Here at FiveThirtyEight, we’ve watched Chip Kelly’s 2015 campaign as Philadelphia Eagles head coach/general manager with a combination of fascination and horror. His offseason roster-shuffling was spellbinding; the season itself, not so much. And on Tuesday night, it all came to a screeching halt with the news that Kelly had been fired in the wake of Philly’s 38-24 loss to Washington, with one game left in the season.Although Kelly had two years left on his contract, the dismissal wasn’t exactly a surprise. According to Elo ratings, our pet metric for gauging team strength, the Eagles have been one of the league’s most underwhelming teams this season. It’s telling that Philly refused to even let Kelly play out the string in Week 17, as if to emphasize the magnitude of the disappointment — this is only the seventh time since 1970 that an NFL team has fired its coach with exactly one game left in the schedule.Despite the fanfare that accompanied Kelly’s transition to the NFL in 2013, Philly couldn’t sustain the flashes of offensive brilliance it showed during Kelly’s first season as coach. The team regressed each year under his watch, to the degree that — according to our expected points added (EPA) grades, which rate teams on a scale in which 100 is average and one standard deviation is worth 15 points — Kelly’s final edition of the Eagles was below average at running the ball and stopping the pass, and downright horrible when it came to passing and stopping the run. PER PLAY, HURRY-UPPER PLAY, NON-HURRY-UP Broncos20132425.7+1.3636.5-0.17 It bears mentioning that the league-average EPA on hurry-up plays is significantly worse than it is on other types of plays (shaded both ways by manic end-of-game scenarios), so we might have predicted that the Eagles would fare much worse while in blur mode, not better. Yet Philly managed to keep its hurry-up efficiency above league average and compounded that edge by running these up-tempo plays about four times as frequently as the typical team. This should have been a tremendous advantage.And if all Kelly did was run fast plays, we might not be discussing his firing right now. But aside from Kelly’s first season, Philly’s offense has been hugely ineffective when it isn’t pushing the tempo. A freakishly efficient showing on fast-paced plays last season helped temporarily mask that issue, but the Eagles offense cratered this year when its hurry-up efficiency dipped, even though it remained above league average. It’s hard to rely solely on your hurry-up when, even at its peak, hurry-up can account for only about a third of your plays.The irony, of course, is that Kelly’s failed Eagles tenure will probably be seen as a sort of referendum on his fast-paced style. And perhaps the types of players necessary to properly execute that formula are less suited to a conventional playing style, or the allocation of practice time required for their execution precludes a team from being sharp in a standard set. But Kelly’s fast-break offense itself turned out to be one of the Eagles’ greatest strengths. The problem, it turns out, was damn near everything else the team did. Eagles20143025.1+0.4434.2-0.17 Giants20151326.4+1.2737.3-0.17 Chargers20151325.1-1.0338.8+0.13 Texans20141425.9+2.0238.3-0.36 Source: TruMEDIA The NFL teams that hurried up the most, 2013-15 Eagles201536%25.1s+0.1033.0s-0.14 201311512195117 Source: TRUMEDIA Patriots20131525.5-0.4637.5+0.14 Bills20132226.9-0.0835.3-0.06 Browns20141525.0+0.6637.6-0.18 20141039897112 Eagles20132726.0-0.0135.0+0.13 201587979791 Packers20151326.1-0.6438.7+0.12 Jaguars20131626.0-2.2837.7+0.25 Patriots20141626.1+1.3837.2-0.17 OFFENSE RATINGDEFENSE RATING The decline of Kelly’s Eagles TEAMYEARHURRY-UP %TIMEEPA VS. AVGTIMEEPA VS. AVG Ravens20131227.0-0.7437.2-0.00 Texans20151225.4+1.2737.9-0.22 NFL Avg.825.7-0.1239.5+0.01 YEARPASSINGRUSHINGPASSINGRUSHING Now, maybe this season’s ignominy was a hiccup, albeit one largely attributable to the slew of personnel changes Kelly made over the offseason. Certainly, teams this bad at passing tend to bounce back the next season. But it’s rarely enough to get back to league average,1Since 1966, only 27 percent of teams with a passing index from 85 to 90 made it to 100 the next season. and even if the Eagles did, they’d still be left with a mediocre offense overall. That’s hardly the revolution Philly had in mind when it lured Kelly away from Oregon three years ago.But here’s the thing: The “blur” offense actually worked for the Eagles. In terms of using the fewest seconds of clock time per play, the 2015 Eagles were the fastest-paced NFL team since TruMedia began tracking the statistic in 2006. They kept pushing the proportion of their plays that used fewer than 30 real-time seconds between snaps — which I’ll define as “hurry-up” plays for our purposes — and those plays largely continued to be more effective (by per-play EPA) than the league average across all plays.