How long can Gerrit Cole keep up his run of dominance?

NEW YORK -- Would you believe there was a point this season when Gerrit Cole looked fallible?

It's true. Check the game logs. During the first couple of months of the season, there were some dicey outings. Here's one: On May 22, Cole was outpitched by Ivan Nova of the White Sox. He gave up six runs over five innings and yielded homers to Jose Abreu and Eloy Jimenez. Cole took the loss that day, a 9-4 Chicago victory.

"I was just off," Cole told reporters afterward. "It wasn't a good night. I wasn't synced up enough in my delivery."

At that point, Cole was 4-5 with a 4.11 ERA, and if there was one label to hang on his season, it would have been inconsistent. Maybe it was due to those delivery-syncing issues, but Cole had peppered in some real clunkers with a few solid starts. Five months later, we know what has come since: nothing but synchronicity.

Taken as a whole, Cole's numbers since that lackluster outing against the White Sox have been nothing short of staggering. Including the postseason, he's gone 18-0 with a 1.66 ERA and 251 strikeouts in 162⅓ innings. The Astros have won 22 of his 24 starts during that span. An average game score, according to the Bill James metric, is 50. Cole has recorded at least a 55 in all 24 of those starts, with an average of 72.

"He's been on one of the most incredible runs I've ever seen, that the baseball world has ever seen," Astros co-ace Justin Verlander said during Houston's American League Division Series win over the Tampa Bay Rays. "If you could name a starting pitcher in all of major league baseball, you'd want him on the mound."

Cole's two postseason outings have been dazzling. He gave up one run on six hits over 15⅔ innings with 25 strikeouts against the Rays in Houston's first postseason series, the K total breaking the record for the LDS round by three. It's possible that, if needed, he could have given the Astros even more.

"I think if I'd have left him in for 130 pitches," Astros manager A.J. Hinch said after Cole's 107-pitch outing in the Game 5 clincher over Tampa Bay, "his 130th pitch would have been 99 mph, paint on the black. That's Gerrit Cole."

In other words, thus far, the post-May 22 Cole has showed up in October like a great scary pumpkin to put the fright in every opponent the Astros encounter. After the Astros outlasted the New York Yankees 3-2 in Game 2 on Sunday to even the AL Championship Series, the horizon looks bright for Houston.

Cole is slated to start Tuesday's Game 3 and would get one more start in the series if it goes long. If Cole remains as unbeatable as he's been since May, then Houston would need to steal only one other game from the Yankees, and Verlander has one more start to go. Not a bad position to be in for an Astros club that is just 4-3 overall in the playoffs and has been outscored 27-22.

When asked how significant it was for the Astros to recapture momentum in the series, Jose Altuve said, "A lot, believe me. Guess what? We got Gerrit Cole on the mound [next]."

You certainly can't blame Altuve for feeling that way. Cole has been awing teammates and opponents alike. After Game 2 in the ALDS, the Rays' Willy Adames called Cole's performance "majestic." It's a good word for what Cole has been doing each turn through the rotation.

Every so often, we see pitchers get on this kind of run, and it's always memorable. Cole hasn't been the only hurler to take on the untouchable look even this season -- the Cardinals' Jack Flaherty posted a 0.91 ERA during the season's second half. There have been other famous instances of "unbeatable" pitchers -- Jacob deGrom only last season. Jake Arrieta in 2015, Pedro Martinez in 2002, Orel Hershiser in 1988, Doc Gooden and John Tudor in 1985, Bob Gibson in 1968. And that's just a partial list.

However, with the Astros pinning so much of their confidence on Cole's perceived infallibility, it's fair to ask: Do pitchers who get on this kind of run typically carry their dominance through to the finish? Are they and their teammates left standing at the end?

Well, obviously we're talking about anecdotal evidence here because the historical sample size is going to be small almost by definition. But maybe looking at these proto-Cole examples will tell us something.

