The Rays were down 0-2 in the best-of-five American League Division Series, and now they're not. That's maybe overly simplistic, but that's where we are, thanks to a Game 4 bullpenning masterpiece and one of the greatest defensive relays you'll ever see.

Now, we're on to Thursday's decisive winner-take-all Game 5, where the Rays will be presented with a nearly impossible task: Beat Gerrit Cole. That's not hyperbole, either; the Astros haven't lost a game started by Cole since July 12 -- and even in that one, Cole left with a 7-4 lead before the Astros bullpen kicked it away. (That happened in his previous-start loss, too, on June 19; Cole hasn't left a game the Astros would lose without a lead since May.)

It's a tough ask, is the point, because Cole is a good bet to win the AL Cy Young Award, and because the Astros outscored the Rays 9-3 in the first two games of the series in Houston. At FanGraphs right now, the odds are 65/35 in favor of the Astros. That makes it less than a coin flip, but it's not impossible, either. Don't forget: The Astros and Rays started the season by playing four games, two started by Cole and Justin Verlander. Tampa Bay won three in a row after losing on Opening Day. These things can and do happen.

So as we look ahead: What if the Rays complete a comeback from down 0-2 to win this series, 3-2? What does it mean not only to do that, but to do it against a 107-win Astros team that very legitimately might be the best baseball team ever?

• TB-HOU: Game 5 lineups, matchups, FAQ

There's no perfect way to quantify that, of course, but we can take our best shot. The first thing we can tell you is this: Coming back from down 0-2 to win three straight is pretty rare, regardless of who is involved. There have been 133 five-game playoff series to go the full five games over the years, and 82 of them started out with a 2-0 lead. (This includes the first 15 years of the LCS, which was a best-of-five through 1984.) Of those 82, they break down like this:

• 72 of the 82 ended in a three-game sweep (88%)

• Of the 10 comebacks, six of them were from the team with three home games

• Of the 10 comebacks, four of them were from the team with three road games

Tampa Bay, those are your odds. We've seen the team without home-field advantage come back from 0-2 to win 2-0 just four times in 133 five-game series, or about 3%. If the Rays want to join the 2017 Yankees (ALDS vs. Indians), 2012 Giants (NLDS vs. Reds), 2003 Red Sox (ALDS vs. A's), and 1999 Red Sox (ALDS vs. Indians), this is how hard it's going to be.

We could stop there, probably. That's impressive enough. But let's try to go a little further, because again, this isn't just any opponent. This is the 107-win Astros, and only a dozen teams in the history of baseball have won more regular-season games than that. If we look at the 10 comebacks from 0-2 to 3-2, would the gap in winning percentage between the .660 Astros and .593 Rays rank at the top?

1) 2017 ALDS: Yankees (.562) over Indians (.630)
Difference: -.068

As you'll see in a second, the Rays would be just .001 behind, though you'd be forgiven for giving the Rays a tiebreaker boost because the 107-win Astros won five more games than the 102-win 2017 Indians. That said, remember that this was the Indians team that had won 22 straight games in August and September, had Corey Kluber winning the AL Cy Young Award, a full season of Andrew Miller dominating in the bullpen, the breakout season of José Ramírez, and on and on. They'd even won the first two games of this one at home, before dropping the final three. Those 102 wins, by the way, were the second-most in Indians history, behind only 1954's 111-win team.

2) 2019 ALDS: Rays (.593) over Astros (.660)
Difference: -.067

OK, we're getting ahead of ourselves, since the Rays still need to, you know, win Game 5. But if they do, this will stand as arguably the most remarkable 0-2 comeback we've ever seen in a five-game series. Remember, this doesn't even give extra credit to the fact that the Rays would be one of the few teams to do this on the road. If Rays can pull this off, they would have completed one of the rarest comebacks in playoff history; they'd be the largest underdogs, by regular-season winning percentage, to ever turn around an 0-2 deficit into a 3-2 win.

For what it's worth, it goes a little deeper than that, too. Back in Game 3, when the Astros were up 2-0 in the series and up 1-0 in the game after Jose Altuve's first-inning home run, Charlie Morton issued a second-inning walk to Robinson Chirinos. At the time, based on the inputs -- up 2-0 in a best-of-five, up in the deciding game -- the Astros had a high-water series win probability mark of 89.4%. That is, teams in similar situations go on to win the series nine times out of 10.

