Why Has Regular Season Become Less Predictive of NFL Playoff Success?
Historically, regular season point differentials predicted the likelihood of reaching the Super Bowl very well. In 2010, Green Bay may have been a “Wild Card” entrant based on their record, but their aggregate differential of +148 put them on top of both conferences. Since 2002, however, regular season differential have become less meaningful. The NY Giants’ regular season aggregate point differential of -6 ranked 10th out of the 16 NFC teams — Arizona’s ranking in 2008. Since the 2003 season, four NFL teams whose point differentials placed them 7th or lower in their conference reached the Super Bowl with another 2 with rankings of 5.
Prior to 2003, the Raiders held the award for the lowest ranked team to reach the Super Bowl in terms of point differential at 6th in the AFC in 1979. From 1978 to 2002, four teams ranked 4th made the Super Bowl. On the flip side, of the 50 teams making the Super Bowl in the time frame, 40 (80 percent) ranked either 1 or 2 in their conference in terms of point differential. Since 2002, 50 percent have ranked 1 or 2. Some of the Super Bowl winners with eye-popping regular season differentials have been ’85 Bears (+258), ’84 49ers (+248), ’99 Rams (+284), and ’96 Packers (+246). The 2003-present era contains the ’07 Patriots with a whopping +315. (Stats from Pro Football Reference).
What’s driving this shift? I have read suggestions of greater parity but that merely restates the question in a different form. The most obvious change has been the playoff format. In 1978, the league introduced Wild Card” playoff format in 1978 and adjusted it in 1990 and 2002, suggesting a possible explanation. This maybe the answer, but it’s not simple even with it. The top two teams have received Wild Card “byes” under under all three WC systems (1978 2 Wild Card play each other, 1990 — three Wild Card teams + 1 of three division winners, 2002 — two Wild Card teams and two of four division winners). The format change and league divisional restructurings have increased the number of teams in the playoffs, providing the opportunity for lower point differential teams to play whether they enter as a WC team or a divisional winner. This may increase the likelihood of a team with a marginal overall regular season but jelling at the end of the season reaching the playoffs and disturbing the outcomes.
Yesterday (1 Feb 2012), Federal Court of Australia Justice Rares ruled in favour of Optus (SingTel), Australia’s second-largest telecommunications company, in a case that upholds their right to offer an internet/cloud-based service allowing subscribers to record free-to-air (FTA) television for private playback on mobile devices. The service allows recorded programs to be played back with as little as a two minute delay from the FTA telecast.
This threatens the viability of the exclusive internet broadcasting rights agreements negotiated by Australian sports. Telsta, Australia’s largest telecommunications company, holds exclusive rights to the live internet streaming of two largest Australian sports, the Australian Football League (AFL) and the National Rugby League (NRL). Optus was opposed by Telstra, the AFL and the NRL in the case, which tested the Australian Copyright Act 1966 (Cth) and the right of individuals to make recordings for private or domestic use.
Such internet rights were evidently a AUD $153 million component of the AFL’s AUD 1.25 billion broadcasting rights deal (2012-2016) negotiated last year. Four of nine regular season AFL games will be broadcast on FTA TV each week, along with the entire AFL Finals Series. The Optus service allows all such FTA games to be time-shifted. Negotiation of the next round of NRL brodcasting rights will formally commence some time this year. NRL executives and analysts had previously talked up the NRL rights as being worth a similar amount to the AFL rights.
The case will almost certainly be appealed by Telstra, the AFL and the NRL to the Full Court of the Federal Court. The Australian Government may also amend the Copyright Act in due course and the sports will lobby the Australian government hard if future judicial decisions uphold the decision of Justice Rares. In 2011, the Australian Government announced a review into media policy in light of the convergence of technology, so this case will doubtless be considered.
From an economic perspective, interesting questions arise as to whether Optus is a free-rider, and the appropriate public policy response to such behaviour (the argument of the sports and Telstra), whether the decision is good for sports consumers (the argument of Optus) as well as whether courts and/or legislatures should construct walls in the form of property rights that individuals have been unwilling, or unable, to construct themselves via contract.
If Optus prevail in the long-run, this decision will certainly put a significant dent into the value of sports internet media rights and exclusive internet rights will be worthless, at least until sports devise new ways to protect such rights. It could feasibly also diminish the value of FTA rights. Whether the benefit to consumers is offset by the utility lost to those who benefit from such revenue flowing to a sport is open to debate.
Around the web, “epic” consistently describes the Nadal/Djokovic 6-hour marathon match in the Australian Open final. The effort, stamina, and skill of these competitors deserves admiration. The final followed on the heels of a nearly 5-hour semifinal between Andy Murray and Djokovic. A 353 minute match? That’s 203 minutes beyond Bill Simmons’ suggested 150-minute limit for almost anything (See “Less is More”). For me, even the Murray-Djokovic match went from epic to ridiculously long, and I didn’t start watching until the third set.
