Parker Messick went into Sunday’s Game 5 of the ALCS with a 3.09 postseason ERA, and on the surface the Guardians had little reason to worry. He lasted two innings. The Rays scored in each of the first three innings, Drew Rasmussen allowed one run over seven, and Tampa Bay won 3-1 to take a 3-2 series lead back to St. Petersburg. Cleveland needed six pitchers to get through the night.

It is tempting to call it a bad night and move on. However, the warning signs were there before he threw a pitch. And with the Guardians facing elimination tonight without their best starter, the way they arrived at this spot matters.

What the ERA was hiding

ERA tells you how many runs a pitcher allowed. It does not tell you whether he deserved to allow them. That is what FIP and xERA are for. FIP (Fielding Independent Pitching) only looks at the things a pitcher controls himself: strikeouts, walks and home runs. xERA uses Statcast’s quality-of-contact data to estimate what a pitcher’s ERA should be based on how hard and at what angle hitters are squaring him up.

According to FanGraphs, Messick’s 3.09 postseason ERA before Game 5 came with a 4.13 FIP and a 5.25 xERA. That is a gap of more than two runs between what happened and what the underlying numbers said should have happened. When you see a split like that, the ERA is usually the number that moves.

The biggest driver was command. Messick walked 15.1 percent of the batters he faced this postseason, compared with 6.6 percent during the regular season. A walk rate that more than doubles puts runners on base for free, and in a postseason where every run is hard to come by, free baserunners turn into runs. Sunday followed the script. He allowed five hits and two walks, struck out three, and needed 57 pitches to get six outs.

Don’t overcorrect

Here is the part that gets lost when a pitcher has a short outing in a big game. The same logic that said Messick’s postseason ERA was too good also says we should not read too much into his postseason at all. Before Sunday, he had thrown 11⅔ innings this October. That is a tiny sample, nowhere near enough to tell us that something has fundamentally changed.

The larger sample says Messick is very good. Over 32 starts and 186⅓ innings in the regular season, he posted a 2.56 ERA with 193 strikeouts and 49 walks. He got better as the year went on, with a 2.30 ERA over 13 starts after the All-Star break. Stephen Vogt argued he belonged in the Cy Young conversation, and that was not a manager just talking up his guy. A pitcher who walked 6.6 percent of hitters over a full season did not forget how to throw strikes in two weeks.

Regression works in both directions. The postseason ERA was likely to get worse, and it did. The walk rate is likely to get better, because the regular-season version of Messick is the real one. This is October noise, the kind of small-sample swing that happens to good pitchers every postseason. The problem for Cleveland is that October is all small samples, and there is no time to wait it out.

What it means for Game 6

The short start is a big reason Cleveland is in a tough spot tonight. The Guardians have already used 12 different pitchers in this series, and Sunday’s bullpen-heavy night came one game before an elimination game. Gavin Williams, who has struck out 33 batters in 16⅔ postseason innings, would have been the obvious choice, but Vogt said pitching him on short rest is “not an option.” Williams threw a career-high 184⅓ innings this season, and Cleveland is saving him for a potential Game 7.

That leaves the Guardians with a bullpen game, with Slade Cecconi a possible opener, against Freddy Peralta. Peralta has been nearly untouchable down the stretch. He is 6-0 with a 0.79 ERA in nine appearances since August 26, and he threw six shutout innings in Game 2 at Tropicana Field, where he hit 99.4 mph, the hardest fastball of his career. Cleveland’s offense managed five hits on Sunday. It will likely need more than that tonight.

History is not on Cleveland’s side either. Under the current 2-3-2 format, road teams that won Game 5 of a series tied 2-2 have gone on to win 21 of 28 times. The Guardians, who have not won a World Series since 1948, need two wins in Florida to keep that drought from reaching another year.

Whatever happens tonight, Messick is not a problem Cleveland needs to solve this winter. He is a talented young pitcher who ran into a small sample at the worst possible time. The ERA was lying in October. The 186 innings before it were not.

Data: FanGraphs, MLB.com, RotoWire, NBC Sports, theScore.