an inning is would be to look up which section of the batting order bats most often in that given inning. How often do hitters one through five bat in the ninth? What about six through nine? The latter would generally make for a much easier inning than the former.

    • macomeau says:

      It might also help explain why the first inning is the toughest. In addition to the pitcher not being quite “ready” or “into the game”, he also has to face the best part of the lineup. That’s less likely in any other inning (unless the pitcher has a perfect game going, but then the batting team has other problems).

    • Unknown says:

      Yes, but that’s the point. As a manager you are likely to have that fresh pitcher, which means for you as a manager, those last three outs are easier to get not harder. Looking at it from both teams’ points of view shows that it’s easier to get the last three outs, not harder and certainly not the hardest. You might say “having that great one inning pitcher is the hardest thing to have on your pitching staff” instead.

    • macomeau says:

      Right, my point was just that the stats shown are likely to be unduly influenced by the fact that the pitching team’s manager gets to choose his best pitcher. The batting team is stuck with whichever part of the batting order is coming up (or pinch hitters, which have their own perils). If David Robertson gets the 3/4/5 hitters out in the eighth, and Mariano Rivera gets the 6/7/8 hitters out in the ninth, which of them likely had the harder outs to get?

      The argument with ERA and the slash lines can be boiled down to “the last three outs aren’t the toughest because the best pitcher is pitching.” My argument is “the last three outs aren’t the toughest because you’re unlikely to be facing the toughest hitters.”

      Call it another argument about the stupidity of the save rule.

    • spencersteel says:

      The first inning is undoubtedly the toughest as it is the only inning in which the batting team has total control over who the first three hitters will be, and unless you’re Jim Leyland and have some sick fetish about batting Don Kelly and his 285 career OBP leadoff, those hitters are typically going to be your best.

    • Rob says:

      I like the concept of batting your best hitter first, second best hitter second, etc. That way, over time, your best hitter gets the most at bats, your second best hitter gets the second most at bats… and so on. So, you end up with the best overall production, and theoretically, more runs & more wins.

    • Dodger300 says:

      How could it be that in the first inning the pitcher suffers because he is not “ready,” but in the ninth inning the pitcher benefits becasue he is “fresh?”

      Sure sounds contradictory to me. It’s like finding a passage in the Bible to support your own bias, whatever your bias happens to be.

      Or as Joe writes “The marvelous thing about baseball is that because the season is so long, you will see just about anything you want to see over 162 games.”

    • flynnie says:

      Because a starter hasn’t pitched in 4-5 days and needs time to settle in, while the closer gets ready quicker and knows he’ll only be out there for an inning, so he needs no need to pace himself. As McCarver would say, “You better get to this guy early!” As to the gratuitous Bible-bashing, I’m glad that the “God is love and they who abide in love abide in God and God in them” won out, so the authorities are not hunting you down to burn you at the stake for a heretic.

  6. Jere says:

    given that the first 3 outs are the hardest, doesn’t that remind you of the cliche that ‘if you’re going to get ‘to’ this pitcher, you gotta get him early’…there’s noting earlier than the first inning

  7. No one ever said “The last 3 outs are the toughest” as Mariano Rivera was mowing down the side in order. It’s always right after the closer lets a guy or two on base, which is a bit of confirmation bias.

  8. Agreeing with (half of) your point, a lot of long-time baseball pitching coaches state that the two toughest innings are the first and the last.

    That said, I don’t know that ERA by inning since 2000 is the way to go in proving the point that the 1st inning is the toughest. You could come up with point after point why that ERA is so high (and if you wanted, start with simply providing a list of every Royal pitcher who has started since 2000 – just the names, no numbers needed!). Also, just because the 9th inning has the lowest ERA doesn’t necessarily mean it’s the toughest either — it is entirely possible that it is the toughest, but teams have their best pitcher throwing it.

    The other inning one needs to consider as the “toughest” three outs is the inning after your team takes a lead. Let’s get some intern stats on that one…

    • John says:

      Your point about having the best pitcher pitch the 9th is exactly what I thought of. The hitters’ triple slash in the 9th is their worst since 2000, but how would that line look if they weren’t facing Mariano Rivera?

  9. B-Tank says:

    I’m not sure I agree with the “Runs, as has been written before, are scored by a chain of events. Sometimes it’s a short chain — Raul Ibanez hits a home run.” I don’t think a single event counts as a chain. That’s just a chain link. It may be splitting hairs, but nobody would look at a single link of chain lying on a table and think, “That’s a one-link chain.” Rather, they would just think, “There’s a chain link.”