Why does men’s tennis use a best-of-5-set format in grand slam events? The quick, Econ 101 answer is that it must be profit or revenue maximizing. I’m willing to be proven wrong, but my default position is that 300 minute matches don’t maximize viewership in the short run or long run. Sure, there are hardcore fans that may stick in there, but I am the marginal tennis fan. I watch quite a bit of women’s tennis and very little men’s tennis for the very reason that I enjoy a 1-2 hour match such as Clijsters-Azarenka, but at 2 hours, a competitive men’s match has barely started. I once referred to the NBA, where hardly any game matters. In men’s tennis, hardly any set matters in terms of sets played where both players face elimination.
A longer version of the Econ 101 answer might go like this: fans like to see the best players matched up; 3-set matches lead to more upsets and fewer semis and finals between highly-rated players, and therefore, less viewers. Ok, Matchups between highly-rated players in semis and finals raise viewership, but that a partial impact. Fan enjoyment of upsets and uncertainty offsets it to some extent, and match length certainly pulls viewership the opposite direction. Why not play 9-set matches? For that matter, why not make tennis like cricket and extend matches over 5 days? If 5 sets over 6 hours is epic, wouldn’t 25 sets over 30 hours reach “mega epic.”
My guess is that the answer goes beyond Econ 101. Political economy enters the picture. Why does FIFA play the World Cup in South Africa and not only endure but sometimes seem to promote faking. (FIFA: Money, Control, or Both). Yes, the FIFA aristocrats like money (that’s obvious) but they have other interests and internal disputes between, for example, the English Football Association and continental associations. Four different associations organize the tennis grand slam events. All of them care about money, but all are also subject to a variety of internal politics and by somewhat aristocratic organizations, with the exception of the Australian Open.
The Women’s Professional Soccer (WPS), the 5-team American league that just finished out its 3rd season, cancelled its 2012 season today while holding out hope for a return in 2013 or beyond. The WPS represents the second women’s soccer league to fail in the United states in the past decade, following the demise of WUSA in 2003.
WPS’s announcement follows one day after CONCACAF finished up its women’s soccer Olympic qualifying tournament in Vancouver, BC that attracted record crowds that exceeded 25,000 fans for the U.S.-Canada final. Other international women’s soccer matches have also demonstrated an impressively broad level of popularity. Last summer’s Women’s World Cup final, for example, attracted a television rating of 8.6, a figure that exceeded the average rating for games in the 2010 World Series between Texas and San Francisco.
Clearly a market exists for women’s soccer, but attracting fans outside of international events like the Olympics and the World Cup remains an elusive goal. I don’t really have a brilliant economic point to make here other than to note that the saga of women’s soccer in the U.S. serves as a testament to the impressive ability of existing leagues in other sports to regularly attract large crowds, even for those meaningless late season games when both teams are out of the hunt, and without the pageantry and national fervor that women’s soccer has managed to capture on occasion.
Played last night into early morning, the epic almost six-hour, five-set Men’s Singles Final of the 2012 Australian Open between top two seeds Novak Djokovic and Raphael Nadal will be long-remembered by tennis purists for many reasons.
For economists, however, one relatively minor incident may just stand out, if for nothing else as a game-theoretic oddity: Nadal challenged the call on his own serve [watch footage], which had been ruled in by the line judge (and not overruled by the chair umpire), ostensibly because Djokovic rifled a rapid-fire winner in return, at a not-so-critical moment in the second set…and it almost worked to earn a replay of the point for him, as Hawkeye ruled the serve to be in only by a hair’s breath.
Of amusement is how neither of the host broadcaster commentators Lleyton Hewitt nor Jim Courier (both former world number ones) appeared to know whether this was indeed permissible under the current rules of the Challenge System, especially since the former is still a tour professional (he had bowed out in four sets to Djokovic in the fourth round). Fortunately, the chair umpire knew better, although Djokovic protested briefly on the basis that it took Nadal too long to indicate the intention to challenge, however, given the (weak in one’s opinion) standards of enforcement of time limits to indicate, the chair umpire was probably correct in allowing it.
Perhaps it worth wondering whether ITF officials and collaborators foresaw this (and other hitherto unobserved unusual scenarios) when drafting the specifics of the system, and indeed whether in light of this occurrence they will decide to ‘close the loophole’, though I’d doubt it, since there seems to be nothing fundamentally wrong with this type of challenge in principle.
Since less than half of all challenges currently are upheld, if there is a reform worth considering, it would be the limiting of challenges to two (incorrect) per set, as originally in the 2007 Australian Open, to reduce the number of more frivolous and tactical challenges, which appear to have been rising in incidence more recently as players have become more used to - and more able to exploit - the system.
