Is the Blue Jays' rotation full of upside or just really lucky?

— --
You're asked to put together a rotation for 2017, and because you're in no way qualified to run a major league franchise, you predictably construct a mess.

One pitcher is a recent top prospect who has pitched well in relief but, when his team has tried him as a starter, he has produced some of the worst strikeout, walk and ground ball rates in the league. He's like a late-model Trevor Cahill, but a lot younger, or like Aaron Blair, but maybe a little better.

A second guy is in his 30s, long removed from the days when he would have been a top prospect if he had been one, which he wasn't. Thanks to a strong second half, he was, on the whole, about average last year, but before that, it had been a half decade since he'd come close to that standard. Nine months ago at the trade deadline, he was swapped for a nondescript minor leaguer, straight up. He's Ricky Nolasco, but maybe a little better.

A third guy takes the ball every fifth day, and has for years. He does it with a league-average ERA and has for years. He's also one of the oldest active players, so you worry about him like you worry about taking expired antihistamines. He's R.A. Dickey, more or less, except this guy is only 41, and Dickey is going to be 42.

A fourth pitcher is 32 and, until this past year, he had never thrown enough innings in a season (162) to qualify for the ERA title. Sometimes he has been hurt, sometimes he's been bad, sometimes he's been a low-leverage reliever, and sometimes he's been sent down to the minors. He had a good year last year, though he did it with one of the lowest ever batting averages on balls in play. Even you know historically low BABIPs are flukes, and you don't seem to know anything! Say you acquired Dan Straily.

Finally, your ace: Another former top prospect, still young, and he has been successful in the majors. But he missed almost all of last season with an injury, and his entire major league career is less than 200 innings old. Let's call him Aaron Nola.

So that's your 2017 rotation: Aaron Nola, Dan Straily, R.A. Dickey, Ricky Nolasco, Aaron Blair. You are about to get pounded. You should have tried harder.

In the early spring, Baseball Prospectus' projection system PECOTA looked at a version of that rotation and said, more or less: You are about to get pounded. Aaron Sanchez, J.A. Happ, R.A. Dickey, Marco Estrada and Marcus Stroman were projected to start 134 games and be, collectively, the worst rotation in the American League East, producing less value as a unit (about 4.1 wins above replacement player) than the erstwhile Blue Jay David Price was projected to produce (4.8 wins) all by himself in Boston.

Instead, as Toronto prepares for an American League Championship Series matchup again the Cleveland Indians -- the Cleveland Indians who appeared to have the league's best rotation entering the season -- all the roles have switched. Cleveland limps in without two of its three best starters, while the Blue Jays lean on a stellar starting rotation covering for an offense that scored 132 fewer runs than it did in 2015. The Blue Jays' rotation had the best ERA in the American League, and it did so while starting 81 times in one of the league's most merciless ballparks. Their starters threw more innings than any other team's, which was especially important because the Jays' bullpen had the fourth-worst ERA in the AL. The projections could not have been more wrong about the Blue Jays' starters, which might have been the point all along.

Luck is an uncomfortable thing to talk about in baseball. Ballplayers generally don't want to hear that they've been lucky -- because it devalues their successes -- or that they've been unlucky, because it spotlights their failures. But when we talk about luck, we should usually not be talking about the players at all.

Luck as an explanation usually comes up most often when we talk about unsustainable performances. The Texas Rangers won a disproportionate number of one-run games this year. That meant they had an excellent record even though they weren't better at hitting or pitching than their opponents. That's usually unsustainable. The best team, going forward, is usually the one that is better at scoring runs, better at preventing them, or better at both. "Winning" is virtually always a byproduct of these skills, not a discreet skill itself.

Luck comes up too when we talk about clutch performance. Some players have a record of hitting better in taut situations, or pitching better in the same. Do those same players have a higher chance of hitting better in the next tight situation? Most research comes up empty. It's usually not sustainable.

The Rangers, dispatched by these Toronto Blue Jays in the division series, won a ridiculous sum of one-run games this year because they performed well in clutch situations. During low-leverage and medium-leverage moments, they were an average offense and a below-average defense. But in high-leverage situations, as defined by Baseball-Reference, and generally corresponding with close games, the offense and defense were both much better. The Rangers were clutch, so they won the close games. Did they get lucky? Was Adrian Beltre just lucky?

Of course not. They might not be able to sustain it, but they actually did those things -- the situation asked, and they delivered. They are the actors in the situation. They had agency. They, and their opponents, were (excluding such irregular interferences as wind gusts, baseball gods and Chicago gamblers) the only ones who could affect the situation, and they did so. History screams "unsustainable," but history also records those wins dutifully.