  10. Jeff says:

    In regard to things that sound right, I heard Tom Verducci say the other day that “Derek Jeter pops up to the infield maybe once a season.” Now, I’m sure that Jeter doesn’t pop up a lot–most great hitters don’t–but once a season? Come on. Why not just say “Jeter doesn’t pop up much”? Why exaggerate like that?

    • Come to think of it, rarely do you even see infield pop-ups so why did he even mention something like that?

    • Nick O says:

      FWIW, per Fangraphs, Jeter’s infield fly ball % is 2.6% for his career, so that would make about 15 infield pop ups per year. The league average was 10% more or less every year going back til 2007.

      The guys with the highest percentages are guys who are really bad (Jack Wilson, Jeff Mathis, Willy Taveras) or hacktastic mid-range power guys (Chris Young, Scott Hairston, Edwin Encarnacion) who hit a ton of fly balls. The guys with the lowest percentages are generally line drive hitters with high BABIPs (Votto, Howard, Kendrick, Jeter, Hawpe, Michael Young, Joe Mauer). So Verducci was right that Jeter does not hit many infield flies, but was wrong on how many he hits a year.

      He would have been right if he had said it about Joey Votto though. Votto has hit exactly one infield fly since the start of 2010!!

    • Rob says:

      Jeter is a slap hitter, so he’s much more likely to hit a weak ground ball to a middle infielder. That’s the better point.

  11. Matt says:

    Of course the fewest runs are scored and the lowest ERA is in the 9th inning. Every walk off win has potential runs left on the table. The home team technically could score more because there hasn’t been 3 outs yet.

    • I thought about this … I’ll bet the effect is almost none. Because we’re only talking about walk-off wins — there aren’t many of those — and in those while there are some runs not being scored, there are also outs not being recorded.

      If a team scores a the winning run with out out in the ninth, thats 1 run in only 1/3 of an inning. That’s a 27.00 ERA. I doubt that winning teams in walk off games would average three runs per ninth inning.

    • rlc says:

      See, this is one situation where the cliché is completely accurate. It’s not just “tough” to get the third out in a walk-off, it’s literally impossible. (And since this whole argument turns on people using different senses of a word, I do mean literally). I’ve never seen a major league team denied the opportunity to get the second out of the first inning just because they had allowed too many baserunners, but almost every fan has seen it happen in the bottom of the ninth.

      And when it does…well, that’s the source of the cliché. It’s certainly the case that not getting the last three outs of a game – especially a game where you held a lead going into the last half inning – is emotionally “tough”. It’s a good candidate for the “toughest” outcome you can have. From there it’s an easy rhetorical jump to saying that avoiding that outcome is a “toughest” job.

  12. Robert says:

    I expect every inning is of roughly equal difficulty when you compare against the proper expectation. The best hitters always bat in the first (check), the bottom of the order is USUALLY coming around in the 2nd (check), things randomize over the next several innings with a small uptick as starters tire or the crappy part of the bullpen comes in (check), we noticeably trend downward as we get to the stronger part of the bullpen (check), and hitting falls off a cliff when the “best” reliever comes in, terrible defenders get substituted for, pitchers almost never hit, and we start losing ABs from home teams (check).

    Interesting to see the numbers. A fairer compare would probably control the 9th inning for home/road, but I wouldn’t suspect that that would change it being the most pitcher dominated inning. Would also be interesting to see AL/NL splits to see if/how badly the pitcher spot hurts run expectation in the 2nd and 3rd.

    • John Gregory says:

      In addition to losing ABs from home teams, when the home team *does* come to bat it’s because they are not in the lead. That would correlate with the home team not being the better of the two teams, on at least that day.

  13. sammyg says:

    Usually love your stat based arguments, Joe. And I totally agree with you that people frequently make up nonsense arguments to account for random events and are all too frequently ready to make silly statements based on small sample size. But this argument just doesn’t work for me.

    I love sabermetrics, but the one thing about it that always has rubbed me the wrong way is that sometimes it seems that people make arguments that seem to argue that there’s no emotional element to the game. That pressure simply doesn’t exist. That players don’t feel something tight in their stomach when the situation is big and the crowd is going nuts. I simply don’t buy that. It’s an argument you would never make for golf where almost every single stroke is seen as a reflection of how well the golfer is dealing with “the moment,” so why do people try to make it for baseball?

    Pressure is real. This is true in baseball as it’s true in all facets of life. Big moments will elicit an emotional response. Some people respond favorably some people crumble. Some people it’s a wash.