The new NFL overtime rule, see page 111 (introduced last season for playoff games only) finally became binding in the current playoff series two weeks ago, and then for a second time yesterday. Of most interest to economists about the rule change is its potential for teams to alter their overtime strategies in response to the element of the rule that both teams are now guaranteed a direct opportunity to score.
While it’s far too early to make firm inferences about this, it’s worth recounting that the primary intention of the rule was for the result of the coin toss determining which team gets to receive the kick-off (and consequently first possession) to have less power in determining the match winner (previously almost 60% of the nearly 500 overtime games since 1974). At the margin, we might also expect touchdowns to be the winning method of scoring slightly more often than previously, since it is now the only way that the team with first possession can effectively kill the game without the opposition getting the chance to equalize.
On the basis of these first two observations, it is interesting to note that the qualitative outcomes were very different - while winning the toss allowed Denver and Tim Tebow to end the contest on the first play via a touchdown, both NY Giants and San Fransisco each failed to score on their first possession, triggering reversion to sudden death as previously, which was won eventually by the former. [Disclaimer: I have only recently followed the sport, so feel free to critique this via comments.]
While this (admittedly premature) anecdotal evidence suggests that the coin toss is more influential than before, the reality is about as mixed as could have been expected. Someday – perhaps in a few decades or even within the decade should the rule be extended to the regular season, one of our esteemed TSE colleagues will no doubt crunch the numbers on this rule change when there is a sufficient (presuming due restraint exercised) sample size.
Brian Burke (admin of Advanced NFL Stats) posted a very insightful piece on Deadspin (and also Slate) regarding the slow evolution in the punt/go-for-it decision on 4th down. Readers might recall the outcry two season back across media-dom from Bill Simmons, Teddy Bruschi and others when the Patriots shunned punting from their own 30 late in a game against the Colts. (Belichick Understands Probability). Sacrilege. Insulting to the defense, *&%^& statheads, yada, yada, yada. Burke offers a very persuasive thought experiment:
There are many doubters when it comes to four-down football. If you’re in that camp, indulge me in a quick thought experiment. Let’s imagine a football world where the punt and field goal had never been invented. (Sorry, Ray Guy and Jan Stenerud.) In this universe, there would be no second-guessing: Teams would go for it on every fourth down. Then one day, some smart guy invents the punt and approaches a head coach with his new idea. “Hey coach,” he’d say, “instead of trying for a first down every time, let’s voluntarily give the ball to the other team.” Our coach would be incredulous at this suggestion. “You want me to give up 25 percent of our precious downs for just 35 yards of field position? Do you have any idea how difficult it would be for us to score?” And the coach would be right.
His experiment amplifies the critical cost of punting — voluntarily giving up valuable possession of the ball. Yes, going-for-it on 4th down and failing hurts, but so does handing over the football to a high-powered offense. The story is couched so that the typically explicit and implicit costs switch places, thereby highlighting the silliness of coaches who treat the 35 yard change of in expected field position as the only real cost while discounting the implicit cost of the lost possession. Moreover, this discounting of the implicit cost of possession seems closely tied to what is generally accepted practice that made more sense in an era of defensive dominance and low scoring. The 1974 Steelers gaining 35 yards in field position for their defense against almost anybody is much different than a than handing over possession to Brady, Manning, Brees, or Rodgers with time on the clock where they have already racked up 400 yards of offense or more.
A couple of articles, both related to scheduling, caught my attention recently. First, Bob Harig at ESPN ponders “Why the No Shows for the PGA Tour Opener?”. Eleven of the players eligible for the elite, winners-only kickoff to the PGA season skipped the event, including 3 of the 4 major winners from last year, Rory McIlroy, Charles Schwartzel, and Darren Clarke, along with big names like Phil Mickelson. That’s over 25% no shows for a no-cut, guaranteed payoff event (last place paid out $64,000). As Harig says,
No sport starts its season with as little fanfare, as little buzz, as the PGA Tour.
He identifies several contributors: injury, short offseason, many other opportunities to make money, competition with NFL playoffs. I want to single out the long season/number of events for scrutiny. The choice of more (or fewer) events presents a very difficult tradeoff. A long season provides many hours of TV exposure, and dedicated intermediaries like NBC’s Golf Channel like this. (Of course, the long schedule preceded the Golf Channel’s tour coverage). The down side is that it means more opportunities for players to cherry pick their favored events, with fewer times where the best play against the best which enhances the quality of watching for consumers. In addition, a longer season pits the tour on weekends in January against big NFL games. The tour made a shift in the Kapalua event this year to a Monday finish to try to mitigate this impact, but this loses the bigger Sunday audience. As I’ve written before on the NBA, leagues and or their TV-associates seem very reluctant to experiment with shorter seasons.