There were other interested parties who didn't have agency. The Rangers fans, most important, didn't have agency. The Rangers front office arguably didn't have agency once the ball was in play. The executives put the team together, they helped make them good, they told everybody where to stand, but what happened in the batter's box with two outs and runners on second and third was entirely out of their control.

So the Rangers, as players, weren't lucky. The general manager and front office staff might have been lucky; probably were. The fans definitely were. They got to watch a winning team that might have been one April butterfly flap from finishing in fourth. There is no shame in saying Rangers fans were extremely lucky this season. The wins provide just as much fun.

Back to the Blue Jays rotation, though.

PECOTA, as noted, did not like this group of five as much as, in retrospect, it should have:

That actually undersells things a bit. Offense went up about a quarter of a run per game across the majors this year. If PECOTA had known that ahead of time, everybody's projected ERAs would have been a bit higher. Dickey and Stroman, then, were cumulatively just a little worse than their projections. As for the three starters tapped to fill out the rotation: Estrada, re-signed as a free agent last winter, was quite a bit better, while free agent signee Happ and converted reliever Sanchez went from fringe-starter projections to a couple of actual aces. The Blue Jays let David Price leave as a free agent and found, for less than what Price made this year, three pitchers who were about as good as Price was supposed to be.

Isn't that lucky? The Blue Jays didn't, after all, sign J.A. Happ instead of David Price because they thought Happ would be better, but because Price was asking for about six times as much money. If Price had told the Jays he would have given them a 90 percent hometown discount, the Jays would have obviously signed him instead. And they would have been worse. Luck is weird.

But the key thing here is that this is not uncommon. There were 119 pitchers who threw at least 120 innings in 2015. Some retired, some got hurt, some moved to the bullpen, and some were not invited back, but most -- 93 -- returned as major league starters in 2016. I broke down those 93 into five groups, from the best 20 percent all the way down to the worst 20 percent, based on their collective 2015 ERA .

(ERA is ERA adjusted for league and ballpark and scaled to the league average. An ERA of 100 is average; higher is better.)

That, as you can see, is a large spread. The aces are about a mile ahead of the tier below them. How did each tier do this year, though?

For major-league pitchers, there are powerful forces pushing their performances into extreme territories over the course of one season -- BABIP, health, the randomness inherent in small samples. But there are even more powerful forces bringing them back toward the league average in the collective long run. Even the best pitchers can be felled entirely by a weakened shoulder, and even the worst can have a season when the line drives all seem to be at somebody. These 93 pitchers' 2015 ERAs moved (in either direction) by an average of one full run in 2016. What J.A. Happ did -- or, for that matter, what David Price did -- is not an outlier. It's just pitching.

So if you know you're at the mercy of wild swings of "luck," it's not the worst idea to bet smaller sums on upside, which is a good way of describing what the Blue Jays did.

Sanchez was the 51st-best prospect in baseball heading into 2015, according to ESPN baseball analyst Keith Law, who had ranked him even higher the year before. Sanchez throws one of the hardest sinkers in history. He had struggled in a handful of major league starts, but it was only a handful, and as a reliever, he had dominated in the majors. He came to spring training boasting of a " total transformation" of his body.

Happ had been rebuilding his pitching style for years, and after a trade to Pittsburgh in July 2015, it clicked. Over the final two months, he had the third-best ERA in baseball and the 14th-best K/BB rate, wedged on the leaderboard between bona fide aces Corey Kluber and Jacob deGrom. It was way out of line with the rest of his career -- with nearly 1,000 major league innings -- but it demonstrated that he was at least capable of pitching like an ace for months at a time. That's something you'd never say about hundreds of major league pitchers.

Estrada had relied on a historically low BABIP, but his fastball was an outlier too -- a " cue ball" that gets more "rise" than any other fastball in baseball. The abnormality of that fastball movement seemed to defy normal BABIP laws by consistently missing barrels. Estrada allowed line drives less frequently than any player in baseball. Maybe the low BABIP he allowed wasn't a fluke but a skill -- like control or high strikeout rates -- that he could build a pitching career around.

Then throw in Dickey, but focus on his consistency, and Stroman, but focus on his youth and success and pedigree, and suddenly this is a rotation in which Dickey is the fifth starter and everybody else is good enough to start a postseason game.

It's an untraditional version of upside. Estrada throws one of the slowest fastballs in the game, and Happ has been for years the sort of generic low-90s veteran you expect to make fun of the Rockies for signing -- and it took imagination to see. And it could just as easily have amounted to nothing, in which case the Blue Jays might have missed the postseason with the AL East's worst rotation. But it was also a group capable of making the Blue Jays look very smart, and making them feel a little lucky.