    SO, with all that said. I do feel like the pressure of being ahead in the 9th inning with the potential to blow the game is much much stronger than already being behind with expectations that you are gonna lose. In a must win game (game 7 or elimination game), you could probably say the pressure cuts both ways. But, I’ve always felt that the 9th inning pressure was much stronger on the winning team in the 9th inning.

    Maybe people will disagree. I’ve always felt that the pressure that you could “blow it” versus you could “be the hero” is a stronger more stressful pressure. It’s all about expectations.

    Also, like others have already stated. If you’re only down 1 run the offense team has a strong advantage. Scoring exactly 1 run is much easier than trying to score multiple runs.

    • Blotz says:

      “Pressure is real. This is true in baseball as it’s true in all facets of life. Big moments will elicit an emotional response. Some people respond favorably some people crumble. Some people it’s a wash. “

      I don’t think you are wrong, it’s just that I don’t think that said “9th inning pressure” has any significant measurable effect on how the vast majority of relievers perform in the 9th inning. For every stone cold hall of famer like Rivera, there is the stone cold fluke of Bobby Thigpen. In general I find that pitchers who can “handle the closer role” are much more common than people think. What colors the discussion of closers are the extremes. On one hand there is Eck and Mariano… the extremely rare elite pitcher who does it for a long time. At the other end are the guys bad teams like the Royals or Cubs shuffle through on a biannual basis. In the middle are the guys who can gear it up for a couple of years as really effective relievers no matter where you stick them, they get dubbed “closers” because that’s a position now, and they get overpaid by the Reds because every team is supposed to have one, I call this class of pitchers “Fransisco Cordova”. History has shown us that a whole lot of pitchers can be CoCo if given the ball in the often favorable situation of a 3 run lead in the 9th against the bottom of the line-up.

    • Chris says:

      Doesn’t the pressure go both ways though. I would expect that both the pitcher and the batter are feeling the heat in a 1 run game in the ninth inning. The game is within reach for both teams.

      I think the the reason sabremetrics doesn’t pay much attention to pressure or clutchiness, is because given a large enough sample size they just don’t show up in the stats.

      The most common example is someone like Derek Jeter, someone who in my opinion is incorrectly tagged as some great clutch hitter who always comes through when it matters.

      As a Yankee he’s had plenty of playoff experience, playing in 152 games or essentially a full season of baseball. His OPS over those 152 games: 0.839

      Compare that to his career regular season OPS: 0.833

      So it isn’t so much that he thrives under pressure, its that he’s a hall of fame shortstop who performs as such in the playoffs. He comes through because he’s a great player, not because of some mythical clutch power.

      Take even his career splits in the Clutch section on baseball reference. All are in the vicinity of his career 0.833 OPS, hardly raising his game.

      Pressure and emotion are part of the game for sure, I just don’t think they play the role that others seem to think they do.

    • macomeau says:

      I believe there is such a thing as pressure, and that it has an effect, but I also think that effect is probably overstated and overemphasized by people using it in arguments.

      If a player thinks it’s tougher to pitch in the ninth, it probably will be tougher for them to pitch in the ninth. Mechanically, it isn’t any tougher than any other inning, but the mind is a funny place.

      That said, people making the pressure argument seem to make out that the ninth is a million times more pressure-packed than the eighth, seventh or sixth. I don’t think that’s so. I don’t even think players who talk about pressure think that’s so.

      Ultimately, I think the difference just comes down to ability. Corey Wade doesn’t come in to a one-run lead in the sixth thinking, “Welp, we’ve got a few more innings, so I can give up a run.” He comes in and tries to get every batter he faces out. So does Soriano. So does Robertson. So does Rivera. Even Clay Rapada does that. The difference is simply that Robertson and Rivera are much better than Wade and Rapada at getting batters out.

      Maybe the guys who “can’t handle the pressure” of being a closer are really just the guys who don’t have good enough stuff. Maybe they even realize that themselves.

    • Bacong says:

      I agree with you that pressure is a real thing. Like the post above me, I also agree that is not a measurable. A good pitcher will perform. Pressure can get to a player from time to time but if the player is good then something like that will even out.

      Just because a 3 run lead in the 9th is lost doesn’t mean the closer was feeling the pressure. There are far too many variables in baseball to blame it on something completely subjective like they’re nervous or whatever.