The second scheduling article appeared in today’s Wall Street Journal , where Jared Diamond details the small (or non-existent) number of true road games played by the big dogs among NCAA basketball teams (the same issue is also prevalent among football teams.) Kentucky, North Carolina, and Duke have played no non-conference road games, Kansas one. Louisville had also played no road games until its Dec. 31 rivalry game against Kentucky. Last night the Cards lost by 30 at Providence.
The financial incentives are obvious. The issue that I would raise is the lack of adjustment in things like RPI or polls for such home-cooking. The NCAA tournament selection committee often hides behind “quality wins” and records against top x (50, 70, …) teams to justify their selections and seedings. However, these matchups in non-conference games are heavily biased toward teams playing at home all of the time. That’s why games from prior year tournaments on neutral (or somewhat neutral) courts are really more informational than the selection committee seems to utilize. Of course, the same financial incentives at work in why the big dogs play at home so much are also at work, maybe more subtly, in the politics behind selection committee decisions.
Orange Bowl: 70-33; Rose Bowl: 45-38; Fiesta Bowl 41-38. All of these scores and, of course, the insane 67-56 scoring orgy in the Alamo Bowl 67-56 between Baylor and Washington make me contemplate Guns N’ Roses eternal question: where do we go from here?
In 2008, “Spreading the Wealth” mulled the points explosion with Longhorn coach Mack Brown’s remark
“I’m sure [defense will catch up], it always has. But we’re not sure how yet.”
I long subscribed to Brown’s default position — offense and defensive advantage ebbs and flows — but I’m beginning to doubt myself. The points-monster just keeps growing. From 1971-82, Rose bowl winners topped 30 points 2 times. From 2001-12, 9 winners scored more than 30. The total score stuns more than the winner. Over the past two seasons, teams combined for 378 points (63 points per game) in the Orange, Rose, and Sugar Bowls. In my junior and senior years of high school, 1977 and 78, they totaled 222 or 37 points per game. (Data available at Sports-reference.com’s Bowl Index).
These shifts appear in pro and high school football also. Over the last two years, the semi-final and final games in the top division of Texas high school playoffs (2011 and prior years) have put up 63 points per game versus 21 in 77 and 78. In the NFL, two teams with defenses ranked near the league bottom, New England and Green Bay, enter the playoffs as number 1 seeds.
In the NFL, the evolution of rules and their enforcement has inexorably led to more effective passing starting with the elimination of “bump and run” in the late 70s to crackdown on defensive holding in the early 2000s to the ever-greater protections extended to QBs. At all levels, strategic developments on offense have effectively expanded the square footage that defenses must defend on each play. While all of these developments have contributed, their long run impacts have been limited by labor supply — young QBs developed in high school and college as effective throwers. After a generation, there are now scads of HS and college quarterbacks running relatively sophisticated passing offenses, developing not only their arms but their minds and eyes. Some of these shifts may be countered by defensive strategy and development of defensive players, but I wonder if some of the rule and strategy changes permanently shift the relative balance between offense and defense.
From a league/consumer standpoint, are these developments “bad”? It’s easy for some middle-aged or older like myself to pine about some bygone golden era and bash the current state of play. What is the optimal amount of scoring? I don’t have reliable data on this question, but I do suspect that fans tire of no scoring (unless they like soccer) or pinball-like scoring. I disliked the heavily run-oriented styles of the 1970s, but I also enjoy seeing good defensive plays, of which there are few are none in a 67-56 game.
While these trends may end, it’s hard to see that happening very soon. In high school and college, teams with strong defenses and much better overall ability than an opponent (say LSU or Alabama) can shut down offenses relatively well. In the pros, with more equalized talent, the tipping point seems determined only by whether a team has a passing proficient QB or not. There’s no “stopping” Brady, Rodgers, or Brees — only hopes of intercepting a tipped pass or pushing them back with a penalty. Even the supposed defensive or run-oriented teams like Pittsburgh often resort to second half passing shootouts against stronger opponents with efficient offenses.
Further to an earlier post one year ago to the day, the same result occurred upon conclusion of the 2011 NFL season as in EVERY previous season since the re-alignment to 32 teams in 2002 – that the NFL is even more competitively balanced when the standings are adjusted for strength of schedule (than on the basis of raw standings) .
Even though this season was noticeably less balanced than (the recent) average without adjustment, the difference from the adjustment was even more pronounced than any other season during this period (with the exception of 2003), making adjusted competitive balance about average over the same period.
On this occasion, the actual-to-idealized standard deviation ratios are 1.611 (unadjusted) and 1.462 (adjusted); the Gini coefficients are 0.292 (unadjusted) and 0.260 (adjusted); and the Herfindahl indexes of CB are 1.162 (unadjusted) and 1.134 (adjusted).
Ultimately, this result reinforces further the need to account for strength of schedule in producing standard competitive balance metrics for various empirical studies!
