Preseason 'Dress Rehearsals' Can Be More Misleading Than Enlightening

The 2001 Patriots lost their “dress rehearsal,” their third preseason game, 20-3 to the Buccaneers. The starting offense recorded just three first downs. The first-team defense allowed two touchdowns. Tom Brady played the final series of the game, completing one of three passes for four yards. It was a dull, listless performance.

Winslow Townson/Associated Press

Those Patriots finished 5-11 the previous season. Brady was an unheralded third-stringer. No one at the time expected them to come out in the dress rehearsal and play like the next great dynasty or a Mount Rushmore-caliber Hall of Famer.

But in retrospect, they should have shown us something back in 2001. If the third preseason game really is the bellwether of a team’s fortunes, shouldn’t the Patriots at least have looked pretty good? Shouldn’t Brady have led a late drive or something?

We will spend a lot of time over the next week or two sifting through this year’s third round of preseason games carefully, searching for signs and portents about who will rise or fall this year. But if we couldn’t tell that the Patriots were about to become THE PATRIOTS in their third preseason game of 2001, it doesn’t seem like we have much of a chance figuring out whether some team will go from 6-10 to 10-6 based on this weekend’s results.

But maybe the 2001 Patriots were an isolated case. Let’s take a look back through preseason history to determine if the third preseason game can really help us predict anything about a team or a player’s fortunes.

Super Bowl Tuneup

Let’s start with a simple question: How do eventual Super Bowl winners look in their dress rehearsals? The table below shows the breakdown. Notice that we will be focusing on halftime scores, for obvious reasons. I also threw the 2007 Patriots into the data, because if the third preseason game can help show that some team is about to go 17-0, we’d all like to know about it.

Season	Team	Game Result	Halftime Score
2001	Patriots	Lost 20-3	Trailed 7-3
2002	Buccaneers	Lost 40-10	Trailed 16-7
2003	Patriots	Won 24-12	Led 17-3
2004	Patriots	Lost 20-17	Led 14-10
2005	Steelers	Lost 17-10	Tied 10-10
2006	Colts	Won 27-14	Led 17-0
2007	Giants	Lost 20-12	Led 12-6
2007	(Patriots)	Won 24-7	Led 10-7
2008	Steelers	Won 12-10	Trailed 7-3
2009	Saints	Won 45-7	Led 31-0
2010	Packers	Won 59-24	Led 29-17
2011	Giants	Lost 17-3	Trailed 7-3
2012	Ravens	Won 48-17	Led 13-3
2013	Seahawks	Won 17-10	Tied 3-3
2014	Patriots	Won 30-7	Led 13-0
2015	Broncos	Won 19-12	Led 8-3

Eventual Super Bowl winners are 10-6 overall in third preseason games and 10-4-2 at halftime. So it’s better to win the dress rehearsal than not. It’s a pretty minor trend, one with no real predictive value.

It’s worth noting that eight of the last nine Super Bowl winners also won the rehearsal. Also, no eventual Super Bowl winner looked truly awful in the rehearsal, with the possible exception the 2002 Buccaneers. Brad Johnson threw two first-quarter interceptions for the Buccaneers in that game; he would throw six in the entire regular season.

Some of those wins by eventual champions were pretty convincing, like the Saints obliterating the Raiders in 2009 or the Packers pulling away from the Colts in 2010. That Saints blowout looks significant in retrospect: They were coming off a series of non-winning seasons, and a 31-0 halftime lead was a clue that they may have taken a major step forward.

On the flip side, the Giants lost a pair of “Snoopy Bowls” to the Jets. Granted, Big Blue had a habit of looking pretty bad until the fourth quarter of the Super Bowl, but the listless losses gave us no hint that those Giants teams would be different.

Also, our data is full of Patriots teams, and no one has needed a preseason game to determine whether the Patriots would be good since 2001, when it didn’t help. The 2006 Colts looked great in their rehearsal, but they went 14-2 the previous year, so we weren't reading preseason tea leaves to determine if they were any good. You get the idea. If the rehearsal can only tell us that already great teams will (sometimes) look great in the preseason, then it isn't telling us much, particularly when surprise winners like the Giants, Buccaneers and 2001 Patriots don't do anything surprising.