    • spencersteel says:

      The trouble with “pressure” as we laypeople use the term is that professional athletes, while not entirely immune, are different animals. I always love the interviews with the first-time postseason pitcher about to take the hill and, well, how’s he going to handle the heat. The answer is he doesn’t care because he isn’t thinking about it. It’s akin to asking a world-class neurosurgeon if he can handle the OR. I’m not saying they don’t understand the gravity of the moment, or even feel butterflies. What I’m saying is that those things have a negligible effect on their performance, whereas I would likely wet myself just throwing out the first pitch. The guys who couldn’t handle the pressure? They washed out in the Sally League.

    • Bingo bango, they’ve been playing baseball their entire lives and see pressure in all innings, now all of the sudden the 9th is going to give you a heart attack? NTM they are being paid tons of money to get 3 outs. Also to the guy saying “sabremetrics”….all Joe did was look at the BA of each inning….

    • To be fair, I think that sammyg does have a point, in that numbers-based critical thinkers have a bias against factoring in “emotion” and “pressure”, particularly for baseball. This bias is usually justified, since everyone who constitutionally hates stats loves to spend a lot of thought on emotion, and use it as a red herring in arguments with stat-heads. The infuriating thing is that appealing to emotion too quickly can cause you to miss some rather enlightening low-hanging fruit that you could reach by checking to see if there are (maybe more sabr-y) numbers that will explain whatever it is that you explain by emotions. I remember a FJM post where somebody speculates that maybe one of the reasons for resistance to newer stats is that you get used to telling emotional stories to fill the gaps left by the older stats.
      It’s possible that now that the argument has progressed enough so that people are seeing the benefit of considering more than half a dozen numbers and the memories of our lyin’ eyes, there will be fewer people who refuse to consider like BABIP and WARP and VORP and stuff and insist that stories about “clutchness” be used to fill the gap instead. That way, we can talk about emotion and confidence without getting angry and dismissive.

    • An example of the low-hanging fruit I’m referring to: imagine a player that is improving, but “basic stats” don’t explain why. Maybe his rbi total is improving, and/or people are noticing him killing less rallies than before, but his BA and H numbers are the same. Nothing in the other two triple crown stats would explain why more guys seem to get around when he’s at the plate. If you give up on the numbers here, and appeal to his mental/emotional state, you could say: He could be getting more “clutch” as he gains more experience in the game!
      But that’s giving up too quickly. What if, of the balls he’s getting in play, more are fly balls and less grounders? To compensate for better contact, maybe he’s striking out a bit more; but his new style leads to much fewer GIDP. That would keep more rallies alive, and end less innings on run-cancelling forceouts, maybe even boost his RBI totals. So he is improving, but if all that is true, it’s harder to believe that his new confidence has mystically allowed him to magic his way to more RBIs – instead, probably he made a mechanical adjustment to his swing. Whatever his emotional state is, he’s improving because he is now a better baseball player, although if you don’t look past triple crown stats, it’s hard to explain why, since a difference of a dozen or two fewer GIDPs might not register on the casual memory.

  14. Aaron says:

    How does the average Leverage Index compare for each inning. If the LI is highest in the 9th on average you could somewhat make that argument depending on your definition of tough…

  15. Boy, it must be tough being Tim McCarver. Even when he ends up saying something RIGHT, major sportswriters still take him to task.

    Oh well, I guess (as Tom Lehrer once said) it’s more important to know what you’re doing than to get the right answer.

  16. TMcD says:

    As always, an interesting and enjoyable post but I’m more skeptical about this one than most of your work (other than why the first inning is a really tough inning–that I understand/see). The reasons for my skepticism are several and have largely been mentioned: in today’s era the “best” relievers pitch the last inning of close games so I would expect hitters to perform less well in general (or those pitchers won’t keep pitching in that inning); the data is a bit skewed because home teams don’t bat as often in the 9th (because they win more–which again, the “better” team’s data isn’t accounted for here); the walk-off factor (which I think you underestimate in terms of how often that happens; this could be found I imagine) ending a game without more runs scoring; and (I think) most importantly the difference in how hard it is to get those last three outs in a close game (use the save rules–3 runs or less) vs. other games. In a “close game” I suspect it is harder to get those outs but I’m not sure the stats will show it because of the closers who are in the game at that time (which makes much of the effort to prove this point–yours and mine–a bit circular).

    • adam says:

      Just curious, why do you suspect it is harder to get the last three outs of a close game? Do you believe the hitters get better? Do you believe the pitchers buckle under the pressure more than the hitters do?

    • Also a hitter has been out there all day where as the reliever has been resting, so it does make sense the pitcher has the advantage in the 9th.

