Since mid-2015, when the recent Phillies hit their lowest point -- bad, expensive and with a barren farm system -- they have had four managers, not to mention five hitting instructors and five pitching coaches. "Managers," Roger Angell once wrote, "are changed whenever it becomes apparent that something must be done, even though it is almost always plain that nothing can be done."

But general managers are different. They're changed when it's apparent something must be done and something can be done -- which, these days, usually means tearing everything down, getting the team to be as bad as possible and then rebuilding. New GMs generally come with five-year leashes and five-year plans, and that plan is both demanding and forgiving: They don't have to build five good teams, just one -- the one that's playing at the end of the rebuild.

The Phillies hired Matt Klentak for that rebuild at the end of 2015. Like bank robbers digging underground to reach a vault from below, the Phillies dug and dug and dug -- but when they emerged from their tunnel, they were in the Rite-Aid bathroom across the street from the bank. To date, they are the first team in what we might call baseball's tanking era to fail at it.

The Phillies, still without a winning season since 2011, have reached the point where the proverbial final pieces -- the expensive veteran stars -- have not only been added but are leaving: catcher J.T. Realmuto is a free agent; starter Jake Arrieta's three-year contract is spent; closer David Robertson's two-year deal -- remember that? -- just expired; and infielder Jean Segura is reportedly coming up in trade talks. The Phillies might still get good, but they're no longer:

• young, as their hitters and pitchers were both older than league average in 2020;

• cheap, as their payroll is back in the league's top 10, where it was before they stripped down;

• rich in prospects, as their farm system ranked in the bottom 10 last year;

• clearly contenders, as (way premature) ZiPS projections for next year see the Phillies as currently constituted winning around 76 games.

They're also no longer led by Klentak, who stepped down as GM this offseason.

The tear-down-and-rebuild strategy worked for newly hired GMs in Houston, in Chicago, in Atlanta, and to a lesser but still consequential extent in Milwaukee, and it appears to have worked in San Diego and the South Side of Chicago. All those teams made the playoffs on schedule, and most ended up with their best teams in decades. So what happened to the Phillies, and what does it tell us about rebuilding?

The Phillies didn't botch their rebuild in any clear and obvious way. If teams embark on such a path -- and if they can justify it to themselves and their fans -- there are several steps where it can go wrong:

They can start the rebuild once it's too late, once there's little to trade for starter yeast. The Phillies started a little late -- it was clear at the 2014 trade deadline that they should have been sellers, but instead they held on to superfluous veterans such as Antonio Bastardo and Marlon Byrd. Even the pre-Klentak GM, Ruben Amaro Jr., said recently that in hindsight, he would have pushed to start the transition earlier. But they traded Bastardo and Byrd just a few months later, and by the time they tore down in earnest the next summer, they still had a lot of players other teams coveted: Cole Hamels, Ken Giles, Chase Utley.

They can fail, for whatever reason -- bad trades, bad drafts, bad development -- to improve the farm system. But the Phillies improved the farm system a ton. Their system was ranked 29th out of 30 organizations in 2012 (and 20th in 2015) and was a top-five system in 2016, 2017 and 2018. In 2018, when the club was on the brink of competing for the NL East, Philly probably had the second-best farm system in baseball, plus young major leaguers Rhys Hoskins and Aaron Nola, who had both recently graduated from that system.

When they start spending money on veterans again, they can get veterans who don't end up playing well. But while any collection of transactions will have hits and misses, the Phillies have generally added good major leaguers who played well for them: Bryce Harper has been one of the 25 best hitters in baseball as a Phillie, Zack Wheeler just got down-ballot Cy Young votes in his first year there, Realmuto was the best catcher in baseball for the past two seasons, Didi Gregorius had the bounce-back season the Phillies were betting on. Segura, Arrieta, and Andrew McCutchen have all been average major leaguers as Phillies, as was Carlos Santana for a season. That's about what those players were acquired to be. And when the Phillies were competitive at the trade deadline in 2019, the three cheap rentals Klentak acquired to improve the offense -- Corey Dickerson, Jay Bruce and Brad Miller -- all improved the offense, finishing with the club's three highest slugging percentages that year.

They can, as a team, badly underperform expectations. But the Phillies haven't really done that either: Over the past three seasons, ZiPS has projected the Phillies to win 189 games. They have won 189 games.

So what is to blame? It comes down to two things:

1. The prospects they developed didn't develop all the way

In other words: The Phillies made them prospects, but they couldn't make them major leaguers.

You can actually see this failure in two stages. In the first stage -- the pre-rebuild stage -- the Phillies had a weak farm system and nearly every player at the top of their system failed. Consider, for example, their prospects in 2013, when they were widely recognized to have one of the worst systems in baseball. This is one expert's ranking of their 10 best at the time:

1. Jesse Biddle
2. Maikel Franco
3. Adam Morgan
4. Roman Quinn
5. Tommy Joseph
6. Ethan Martin
7. Cody Asche
8. Jonathan Pettibone
9. Carlos Tocci
10. Shane Watson

That's obviously a bunch of duds. Collectively, they've thrown about 500 innings and batted about 5,000 times in the majors, and they've produced 0.2 WAR -- and not all of those 0.2 WAR for the Phillies, even. But nobody thought it was a good system, and "a bunch of duds" wasn't much worse than was expected. That's why you rebuild: to surround a bunch of duds with a rich farm system.