August All-Pros

Perhaps we should come at this from the opposite direction. Super Bowl teams may just look kinda OK in the dress rehearsal, but what about teams that look like soon-to-be world champions in the third preseason game? Does a blowout win presage a tremendous season? Does a blowout loss predict a terrible year?

I combed through the last decade seeking third preseason games where the halftime score differential was 21 points. These are rehearsals in which one team looked overwhelmingly awesome and the other sent its fanbase to StubHub to sell its tickets. The following chart shows how both the winners and losers of those games fared in the regular season:

Year	Winner	Halftime Score	Winner's Record	Loser	Loser's Record
2015	Eagles	39-14	7-9	Packers	10-6
2014	Buccaneers	24-0	2-14	Bills	9-7
2014	Seahawks	31-0	12-4	Bears	5-11
2013	Bears	27-3	8-8	Raiders	4-12
2009	Saints	31-0	13-3	Raiders	5-11
2008	Eagles	24-3	9-6-1	Patriots	11-5
2008	Panthers	34-0	12-4	Redskins	8-8
2006	Raiders	21-0	2-14	Lions	3-13
2006	Bengals	34-7	8-8	Packers	8-8

The Super Bowl-bound 2009 Saints make our list as big winners. So do the 2014 Seahawks, who would take the Patriots to the 1-yard line in the Super Bowl. But so do the 2014 Buccaneers and 2006 Raiders, destined to win just two games each.

The 2008 Patriots won 11 games without Brady despite a dud of a rehearsal. Granted, Matt Cassel started that “rehearsal,” but he ended up starting the whole season. The preseason warm-up gave no indication that the Patriots would still be pretty darn good.

Mike Roemer/Associated Press

The Eagles obliterated the Packers last year in the most recent example of a misleading preseason blowout. That game had me fooled: The Eagles were coming off a pair of winning seasons, after all, so 39 first-half points by a redesigned offense looked like a sure sign that they were Super Bowl contenders. It was really a sign that teams don’t game-plan for the no-huddle in August.

Teams that blow out opponents in the dress rehearsal go on to win an average of 8.2 regular-season games. (Note: The Eagles' tie in 2008 counts as half a win.) The blown-out teams win an average of 7.o games. That’s a tiny nudge off .500 in each direction based on a small 10-game sample. In other words, it’s statistical bean dip.

Teams Fall Apart

Maybe we are asking too much of 30-odd minutes of August football if we expect it to predict a particular record or result. We may have a better chance of spotting some major franchise-shaking trend. For example, if a perennial contender is about to have a catastrophic season, it should theoretically be evident in the dress rehearsal game. If the rehearsal can really tell us anything, that is.

The next table shows how longtime playoff teams performed in the dress rehearsals for their “disastrous collapse” seasons. I wasn’t interested in up-and-down teams who went 11-5 once and fell back to earth or contenders that rolled to a stop slowly with a year or two of .500 or near-.500 seasons (like the Saints are doing). These teams were in the playoffs or had 10-plus-win seasons for three or more seasons, then fell off a cliff. Here’s how they fared in the first halves of their dress rehearsals:

Year	Team	Halftime Score	Previous Record	Eventual Record
2000	Jaguars	Led 10-6	14-2	7-9
2001	Vikings	Led 14-10	11-5	5-11
2003	Raiders	Trailed 14-3	11-5	4-12
2004	Dolphins	Tied 3-3	10-6	4-12
2004	Titans	Led 14-13	12-4	5-11
2005	Eagles	Led 27-3	13-3	6-10
2005	Packers	Trailed 20-3	10-6	4-12
2008	Seahawks	Trailed 10-7	10-6	4-12
2011	Colts	Led 14-10	10-6	2-14
2013	Falcons	Trailed 13-6	13-3	4-12
2013	Texans	Trailed 17-16	12-4	2-14
2015	Ravens	Trailed 14-13	10-6	5-11

Curtis Painter completed 11 of 21 passes for 171 yards and two touchdowns, including a 57-yarder to Reggie Wayne, against the Packers in the 2011 Colts dress rehearsal. Colts fans must have walked away from that game feeling that the team would at least be competitive and exciting without Peyton Manning, not a 2-14 calamity.