    • TMcD says:

      I’m not sure that hitters or pitchers in general buckle more than the other–but you are right that my post reads like it assumes that hitters get better. Some hitters “get better” and so do some pitchers (and vice versa). But the pitchers who are most called on to get the final outs in the “close games” are the ones who have been proven to not buckle as much. Granted I am putting too much faith in the notion of what a closer does and how he is used.

  17. ralber says:

    Far be it from me to defend the woeful Tim McCarver, but I’ve always thought that the “toughest three outs” cliche referred to the pitcher’s viewpoint. Protecting a lead when the opponents are down to their last chance is bound to be a harder task than pitching earlier in the game, regardless of whether pitchers succeed or fail more often than in other innings.

    Also, a factor that went unmentioned in Joe’s post and the comments is the increasing (I think) use of defensive replacements late in the game. Surely a team’s batting success is going to be affected by replacing the slugging but slow LF with an all-glove, fast youngster.

  18. Dinky says:

    While I agree with your premise (and heard probably the same announcer use those words during a national broadcast this weekend, although mine clarified to say “The toughest three outs are the ninth inning of a one run game”) I must also note that your samples are almost entirely from the era when relief specialists in general and closers in particular were used. I suspect that if you go back to a decade (say, the 1930s, or perhaps the 1950s, trying to avoid WW2 skewing) when starting pitchers considered it a point of price to complete games, you’ll see some reversal of those numbers. Sure, the first inning is always going to be high, as starters try to adjust to the mound or maybe learn they just don’t have their stuff, but the last inning a starter pitches (pre-relief specialists) tends to be his worst ERA; it’s why he gets pulled. Thus, many times tiring starters began the ninth, facing hitters who have seen and adjusted to his stuff for three or even four at bats, and then he gave up some runs. I also expect that in the hitter’s eras before the 1960s, there were more blowouts, and starters with big leads were allowed to pitch the ninth where even giving up a couple of runs did not jeopardize the victory, back before pitch counts became recognized.

    That leads to a couple of thoughts:

    The toughest three outs is an artifact of an earlier age, like batting average. A more (or still) accurate phrase would be “The toughest three outs for a starting pitcher to get are the ninth inning of a one run game”.

    Relievers in general became more important as managers realized that starters don’t always (or even don’t usually, or perhaps shouldn’t) last nine innings.

    Once managers got over the hump of not using relievers, closers evolved as managers realized that they should use their best reliever in the inning which guarantees a win.

    I suspect that pitcher usage will continue to evolve as managers realize that the best time to use their best reliever might be with the bases loaded and nobody out in a two run game in the eighth inning (high leverage situation). Some relievers who put little strain on themselves (e.g. Mike Marshall, Charlie Hough, Dan Quisenberry, Goose Gossage, Mariano Rivera) can be trusted to pitch more than one inning, while others (Jonathan Broxton) can’t. There will always be early adopters who take the time to study and become convinced by new analysis of how to win baseball games (e.g. Earl Weaver, Davey Johnson, and now Joe Maddon). There will also still be late adopters whose talents lie in motivation rather than game management (such as Tommy Lasorda). The beauty of baseball is that there can be multiple paths to success.

    I’d prefer to get more like Lee Trevino, i.e. the toughest three outs to get are when one more bad outing will lead to you getting sent back to the minors and there’s no way you can afford your house and cars on a AAA salary.

    • spencersteel says:

      The current reliever usage pattern is directly attributable to La Russa’s decision to make Eckersley a one-inning closer around 1988. It worked, because Eck was well-suited to the role. Moreover those A’s teams were great and since they were winning 100 games a season, the idea was copycatted by everyone. Trouble is – as you point out – La Russa was pitching to Holtzman’s save rule and not to leverage, and it seems like MLB has been awfully slow to figure that out. I don’t even look at saves – it’s an invented stat that isn’t very instructive. I realize Mariano Rivera is a great closer, but he’s gone and the impact on the Yankees’ record will be small, if there’s any impact at all. Try losing Miguel Cabrera for the year and see what happens.

    • Mark Daniel says:

      The visionary Jim Leyland put his best reliever, Joel Zumaya, in high leverage situations in 2006 while using a lesser pitcher, Todd Jones as his closer. Both pitchers made 62 appearances in 2006, but Zumaya threw 83 innings to Jones 64. Zumaya had an era of 1.94 compared to Jones’ 3.94.
      Zumaya entered games 25 times with men on base, and 16 times with RISP. He also pitched more than 1 inning 31 times.
      Jones came into a game with men on base 3 times, twice with RISP, and pitched more than one inning only 4 times.
      This is a sort of proof of principle season, showing how using the better pitcher in more high leverage situations can work. Then again, Zumaya has had numerous injuries since then and the Tigers did choke away a large division lead in the final months. I think the concept can work so long as there is someone who decent holding the title of closer.

