For starters, Terry Francona had a great regular season as manager of the Indians. He did a great job getting his team ready for each postseason series. He did a lot of things well. However, he might have also cost the Indians their best chance by his bullpen management in Game 4.
Joe Maddon is widely regarded as one of the best managers in baseball. He almost cost his team the series with his bullpen management in Game 6. These two very intelligent managers were both working on the same philosophy. It backfired in spectacular ways. I don't necessarily fault them for it; new things have to be tried from time to time. But while the usage of Andrew Miller and Aroldis Chapman worked for a while, teams would be wise not to replicate it.
Going into Game 4, the Indians held a 2-1 lead in the series. With the series already guaranteed to go back to Cleveland, Game 4 was not as important as Game 3 had been. But Andrew Miller had thrown only 17 pitches in Game 3 (getting the win) and was, as he should have been, available to work Game 4. Then the Indians took a 7-1 lead before Miller had entered the game. Rather than hold Miller back for up to three innings of work in Game 5, Francona decided to pitch Miller anyway. Two things happened. FIrst, Miller was clearly not as effective as he had been in the series to that point. His command wasn't as sharp. That happens. He threw 27 pitches. He gave up a run for the first time all postseason. Most importantly, he practically guaranteed that he wouldn't be available in Game 5.
When Game 5 rolled around, the Indians took an early lead. Trevor Bauer was working on three days rest after throwing 87 pitches (his postseason high) in Game 2. With the bullpen fully available (as they would have been if Francona had turned to Salazar and Merritt in Game 4, instead of Miller, in a game where the Indians led by three before Miller began warming up) Francona could have had Miller warming up when Bauer started the fourth inning, and brought him in at the first sign of trouble. Instead, Bauer was left out to finish the inning, gave up three runs, and the Indians lost 3-2. That got the Cubs back in the series and also resulted in the Cubs seeing more of Bryan Shaw, which would become important later.
In Game 6, Joe Maddon demonstrated that he had not learned from Francona's blunder, bringing in Aroldis Chapman with a five run lead and having him throw 20 pitches. In either the Game 4 or Game 6 scenario, an average reliever gets the job done. These were not high-leverage situations. Maddon can perhaps be forgiven, as it was an elimination game (and his team won, of course) but this was still the wrong decision. Riding one bullpen pitcher through a series was tried by both teams and found wanting.
But this was not the first time that using one pitcher too often led to problems. In a series we should remember well, arguably the greatest relief pitcher of all time found himself in his first ever Game 7. It was his fourth appearance of the series. His first four innings were perfect. In his fifth inning of the series, he gave up back to back singles, intentionally walked the bases loaded, and managed to get out of the inning with a line out and a strike out. While no one would have argued at that point that the Diamondbacks had figured Rivera out, it became apparent three days later that they had. Particularly for players like Finley (who singled off of Rivera in the eighth inning before the ninth inning implosion) and Gonzalez, they had seen Rivera quite a bit through the series. Relief pitchers, unlike starters, are not used to changing their game plan. Rivera didn't, and the Diamondbacks hit him. Miller and Chapman didn't, and they got hit.
So then Game 7 rolled around. The Cubs had seen a lot of Miller, Allen, and Shaw. The Indians had seen a lot of Chapman and Carl Edwards, Jr. Chapman was exhausted. The four "unhittable" bullpen arms in Game 7 allowed 10 hits. They allowed 6 runs. Chapman blew the save but poached the win, Shaw took the loss (a bit unfairly, as Francona made a second blunder by not pulling him after the rain delay.) With a 8-6 lead in the 10th inning, Maddon turned to Edwards, clearly his favorite right handed arm in the bullpen. And the Indians, having seen a lot of Edwards, very nearly came back (and probably would have if Mike Napoli had realized at any point over the last few games that swinging for the fences every pitch wasn't the way to go.)
Unfortunately, despite the failings of this strategy, look to see more of it going forward. Both pennant winners did it, so in a copycat league, everyone else will try to do the same if they get to the postseason. The issue is that very few teams have a dominant reliever like the Cubs and Indians did, and that even if they do, a team with multiple relievers they trust to get outs when needed will always beat a team that is over-reliant on one the way the Indians and the Cubs were.
This wasn't Francona's only failing. In a seven game series, knowing that he would be lucky to get four innings out of his starters in five of the games (Bauer and Tomlin both times, Kluber in Game 7) he underused his long relief. Ryan Merritt, who worked 4.1 innings of 2 hit ball in the ALCS, did not appear in the series at all. Danny Salazar worked just three innings over two games, and both games were well out of hand before he entered. Piggybacking Merritt and Salazar with Bauer and Tomlin would have provided him with more length, and saved Miller. Trevor Bauer should have started the 10th with a clean inning. But the biggest failure he made was his Game 4 use of Miller, and it was only Maddon's making the same mistake that enabled Game 7 to even be close.