The firing and replacement of NFL head coaches has ended with seven slots filled. Three of them (Jim Caldwell, Lovie Smith, Ken Whisenhunt) led teams in the past and four (Jay Gruden, Mike Pettine, Bill O’Brien, Mike Zimmer) move up from assistant coaching-coordinator positions with O’Brien adding the last two seasons as head coach at the collegiate level for Penn State.
While most fans, no doubt, have feelings about whether their team’s hiring shines or stinks, the truth is, it isn’t much different from draws from a lottery hopper or beauty contest. Data analytics have invaded most every aspect of sports, but this is one area where the numbers don’t reveal much other than noise. In the world data miners, I don’t doubt that a persistent digger out there may discover some systematic signals of success, but they are very subtle.
Sure, if a college team could hire Alabama’s Nick Saban, their chances of success are not random. At the pro ranks, teams seldom get a shot at someone with a long, proven track record. Kansas City’s hiring of Andy Reed in 2013 is one of the few examples. The choice between a coach with a modest prior winning percentage and a coordinator with no head coaching experience differs little from a coin flip. The same holds when pulling a successful coach up from the college ranks. For nearly every Jim Harbaugh there is a Steve Spurrier or Greg Schiano.
Prior NFL head coaching experience would seem to be useful indicator. As a friend of mine suggested, a head coach must wear many hats, pointing toward the existence of a learning curve. I suspect that it true, but no such pattern jumps out of the data. If it’s in there, it’s subject to important caveats and swamped by other influences.
Coaching connections seem to be a common indicator used by general managers. If someone coached under Bill Walsh in the 1980s or Bill Belichick in the 2000s, they were more likely to land jobs. In all sports, general managers and owners lean on coaching pedigree.. In the NFL data, at least, they don’t help predict success very obviously. Great coaches spawn about as many dregs as stars.
About the most that can be said is that teams might avoid extremely bad choices by selecting someone like Lovie Smith, who had a solid record at Chicago, but high-end success is not guaranteed. Coaches such as Jimmy Johnson slid backwards in their second stint. On the other side, some coaches with modest to decent records in their first attempts, such as Belichick or Pete Carroll, performed much better in their second chance. Maybe that’s learning or maybe it’s the luck of a hall-of-fame quaterback like Tom Brady emerging out of a late round, shot-in-the-dark draft pick. The bottom line for fans is that whether they are happy or upset by your team’s coaching choice, only time will tell.
In this paper (now forthcoming, JSE: doi: 10.1177/1527002512471538), it was shown that for every single year after the expansion to 32 teams in 2002 (until 2011), the NFL was even more competitively balanced when the strength of schedule was accounted for, without exception, using four common CB measures. Previous The Sports Economist posts on this are here and here.
Since the 2013 regular season has just been completed, we crunched the numbers on the two most recent seasons. The streak remains unbroken, once again demonstrating the importance of adjusting CB measures for unbalanced schedules.
Standard Deviation Ratio: 1.5245 (unadjusted); 1.4645 (adjusted)
Herfindahl Index of CB: 1.1453 (unadjusted); 1.1340 (adjusted)
Concentration (12) Ratio: 1.4010 (unadjusted); 1.3889 (adjusted)
Gini Coefficient: 0.2776 (unadjusted); 0.2647 (adjusted)
Standard Deviation Ratio: 1.5271 (unadjusted); 1.4396 (adjusted)
Herfindahl Index of CB: 1.1458 (unadjusted); 1.1295 (adjusted)
Concentration (12) Ratio: 1.4063 (unadjusted); 1.3904 (adjusted)
Gini Coefficient: 0.2776 (unadjusted); 0.2590 (adjusted)
With cricket’s latest ‘Ashes‘ series having been decided more than a fortnight ago, much of the remaining interest in the final Test, just underway, centres purely on whether Australia can complete a 5-0 Ashes whitewash for only the third time in history (following 1920/21 and 2006/07). Most ‘key performance indicators’ for Sydney are pointing in that direction – the newly-rediscovered ferocity of the Australian pace attack, not to mention the unexpected feebleness of England’s batting top-order (and middle-order, for that matter). England have also selected three debutants.