    • brhalbleib says:


      I don’t know whether Leyland was visionary or not. That kind of situation described has occurred quite often when the better pitcher is the younger, less established pitcher (see, e.g. NY Yankees 1996) and the closer is an “established” closer. Even the Royals are doing that this year with the (so far) lucky Broxton closing while the better pitchers (Crow, Collins, Holland after his stint on the DL) are getting the leverage situations in the earlier innings.

  19. Mark Daniel says:

    Anecdotal evidence suggests that the three last outs are the toughest outs. The game of baseball is well over 100 years old. Over that time, the game has evolved to where it is in its present day form. And right now every team has a closer, a high quality pitcher who specializes in 1 inning outings only at the very end of games when his team is ahead. These closers evolved to exist for a reason, and it may be because the last three outs are the toughest. I sort of think not,though. I think the reason the closer evolved is because not getting the last three outs when you are ahead is painful, and, if repeated often enough, can cause long term damage to a team’s psyche. Losing a game 4-3 when you fell behind 4-0 in the 1st inning is painful. Losing a game 4-3 when you were ahead 3-0 going into the bottom of the 9th is much, much more painful.

  20. wm says:

    How about this way of looking at it: which are the hardest three outs to get?

    How many times since 1970 has a team failed to get the first three outs of the game? Never. Not once has a team failed to get the first three outs of a game.

    How many times since 1970 has a team failed to get the final three outs of the game? Many times… every time the game has ended in a 9th inning walkoff, a team failed to get the final three outs.


  21. adam says:

    I am trying to think about how you could study this independent of context (i.e. ability of hitter, ability of pitcher, pitcher fatigue, etc). Some thoughts:

    1. Adjust for hitter/pitcher quality by using hitter & pitcher stats to determine “expected ERA” and then compare to actual. If it truly is tougher to get the last 3 outs then actual ERA would be higher than expected

    2. Could adjust for pitcher fatigue by only looking at first inning a pitcher appears in the game (only if he is there at the beginning of the inning). Of course this will make it difficult to study innings 2-4.

    3. ERA should account for walk-offs in at least some manner, right? If you score 4 runs over 2 outs the EAR is higher than 4 runs in 3.

    4. Not sure how you account for managers playing for only enough runs to win.

  22. Joe, you might be right, but this post doesn’t prove it. In the era of the closer, I would expect the batting average, etc… to be low in the ninth, that is the point of the closer. Also, there is the psychological aspect. When Rivera enters a game everyone in the stadium knows its over. That has an effect on how the hitters approach their at bats, how managers manage, etc… That does not make the outs easier to get.

  23. EdB says:

    Kind of a gratuitous slap at Tim McCarver. For one thing, his statement was correct — Cano rarely strikes out looking.

    And for another, I’d bet that McCarver has seen more games than most people, including you, Joe. (I don’t have the stats to back me up there, either, but I kind of feel confident in that). Certainly more than me. He’s not perfect, but he’s not a fool, and he knows the game. Jumping on McCarver is kind of a parlor game these days, and you’re better than that, Joe.

    • David says:

      Hear, hear.

      The man spent parts of FOUR FREAKING DECADES squatting behind major league hitters and trying to call the right payoff pitch. His gut is better than yours, Joe.

    • clashfan says:

      But the whole point is that everyone’s gut–even McCarver’s–is not a substitute for actually knowing. See the above note about how often Jeter pops up in a season. That announcer probably has a lot of baseball experience, too, and was off by an order of magnitude.

    • If McCarver knows a lot he should try sharing it with the audience, as opposed to the drivel that has been coming out of his mouth for the past twenty years.

  24. adam says:

    Is the 9th inning of a close game really such a high pressure situation for pro players? A quick glance at baseball suggest between 25 and 33% of games are decided by one run. Almost half of American league games had saves last year. Out of 162 games that’s quite a lot of close games. It would become routine for most people, and I suspect pro athletes are more confident than the average person – typically being the best athlete around growing up.

  25. Nick O says:

    Joe, your first posterisk begged the question: how often does Vladimir Guerrero strike out looking?

  26. Bill Reh says:

    I think that the idea of the 9th inning being more important or harder to pitch than any other is complete B.S. That being said, the reason that the era for inning 9 is lower is because the team that is pitching with the lead will bring in their best relief pitcher (closer) whenever they have a less then four run lead almost all of the time (whenever their closer is “available”) and most of the time when they have a four run lead. So in that sense, I think it’s hard to untangle whether or not the statistics are particularly telling.