Joe Robbins/Getty Images

Donovan McNabb and Terrell Owens put aside their differences to connect five times for 131 yards and a 64-yard touchdown in their preseason dress rehearsal in 2005. There was no evidence that they would spend the season at war or that the Eagles were about to go from the Super Bowl to a 6-10 finish.

Matt Schaub threw for 213 yards and no interceptions in the Texans’ 2013 dress rehearsal. He would turn into a pick-six machine just a few weeks later, and the Texans would go from a playoff team to a 2-14 comedy troupe.

Some teams did show evidence of collapse in the third preseason game. The 2003 Raiders let the Vikings march all over them; the halftime score would have been 20-3 if not for a pair of missed field goals. The Raiders were about to go into a swoon that they still haven’t climbed completely out of.

The second half of a dress rehearsal may reveal some problems with the roster depth. Last year’s Ravens trailed 14-13 at half but 31-13 by the end of the game. It may have been a portent that the team would go off the rails as soon as some starters were hurt.

Overall, however, this is another case of the dress rehearsal coming up short as a predictive tool. If the results can’t tell us that Painter is overmatched, McNabb and Owens are at each other’s throats or Schaub has lost his stuff, what can it really tell us this weekend about 2016? 

Sleeping Dynasties Awake

The third preseason game clearly doesn’t tell us much about teams on the verge of collapse. But perhaps, 2001 Patriots aside, it can provide clues that a powerhouse is about to rise. The following table consists of teams that were coming off losing seasons but were about to embark on at least three straight winning seasons or playoff years. In other words, these are franchises in the midst of significant, long-term leaps forward. Let’s see if their rehearsals offered any clues about what was happening.

Year	Team	Halftime Score	Previous Record	Eventual Record
2000	Eagles	Led 21-6	5-11	11-5
2001	Patriots	Trailed 7-3	5-11	11-5
2004	Chargers	Trailed 23-0	4-12	12-4
2005	Giants	Led 9-7	6-10	11-5
2008	Ravens	Trailed 17-3	5-11	11-5
2008	Falcons	Tied 3-3	4-12	11-5
2009	Packers	Led 38-10	6-10	11-5
2011	Bengals	Led 24-7	4-12	9-7
2011	Broncos	Led 10-3	4-12	8-8
2011	49ers	Trailed 24-7	6-10	13-3
2012	Colts	Trailed 14-7	2-14	11-5
2012	Seahawks	Led 23-7	7-9	11-5
2013	Panthers	Led 24-7	7-9	12-4
2013	Cardinals	Led 14-0	5-11	10-6

Many of the teams on this list rose to greatness in part because they found a quarterback. Often, that quarterback made his presence known in the rehearsal.

Russell Wilson threw two touchdown passes in his rehearsal; that he even started the game as a third-round pick, then excelled, was a hint that the Seahawks’ fortunes were about to change. McNabb was a second-year quarterback when he threw three touchdowns in the 2000 rehearsal. Aaron Rodgers was coming off a 6-10 season and still emerging from Brett Favre’s shadow when he threw three touchdowns in the Packers' third preseason game of 2009.

Peter G. Aiken/Getty Images

If only the storyline was always so neat. Drew Brees and Philip Rivers combined for three interceptions against the Seahawks in the Chargers’ third preseason game of 2004; that game provided a pretty false sense of how the next decade of NFL history would turn out. Injuries to other quarterbacks forced Joe Flacco to start the entire 2008 dress rehearsal for the Ravens. He threw a late touchdown, but the Ravens went 1-of-13 on third-down conversions. Matt Ryan also looked very ordinary in his Falcons rehearsal until the second half.

Colin Kaepernick was 6-of-16 for 52 yards and an interception off the bench as Jim Harbaugh’s 49ers got hammered, showing no signs that they would go 13-3 after many years of dormancy. Eli Manning didn’t even play in the third preseason game in 2005. That 24-7 halftime lead for the Panthers in 2013 was not a credit to Cam Newton’s development; the Panthers scored four touchdowns on defense and punt returns.

Dynasties on the rise are 8-5-1 at halftime of their third preseason games. They outscore their opponents 13-10. That’s a morsel of a trend. If Jared Goff, Jameis Winston or Marcus Mariota throws three touchdowns this weekend, it’s worth taking note. Just remember that Curtis Painter looked great five years ago this weekend.