We needed a list to start with, so using Baseball-Reference.com's Play Index, we used the following parameters: any pitcher with at least 10 starts after the All-Star break who posted a sub-2.00 ERA with an OPS allowed that was no more than half of the league average, to add at least a modicum of league context. Only the most dominant need apply.

Going back to 1920, there have been 62 such hurlers. All the above-mentioned pitchers met those parameters, but that's still a long list. So we added one more item to capture the dominance aspect: those who averaged at least one strikeout per inning. That cuts the list down to 25, but only 14 of those pitchers were on teams that made the postseason. And two of the 14 are Cole and Flaherty, whose destiny we are still working out.

That leaves us with a tidy list of 12, a list that skews modern because of changes in strikeout levels. Sorry, Spud Chandler fans.

And we must note, Cole is an outlier even among this elite group in terms of strikeouts; as ever, we have to bear in mind the league context in which he is pitching. Cole averaged a mind-numbing 14.7 strikeouts per nine innings during the second half -- two more than any of the other 24 pitchers we were looking at before culling the non-playoff pitchers.

After all of that, let's delve into the mini-narratives of our dominating dozen, in chronological order.

Sandy Koufax, Dodgers, 1965

If ever there was a pitcher so good and so thoroughly responsible for carrying a good-but-not-great team to a championship, it was Koufax in '65. He's the first pitcher on our list, and he led the Dodgers to the 1965 World Series title. The last pitcher on our list is from 2017, and he did the same. None of the 10 in between pitched for eventual champions.

Koufax qualifies for the list because he had a 1.94 second-half ERA and thus satisfies the "hot" pitcher criteria we are using. But his ERA for the season was 2.04, and that was Koufax's highest during the span from 1963 to 1966. He was basically dead-on for those four years, the most amazing consecutive pitching seasons of all time. It cost him his career, though, as Koufax wore out his arm and called it quits in baseball's ultimate version of a mic drop.

But Koufax might have been at his apex in the '65 World Series against the Minnesota Twins. He went 2-1 in that Fall Classic, and that he lost a game was a fluke -- he allowed one earned run over 24 innings in the series. After dropping Game 2 at Minnesota, Koufax shut out the Twins on four hits in Game 5 at Dodger Stadium.

Then he did something that we will almost certainly never see again, not from Cole, not from anyone: Koufax came back on two days' rest and threw another complete-game shutout in Game 7, that contest in Minnesota to clinch the title.

Cole is on a mini-version of the four-year Koufax run, and at 29, he's the same age Koufax was then. And Cole is from Southern California, where Koufax achieved his greatest fame. But if Hinch wouldn't let Cole finish Game 5 of the ALDS, when he had 107 pitches on the board through eight innings, he's not letting him throw two complete games in three days in two cities. But man, wouldn't that be fun?

Bob Gibson, Cardinals, 1968

Gibson's second-half ERA (1.19) in 1968 was brought up a lot this season because it was topped by Flaherty. Gibson is on the short list of greatest World Series pitchers ever, posting a 7-2 record with a 1.89 ERA over nine starts in the Fall Classic. He won World Series MVP honors in 1964 and '67. And that total of nine starts? Gibson averaged nine innings in those starts.

Cole could conceivably throw two complete games in a Fall Classic if he is efficient with his pitches, but not three. During the ALDS, Hinch had to consider the next two rounds. Once we get to the Series, he won't have that concern and it will be that much more difficult for Hinch to take Cole out if he's dealing like we saw against the Rays.

What we won't see, most likely, is Cole getting three starts, though it's possible if it were to line up so that he could start Game 1 and Hinch really wanted to push the envelope. More likely is two Cole starts and some kind of valiant relief outing. But the Gibson example, as with Koufax, reminds us that starting pitchers just don't get asked to carry the Herculean workload they once were.

And note that the year that meets all of our arbitrary criteria is 1968 -- when the Cardinals lost the World Series to the Tigers in seven games. Gibson was great in that series, going 2-1 with a 1.67 ERA. But St. Louis ran into a pitcher just as hot: Detroit's Mickey Lolich racked up three complete-game wins and beat Gibson in Game 7.