    To me though the proof of the matter is this: if a pitcher would perform any worse – or better for that matter – in the ninth inning than any other, then what are they doing in the major leagues? If you’ve made it through high school ball, college ball, A ball, high A, AA, AAA, and have been in the major leagues for more than a month, how can you *possibly* not be trying your hardest to get every batter out? These guys – hitters and pitchers – have been through so much preassure and competition for so long that they have to be both talented and conditioned to deal with preassure or else they simply would not be good enough to stay in the major leagues. The same with “clutch hitting”. They’re *all* clutch hitters, otherwise they’de be in A ball or selling insurance.

  27. The real question, Joe (and this goes out to Neyer and Law and others), in my opinion, is why do you get upset by announcers, or anybody, spouting a harmless cliche? Or even hbelieving in baseball myths or legends or waxing poetic from nostalgia, whether ourmemories are correct or not? What’s the harm to you, at all, in people believing pointless, if somewhat false, cliches? I don’t understand the push to remove every ounce of sentimintality or personal belief from fans of the game.

    • adam says:

      It’s pretty simple. Some people get annoyed when announcers say stupid stuff.

    • Chris says:

      Because they aren’t true? Because they are used in seemingly intelligent discussion in which they have no place? You admit to them being both pointless and somewhat false. That answers your question right there. Why not get information that both has a point and is at least somewhat true?

    • James says:

      In this case, that perception has acclimated itself into how managers manage and how teams allocate their payroll. It isn’t an annoyance that ceases when you mute the broadcast.

    • An announcers job isn’t to give empirical evidence about baseball. By virtue of why they exist, announcers must communicate in hyperbole and cliche and generics. If a fan is relying on a home-team announcer to provide an education on the game, that fan is a fool. Honestly, what does it say about a person who has a visceral reaction when another, totally unrelated and irrelevant person spouts a cliche or genric comment not necessarily rooted in fact? In other words, who cares?

    • Chris says:

      Except Tim McCarver isn’t a home-team announcer, he’s a national broadcaster for FOX and most notably the World Series, in which baseball has the potential for more than the typical fans.

      The color man should be providing us with keen insights on the game, not trotting out well known cliches that probably aren’t even true.

  28. Thought provoking. But noticed nobody has asked a closer, pitching coach, manager, or former loser-turned pitching coach for his answer. I promise you it will be much different than all the theoretical and raw data provided here.
    As one closer once to,d me, the greatest thrill is walking into a clubhouse after protecting the efforts of his 24 teammates. The worst is doing the same after flopping in the final inning.
    Numberrs are great, but until there is a sabermetric for emotion, fear, and adrenaline, I’ll prefer to again base my opinions on what I see to be true, rather than what cold, lifeless numbers tell me what SHOULD be true.

    • adam says:

      The idea behind sabermetrics is to measure what already happened. A sabermetrician would look at this and think “if emotion, fear, and adrenaline make it tougher to get the last three outs, this must show up some way in the stats. how can I measure this?”

      If there IS a difference, it SHOULD show up in the “cold, lifeless numbers” unless the hitter and pitcher are affected equally and they cancel themselves out.

      Of course, there could be statistical differences for other reasons, i.e. players approach the situation differently (such as the manager putting in their best reliever). See the LeBron post & discussion for the basketball equivalent.

    • Chris says:

      Should? What are you talking about. The slash lines Joe gave us are true and do tell us what has actually happened.

      I guess instead we can assume you managed to watch every single ninth inning of every game over the time period Joe used, intently observing the emotion and adrenaline of each player therefore having the ability to judge the toughness of 9th inning outs.

    • James says:

      Oh, I’m sure plenty of people care about the quality of their announcers.

  29. I find it annoying that announcers, both play-by-play and analysts, who are ostensibly professional people, do not make the effort to learn about something that effects their profession. You expect a doctor to keep up on trends in medicine or a lawyer in law. Yet, many of these announcers are not professional enough-and they are making a lot of money-learn even enough about sabermetrics to discuss it intelligently. They don’t have to agree with all of it, but they should make some effort to understand it. As “clients” in a way, I think we have a right to expect announcers to have at least a passing understanding of what sabermetrics is all about rather than making fun of it or spouting meaningless and, in many cases, incorrect cliches. I think it’s the lack of thought and effort that bothers me more than anything. I don’t want to eliminate references to batting average, pitcher wins, and so forth, but why would it be so difficult for announcers to read a little about advanced statistical analysis?