If the Rays win -- and assuming no Game 5 insanity where one team is down by something like five runs in the ninth inning and comes back -- their resulting low-water mark of 10.6% would be one of the 10 biggest comebacks in any five-game series, regardless of being down 0-2 or not. This is a good reminder that the 2015 Astros, the ones who kicked away the ALDS to the Royals, blew the biggest possible advantage: Their high-water mark was a 99.2% series win probability.

3) 2001 ALDS: Yankees (.594) over A's (.630)
Difference: -.036

This is the series you will forever remember for "the Derek Jeter flip play," and correctly so, because it's one of the most iconic plays from a certain Hall of Famer. That one came in Game 3, preserving a 1-0 lead, though it's worth remembering the Yankees weren't just the underdogs in the series, they were underdogs in the season. While the Yankees won the AL East and the A's did not win the AL West, that's not really the entire story: The Yankees won 95 games, putting them 13 1/2 up on the Red Sox; the A's won 102 games, putting them 14 games behind the 116-win Mariners.

4) 1984 NLCS: Padres (.568) over Cubs (.596)
Difference: -.028

Each one of these series seems to have that moment, though you may not remember Cubs first baseman Leon Durham pulling a pre-Buckner by allowing Tim Flannery's seventh-inning hit to go through his legs, allowing the Padres to score the tying run and setting themselves up to score three more times and head to the World Series. This one ranked high on the pre-2016 indignities Cubs fans had to suffer; the 96-win Cubbies hadn't won this many games since 1945.

5-t) 2012 NLDS: Giants (.580) over Reds (.599)
Difference: -.019

This was a weird one -- the road team won all five games -- and the Reds somehow won a Game 1 where Johnny Cueto threw only six pitches. The Giants had won 94 games that year, but this was supposed to be the Reds team -- the one with Cueto, Aroldis Chapman, Homer Bailey, Joey Votto and Jay Bruce at or near the height of their powers. The last time the Reds won more than 97? The Big Red Machine, back in 1976.

5-t) 1999 ALDS: Red Sox (.580) over Indians (.599)
Difference: -.019

AKA, "The one where Pedro Martinez came in to throw six innings of shutout relief in a decisive Game 5." That the Indians blew a 2-0 lead -- with home-field advantage, no less -- was especially crushing, since this was the final truly great year of the late-1990s Manny Ramirez/Jim Thome/Roberto Alomar and friends crew.

7) 2003 ALDS: Red Sox (.586) over A's (.593)
Difference: -.007

We are obligated to refer to the 2003 Athletics as "the Moneyball A's," so, here are the Moneyball A's. If there was any gap here, it's a small one, because the A's had won just a single game more than the Red Sox did, and Boston's +152 run differential was superior to Oakland's +125.

8) 1995 ALDS: Mariners (.545) over Yankees (.549)
Difference: -.004

This was one of the most exciting postseason series of all time, and played between two essentially even teams. Both clubs won 79 games in a lockout-shortened season, though the Yankees had a slightly better winning percentage due to having played one less game. The Mariners had a better run differential (+88 to Yew York's +61), though it obviously didn't matter in the end.

Now, we're moving in the other direction -- the gaps here may have been larger, but these teams were the favorites to begin with, and came back from the 0-2 holes they'd put themselves in.

9) 1981 NLDS: Dodgers (.573) over Astros (.555)
Difference: +.018

This wasn't a true "division series" as we understand it today -- the 1981 season was fractured by a strike, and so this was the result of the champions of a "split season" -- but it was also a little different because Houston had the first two games at home and then the final three were all in Los Angeles. The Dodgers had been fantastic in the "first half" (36-21) before playing just OK in the second (27-26); the Astros had done the exact opposite.

10) 1982 ALCS: Brewers (.598) over Angels (.574)
Difference: +.024

Maybe we don't talk about "Harvey's Wallbangers" enough anymore, but this was one of the best Milwaukee teams in history before their recent run of success. Only twice, when they won 96 games in 2011 and '18, have the Brew Crew bested 1982's 95-67 record.

11) 2015 ALDS: Blue Jays (.574) over Rangers (.543)
Difference: +.031

You remember this one, too, because this is when Jose Bautista bat-flipped his lumber into outer space after turning a 3-2 deficit into a 6-3 lead in the seventh inning of Game 5. Both teams had won their divisions, though the Rangers had done so with only 88 wins, one of the lesser records for a division winner in recent history.