And that last note carries the most relevance for Cole: When you get this deep into the season, the other guys are pretty good too.

Ron Guidry, Yankees, 1981

In 1978, Guidry had one of the greatest start-to-finish hot streaks ever, forever embedding himself in Yankees lore as "Louisiana Lightning" by going 25-3. He was never that good again, though he had some big seasons. His '78 run didn't make our list because of the K-per-inning requirement.

However, Guidry checked all the boxes in 1981 when baseball resumed play after a long labor dispute that ruined much of that summer. Guidry went 6-2 with a 1.74 ERA after the strike, with 64 strikeouts in 62 innings.

He ran out of steam in the playoffs to an extent, posting a 3.22 October ERA. The Yankees advanced to their only World Series of the 1980s, but they lost to the Dodgers in six. Guidry won Game 1 at Yankee Stadium but was outdueled by Jerry Reuss in Game 5 at Dodger Stadium.

So Guidry, that season, was not a red-hot pitcher who carried his team to the finish line, and that was no doubt a disappointment to Mark Cole -- Gerrit's father, who was a huge Yankees fan.

Randy Johnson, Astros, 1998

This run was the result of one of the great trade deadline deals ever, the one that shipped the Big Unit to the Astros in 1998 and helped propel the Killer B's-led Houston club into the playoffs. Johnson went 12-3 with a 1.37 ERA and a Cole-like 12.1 strikeouts per nine innings during the second half.

But ... Houston faltered in the NLDS, losing to the Padres in four games. Johnson allowed just three earned runs in 14 innings in two starts that series -- and lost them both. Kevin Brown got him in Game 1, and Sterling Hitchcock in Game 4.

Houston scored three runs in their three losses that series. Given the mostly down nature of the Astros' attack this October -- something that could end at any moment given the firepower of the players involved -- that illustrates the biggest flaw in the red-hot pitcher model.

At some point, you have to score some runs.

Johan Santana, Twins, 2004

Talk about red-hot: Santana went 13-0 with a 1.21 ERA during the second half in 2004 with 11.1 strikeouts per nine innings. Who else would you possibly have wanted on the mound in a playoff series?

Well, no one, but when we get to 2004, we're closer to a 2019 sensibility. Santana was terrific, allowing one run in his two starts. But one of those was a seven-inning stint, while the other was a five-inning outing on short rest.

In the latter outing, Game 4 of the ALDS against the Yankees, Twins reliever Juan Rincon was bombed after Santana left and the Yankees clinched the series.

The lesson would seem to be that if Hinch really wants to get a Koufax/Gibson effect from Cole's transcendence, he would need to ride him in a way that contemporary managers just don't do. And if he did, Cole's agent, Scott Boras, might be displeased since Cole has a mega-payday looming this winter as a free agent.

Rich Harden, Cubs, 2008

Man, October 2008 was a rough one for the Cubs. They led the NL in wins and the majors in run differential. They made a hugely impactful trade during the season to land Harden, who could be fleetingly and tantalizingly dominant. Harden went 5-1 with a 1.77 ERA over 12 starts for the Cubs.

Then ... October. Harden gave up three runs over 4⅓ innings during his only postseason start, and Chicago was swept in three games in its division series against the Dodgers.

Cole is more consistent and durable than Harden ever was, and given his performance against Tampa Bay, he's already cleared that first hurdle, the one that Harden couldn't -- translating his performance to the postseason. Staying hot when it matters most is exactly what we are searching for in this exercise.

Kris Medlen, Braves, 2012

People already forget how dominant Medlen was in 2012, before arm injuries wrecked his career. Medlen went 9-0 with a 0.94 ERA and a strikeout per inning during the second half that season. The Braves earned a wild-card slot and, of course, Medlen got the nod.