But here's the rich farm system, the top prospects from the 2016 system:

• J.P. Crawford
• Nick Williams
• Jake Thompson
• Franklyn Kilome
• Roman Quinn
• Cornelius Randolph
• Mark Appel
• Jorge Alfaro
• Andrew Knapp
• Ben Lively

Not much better! At The Baseball Gauge, there's a little tool that tracks the career production of each team's top 30 prospects from any year in the past. The Phillies' homegrown WAR has ranked near the bottom regardless of how well the farm system appeared to be doing:

2012: Ranked 29th at the time, 28th in WAR since
2013: Ranked 24th at the time, 23rd in WAR
2014: Ranked 25th at the time, 25th in WAR
2015: Ranked 20th at the time, 11th in WAR
2016: Ranked 4th at the time, 24th in WAR
2017: Ranked 5th at the time, 19th in WAR

There were clear successes from the farm system: Nola is an ace, Hoskins was a fifth-rounder and is now a middle-of-the-order hitter, and Sixto Sanchez -- now starring for the Marlins -- was converted by trade into superstar catcher Realmuto. But compare that to the young major league stars other teams developed (from drafts, international signings or trades) out of their trough seasons:

Astros: Carlos Correa, George Springer, Lance McCullers Jr., Alex Bregman

Cubs: Anthony Rizzo, Kris Bryant, Javier Baez, Kyle Hendricks, Willson Contreras

Braves: Ronald Acuna Jr., Ozzie Albies, Dansby Swanson, Max Fried, Ian Anderson

Brewers: Keston Hiura, Josh Hader, Brandon Woodruff, Corbin Burnes, Devin Williams

Padres: Fernando Tatis Jr., Dinelson Lamet, Chris Paddack

White Sox: Lucas Giolito, Tim Anderson, Luis Robert, Yoan Moncada, Eloy Jimenez

Put another way: The 2020 Phillies got about 3 WAR from players they "developed," a term we'll define to cover any player who originally lost his rookie status with the franchise. The 2020 White Sox got about 8 WAR from those players. Only seven wins separated the two teams' records in the condensed 2020 schedule.

The other thing that worked against them is a bit more random.

2. The fickle nature of fluctuation

Remember, we said that ZiPS projections have nailed the Phillies' collective win total over the past three years, on the dot. But it didn't nail the win totals in each of the three years. In one year, the 2018 Phillies were projected to win 72 and won 80. In another, the 2019 Phillies were projected to win 87 and won 81.

Maybe there's something about .500, about mediocrity, that these Phillies are simply drawn to. But also maybe there's a lot of chance in whether a team outperforms or underperforms in any given year, and if that's it, it's easy to imagine that the 2018 Phillies could have underperformed and won 66, while the 2019 Phillies could have been the overperformers and won 95. In which case, we wouldn't (yet) be talking about the failure of the plan but about how well it worked.

Or, consider this: The 2018 Phillies won only 80 games but were in first place into August, before a September collapse. The 2019 Phillies won only 81 games but were in first place before a summer collapse. The 2020 Phillies, for that matter, were two games above .500 and in the sixth seed in the National League's playoff race going into the final eight games of the season. They collapsed, going 1-7 and falling away from the eighth and final playoff spot.

Maybe the Phillies, as constituted, are built to collapse. But again, maybe there's a lot of chance involved here. If the Phillies' two good halves hadn't been the first halves of 2018 and 2019 but, say, the first and second half of one of those years, they're in the playoffs with 90-some wins. If COVID-19 had shortened either of the 2018 or 2019 seasons, instead of 2020, they'd have been in the playoffs. If COVID had shortened 2020 to 50 games instead of 60, they'd have been in the playoffs. If COVID had shortened it to 70 games instead of 60, they might well have been too.

There's obviously a lot more to five years of a franchise than we've listed here: The apparent failure of No. 1 pick Mickey Moniak and the apparent success of third overall pick Alec Bohm; the unexpected breakout of Rule 5 pick Odubel Herrera and the just-as-sudden collapse of his career around the time the Phillies poked their head out of that hole; Scott Kingery's pre-debut extension and subsequent oscillations; Vince Velasquez's health and the fact that the 2020 Phillies had the worst bullpen in 90 years; and, crucially, the way that having four managers and five hitting instructors and so on might have contributed to those good prospects failing to make the transition to the majors.

You can't predict a single day of baseball, and it's awfully hard to predict a year of it, or to predict what a single prospect will do or how a single veteran will age after signing a three-year contract. But the five-year teardown is a bet that, on a long-enough timeline, with an entire organization, you can tame that unpredictability. In a lot of ways the Phillies did: They got cheaper, they got younger, they got All-Stars and MVP candidates, and they even got quite a bit better than they had been in 2015. They got into three pennant races, sort of. But the unpredictability was always lurking. Nothing against the Phillies specifically, but that outcome should make us happy. Tanking to win later isn't a very fun strategy to cheer on in the first place, but it would be even more obnoxious if its success were inevitable.