Potential to Flash Potential

As noted in the last section, Russell Wilson had an outstanding dress rehearsal game. Perhaps we need to focus on rookie quarterbacks. The third preseason game is often a top prospect’s first chance to play several series with the starters. Maybe it provides a chance to identify future franchise quarterbacks or epic busts. If nothing else, it might be fun to collect the data on all of the first-round picks so we can compare Jared Goff, Paxton Lynch or (swoon) Dak Prescott to the likes of Newton and Andrew Luck on Monday morning.

Here are the statistical lines from the dress rehearsal games for each rookie first-round quarterback drafted since 2009. We could go back further, but once you see the table, you will see that enough is enough when it comes to statistically insignificant porridge:

Year	Quarterback	Att	Comp	Yards	TD	INT
2015	Jameis Winston	15	6	90	0	0
2015	Marcus Mariota	11	7	99	0	0
2014	Blake Bortles	17	10	158	1	0
2014	Johnny Manziel	15	10	85	0	0
2014	Teddy Bridgewater	7	4	40	2	0
2013	EJ Manuel	DNP	DNP	DNP	DNP	DNP
2012	Andrew Luck	23	14	151	1	0
2012	Robert Griffin III	17	11	74	1	0
2012	Ryan Tannehill	27	11	112	0	1
2012	Brandon Weeden	20	9	117	0	0
2011	Cam Newton	19	6	75	0	0
2011	Jake Locker	4	2	13	0	0
2011	Blaine Gabbert	13	6	52	1	1
2011	Christian Ponder	8	3	60	0	0
2010	Sam Bradford	22	15	189	2	0
2010	Tim Tebow	10	5	72	1	1
2009	Matthew Stafford	19	13	160	0	1
2009	Mark Sanchez	20	13	149	1	0
2009	Josh Freeman	16	6	50	0	0

There’s a saying in analytics circles: Stare at the tortilla long enough, and you will see the face of Elvis. Actually, I am the only person who says that, but the point is well-known among statisticians: Search random data for trends, and you are likely to invent trends that aren’t really there.

It appears to be a bad sign if a rookie quarterback barely plays in the rehearsal, based on the poor performances by Jake Locker and Christian Ponder. But 2011 results must be taken with a grain of salt: The lockout prevented rookies from participating in minicamps or OTAs, so many were not ready to play by late August.

In a very broad sense, Brandon Weeden’s stats look pretty bad, while Blake Bortles and Luck look pretty good. But Sam Bradford looks like the Hall of Famer of the bunch. And Johnny Manziel, Cam Newton, Robert Griffin and Marcus Mariota all look about the same. Good luck squeezing a conclusion out of that.

There is no grand trend in this data. Future Pro Bowlers and rookie phenoms rarely look spectacular. Busts aren’t exposed. And for every later-round Wilson who kicks down the doors of the depth chart and wows the crowd in the third preseason game, there’s a Brady mopping up the fourth quarter.

One last thought about the table above: In his rookie rehearsal game, not only did Tim Tebow look like a typical quarterback, he did not even run the ball once.

That’s what you call misleading data.

The Show Must Go On

In summary, the preseason dress rehearsal is a game in which the Patriots can look mediocre, Drew Brees and Philip Rivers can look like a pair of stiffs, Donovan McNabb and Terrell Owens can look like besties, a 2-14 team can blow out an opponent and Tim Tebow can look like a pocket passer.

It makes you want to spend this whole weekend at the pool instead of worrying about how the home team looks in the dress rehearsal, doesn’t it? Sure, a Russell Wilson or 2009 Saints could emerge, but how would anyone separate their signals from all of the noise?

Many fans overreact to the third preseason game. Some fans overreact to the over-reactors by screaming about how meaningless the results are. Some of us are paid to jump to premature conclusions, and we’ll tell you on Monday just how significant everything that happened this weekend was.

None of it is that significant. But it is not meaningless either. There is a gentle tilt in most of the data. Teams destined to be great do fare slightly better in the third preseason game than teams destined to be terrible. It’s not a trend you can take to Vegas. It’s no reason to book any February flights to Houston based on this week’s results. But it’s enough to make the games worth watching and examining.

At least we can all agree that an injury-free dress rehearsal is a good thing.

And if some sixth-round pick throws one four-yard pass in the fourth quarter of a lousy game, you may want to scribble down his name. You never know.