Well, Medlen gave up five runs (two earned) over 6⅓ innings, and the Braves lost to the Cardinals in that wild-card game. But as with Harden, it's not really an illustrative example when it comes to Cole, who has already cleared the hurdle of carrying his regular-season dominance into October.

Clayton Kershaw, Dodgers, 2015

Well, doggone it, Kershaw is a walking spoiler alert: In major league baseball circa the 2010s, one dominant starting pitcher is not enough to carry a team through a postseason run. We've seen Kershaw succeed and falter again and again in October. In 2015, he went 10-1 with a 1.31 ERA in the second half of the regular season. But the Dodgers lost to the Mets in five games in the NLDS. Kershaw wasn't the problem, but he wasn't the solution either, in the Koufax sense.

The issue with Kershaw is not that he hasn't often been very good in October, it's that he too often hasn't translated his regular-season excellence to the higher-stakes arena. Repeating this a third time: The 2019 version of Cole has already made that transition -- so far.

Speaking of translating performance to the postseason, we had these numbers lying around, and this seems like as good a place as any to drop them in.

Consider a pitcher's expected performance to simply be the runs per inning he allows over the course of his career. Then consider the runs per inning he's allowed during his postseason career. Take the difference between those two numbers times the number of postseason innings he's thrown, and you have a career estimate for postseason runs allowed over/under expectation, which we'll call runs saved.

Obviously the number of postseason innings is a key component -- the more innings you throw, the more innings you can improve or harm this measurement.

Here are the top five:

CAREER POSTSEASON RUNS SAVED
Mariano Rivera, 24.3
George Earnshaw, 20.8
Jon Lester, 20.5
Orlando Hernandez, 19.7
Dave Stewart, 18.7

And ... the bottom five:

Clayton Kershaw, -36.2
Greg Maddux, -18.7
Tim Wakefield, -16.0
David Price, -12.9
Jaret Wright, -11.5

Jake Arrieta, Cubs, 2015

Arrieta went 12-1 with a 0.75 ERA during the second half of 2015 and carried that level of regular-season dominance into the 2016 season, giving him one of the most amazing calendar years a pitcher has ever had.

A lot of good it did Arrieta in the 2015 postseason: The Cubs beat the Cardinals in the NLDS and lost to the Mets in the NLCS. Arrieta wasn't bad, going 2-1 with a 3.66 ERA during October. But he was in the exact position Cole is in now in that he had strung together so many lights-out appearances that it felt like you could pencil him in for two wins per series.

October is tough. If Cole can keep what he did against the Rays going against the much more potent lineup of the Yankees, only then can we start to dismiss these kinds of cautionary tales.

Corey Kluber, Indians, 2017

Kluber rolled into October after an 11-1 second half with a 1.79 ERA and 11.6 strikeouts per nine innings. Red. Hot. Pitcher. His postseason ERA: 12.79. Dominance, it would seem, is a fickle beast. The Tribe dropped their ALDS against the Yankees in five games, setting up New York's epic showdown with the then-Coleless Astros in the ALCS.

Stephen Strasburg, Nationals, 2017

This is another example in the Santana genre. Strasburg wasn't fully healthy during the second half in 2017, but he did make the 10-start minimum needed to qualify for this study. He went 6-1 with a 0.86 ERA in those outings. Then he was even better in the NLCS against the Cubs. He started twice, threw 14 innings and didn't allow a single earned run. And he still lost one of the two starts. This is the nightmare scenario for Cole: perfection not rewarded.

Justin Verlander, Astros, 2017

As mentioned, our list is bookended by two pitchers who carried late-season dominance into October and maintained it, then celebrated a World Series crown at the end. In this case, it was Cole's present teammate, co-ace and fellow Cy Young front-runner Verlander, doing the honors.

Verlander was 10-1 with a 1.95 ERA during the second half, with some of those numbers put up for the Tigers before the August trade that sent him to Houston. He won four games during the postseason with a 2.21 ERA, working in relief, completing a game, getting by with dominant stuff but also when he was a bit gassed.