    • Chris says:

      I agree completely. This is what I find so frustrating with anti-sabre people. They have no problem trotting out BA or Wins to make a point. They seem to forget that those are stats in the same way that OBP or OPS are stats.

  30. adam says:

    If you want to argue that getting outs in the 9th inning with a small lead is tougher because of adrenaline, pressure, emotion, etc……couldn’t you make the exact same argument that it’s tougher to hit in the 9th inning when you’re trailing, for the exact same reasons?

    Why would it be tougher to pitch but not to hit?

  31. Gary says:

    I think maybe the adjective is wrong. It isn’t so much that they’re the toughest outs as it is that they’re the most important or the most pressure-packed outs. I know as a coach, I felt more pressure protecting a lead with three outs to go. The reason is simple – if you don’t get those final three outs before the other team takes the lead, you don’t have any more chances. If you lose the lead in the third inning or the fifth or even the eighth, you have a chance to come back. Even if you’re the home team, you still have a chance to score again in the bottom of the ninth. So the most pressure is one you when you’re the visiting team with a lead in the bottom of the ninth.

    However, situations are different. There’s a lot less pressure with a six-run lead than a one-run lead. There’s less pressure when the other team has the bottom of their lineup coming up than the meat of the order. There’s less pressure when you’re able to put a pitcher out there who is throwing well than one who is struggling.

    Are those outs tougher than any others? Probably not. When you’re a coach, every out seems tough to get. But there certainly is more pressure associated with those final three outs than the first three.

  32. Luis says:

    The cliche I love that people keep repeating without a shred of data is: “Pitching is two thirds of the game”, then: “No, it’s three quarters!”. And undoubtedly somebody else will say: “No, it’s 90%”

    • John Gregory says:

      Bill James tackled that one in his early Abstracts, and I paraphrase it now and then when the subject comes up. “If it were 90%, you’d see certain things. You’d routinely see 9 hitters traded for 1 pitcher. You’d see the team with the best ERA win the pennant 9 years out of 10. You’d…” well, I forget the rest; he had a dozen or so, I think.

  33. Ed McDonald says:

    Maybe instead of using a closer baseball teams should use an opener? He pitches every other day and gets you those first three really hard outs and then your starter comes in and pitches into the 7th and after that you use your “clean up” crew.

    • adam says:

      Interesting idea. Innovative so of course no one will ever try. If the opener gets through the first inning in 3 batters maybe he should pitch the second inning as well.

  34. Chase says:

    A useful stat (I don’t know if it is possible to find) might be the average number of hitters per out. It might be skewed a bit by the higher propensity of a manager to sacrifice or intentionally walk a batter by a manager, but it would eliminate some of the other skews.

  35. JIM LOOMIS says:

    Name any other sport that can bring so many people with so many different ideas and observations into a discussion like this. Is this a great game, or what!!

  36. A Fly Moses says:

    Well, I’d argue that if you’re saying “the toughest three outs in baseball are…” then it’s neither the first inning nor the last inning. It’s the three (or four) at bats against the other team’s best hitter. E.g. if you’re playing the Reds, the three toughest outs are Votto’s at bats. Now if you’re saying, “the toughest inning to get three outs without allowing a run is…”, I’d lean towards the first for the reasons Joe mentioned. But technically, I’d say neither is correct.

    As for the argument about what the stats say: yes, the runs/inning are probably skewed low because the team’s best reliever is in, but this is mitigated by two factors (both of which are mentioned above): (1) he’s more often than not going to not be facing the opposing team’s 3 best hitters, and (2) 2/3 of the time it is a pitcher facing road batters, who perform worse than home batters. So more often than not, what you’re calling the 3 “toughest outs” are against the road team’s non-best hitters. Which doesn’t pass the gut or stats test.

  37. Unknown says:

    Is there a statistic that measures how many times the first inning went 1-2-3 versus a ninth inning that went 1-2-3?

  38. Doc Quinn says:

    Billy Wagner once said, “Those aren’t the last three outs. They are MY first three outs.” I guess he would know.

  39. Josh says:

    I tend to think that the last three are the hardest, just not on the pitcher. They’re the hardest on the fan, because if they go wrong your team loses, and that sucks. Therefore, they are the most MEMORABLE.

  40. Jim Norman says:

    a) Ninth inning is the hardest — for the fan whose team is clinging to a one-run lead (and for the announcer for that team).
    b) How often have you seen a pitcher yanked in the first inning? Surely this is a factor.

Leave a Reply

Your email address will not be published. Required fields are marked *