At one point last year, the Arizona Cardinals were arguably the NFL’s most unexpected success story, rising all the way to No. 2 in our NFL Elo rankings before injuries to starting quarterback Carson Palmer (and his backup) tanked their season. But that Cardinals team has nothing on this year’s version, which — according to Elo, at least — is enjoying one of the most surprising starts of any team in the Super Bowl era.1
Going into this season, Elo thought the Cardinals were an OK team, ranking them 10th in the league overall but also projecting them to finish .500 and barely break even on point differential. (Because of this, our Elo simulations assigned the Cardinals a mere 36 percent chance of returning to the playoffs.) Based on the team’s preseason Elo rating, we’d expect the Cardinals to have won 2.9 times in the first five games of the season, with a cumulative point differential of +13.
Instead, Arizona is 4-1 with an NFL-best +100 point differential, making it only the 12th NFL team since 1966 to break triple-digits in scoring margin through the season’s first five games. The Cardinals currently rank sixth in Elo. And that 87-point disparity between the Cardinals’ actual and expected point differentials? It makes them the 12th-most-surprising team, through five games, of the Super Bowl era:
| ELO RATING | POINT DIFFERENTIAL | ||||||
|---|---|---|---|---|---|---|---|
| RANK | YEAR | TEAM | PRESEASON | AFTER 5 GAMES | PROJ. | ACTUAL | DIFF |
| 1 | 1999 | St. Louis | 1408 | 1576 | -23 | +123 | +146 |
| 2 | 1966 | Dallas | 1500 | 1578 | +16 | +138 | +122 |
| 3 | 1968 | Dallas | 1545 | 1641 | +16 | +136 | +120 |
| 4 | 2006 | Chicago | 1516 | 1630 | +7 | +120 | +113 |
| 5 | 1974 | New England | 1420 | 1554 | -18 | +87 | +105 |
| 6 | 1969 | Minnesota | 1523 | 1637 | -7 | +97 | +104 |
| 7 | 2007 | New England | 1633 | 1723 | +25 | +117 | +92 |
| 8 | 2013 | Kansas City | 1354 | 1496 | -22 | +70 | +92 |
| 9 | 2009 | New Orleans | 1517 | 1651 | +8 | +99 | +91 |
| 10 | 1991 | Washington | 1581 | 1661 | +22 | +111 | +89 |
| 11 | 1997 | N.Y. Jets | 1332 | 1429 | -41 | +47 | +88 |
| 12 | 2015 | Arizona | 1541 | 1612 | +13 | +100 | +87 |
| 13 | 2007 | Dallas | 1491 | 1604 | -7 | +80 | +87 |
| 14 | 2003 | Indianapolis | 1506 | 1620 | -9 | +76 | +85 |
| 15 | 2011 | Detroit | 1451 | 1560 | -15 | +70 | +85 |
| 16 | 1996 | Green Bay | 1609 | 1661 | +19 | +101 | +82 |
| 17 | 1995 | Oakland | 1521 | 1595 | +4 | +84 | +80 |
| 18 | 2001 | Green Bay | 1556 | 1628 | +4 | +84 | +80 |
| 19 | 1980 | Detroit | 1381 | 1488 | -19 | +61 | +80 |
| 20 | 1970 | Minnesota | 1641 | 1694 | +26 | +105 | +79 |
Perhaps Elo, being the simple system it is, was underestimating the Cardinals by including its non-Palmer games from the end of last season. They were a markedly better team a year ago with Palmer under center, and Palmer has been shredding opposing defenses so far this season. Then again, relative to other predictive systems, Elo wasn’t an outlier in its opinion of Arizona; power ratings implied from preseason Vegas over/unders ranked the Cards seventh, while ESPN’s preseason Football Power Index ratings put the team at 13th.
Either way, there’s good news and bad news for the Cardinals looking forward. The good: For predicting point differential over the remainder of a season, a team’s actual point differential is 1.8 times as important as its Elo-predicted point differential before the season — so Arizona figures to play more like the powerhouse they’ve been than the mediocrity they were supposed to be. The bad: That means some regression is also due. Our rest-of-season Elo simulations say the Cardinals will outscore opponents by an average of 2.5 points per game, quite a bit less than the 20 points per game they’ve won by, on average, so far this year.
In other words, the Cardinals probably aren’t the 2007 New England Patriots. But they also aren’t the middling team they looked like before the season. The truth, as usual, lies somewhere in between.