Moreover, there is an additional factor that, given recent history, points fairly and squarely in Australia’s favour – merely that they are the home side. The calendar year of 2013 was a stellar one for home teams in Tests. Specifically, of the 44 Tests played last year, a remarkable 30 were won by the home team, 10 were drawn and only 3 won by the away team (Pakistan looks likely to make that 31 against Sri Lanka in the Test that started on New Year’s Eve).
The surprising element of this occurrence, according to some pundits, was that it came after a period, from 2010-2012, in which away teams performed quite admirably against the tide of home-ground advantage. At least on the basis of raw numbers – in these years, a combined 124 Tests resulted in 48 home wins, with the number away wins almost at parity (43). This apparent ‘trend’ towards away teams did not go unnoticed by sports journalists and other non-academic writers. For example, Gideon Haigh remarked to this effect (“The quiet revolution: home ground advantage begins to fade away”, The Australian, 20/12/2012).
However, as economists know all too well, looking at just the raw figures is too parsimonious an analysis for making claims that the nature of home-ground advantage – a phenomenon so well researched, understood and entrenched in sporting culture – has diminished so fundamentally and suddenly. What needs to be understood about the sample of Tests in those years were that they were correlated with factors that skew the chances of victory in favour of the home team to begin with, most obviously on the basis of relative strength of both teams.
For example, easily the best team of 2012, South Africa, played 10 Tests that year. All but one of them were played away from home, with an unbeaten record (4 wins, 5 draws) befitting a World number one. This is a nice example of what in economics (and some other scientific disciplines) is called a ‘selection bias’ – such biases have to be accounted for, since it is not difficult to imagine how the figures would have more-highly favoured home teams at an aggregate level if the Proteas had instead been scheduled to play 9 Tests at home. Another example (this time for 2011) is that minnows Bangladesh and Zimbabwe played a combined total of 8 Tests, 7 of which were on home soil, again skewing the overall record in favour of away teams. With such a small sample of teams and Tests, these selection biases are important and should not be ignored.
Likewise, the stunning reversal back towards home teams in 2013 has to be taken with caution – (the again rampant) South Africa played 7 of its 9 Tests at home, winning 6 and drawing the other. Second-ranked India also played the majority (6 of 8) of its Tests at home, completing a perfect record, not to mention third-ranked England’s impressive Northern summer record (5 wins, 2 draws), prior to their almost inexplicable slide in the current series.
Most Australian cricket supporters will hope that the aberration of 2013 does not continue past this week – their next series is away to South Africa. Otherwise, that tour could prove to be a sobering experience following the current euphoria. Nevertheless, the influence of home-ground advantage in Test cricket does not appear to be under any immediate threat. To this end, if one was the betting type, I would not be shy in punting on a home victory at the SCG.
The Canadian Press has chosen the new television deal between Rogers Communications and the National Hockey League (NHL) as the 2013 business story of the year (Toronto mayor Rob Ford is not a business story). Few in the United States would care, but in Canada, hockey is a religion so the deal has received much press. Rogers is also an owner of MLB’s Toronto Blue Jays and its stadium, Rogers Centre, as well as a large network of local and satellite television stations, radio stations and print and on-line magazines and newspapers. Besides Canada’s obsession with the “national sport”, the deal itself is staggering. Rogers agreed to pay the NHL $5.2 billion for the exclusive television rights to the Canadian market for 12 years. To put this in perspective, NBC signed a 10-year TV rights deal with the NHL in 2011 for about $2 billion. Of course the new deal pales in comparison to the ESPN-NFL 10-year TV deal signed in 2011 for $15.2 billion, but the NHL and Gary Bettman are still very pleased.
The Canadian market was definitely undervalued by the NHL previous to the Rogers deal. It was served by a number of different networks (TSN, SportsNet, CBC) that each received a share of games that would not be sold to other networks. Only a handful of games were televised that appealed to Canadian markets, so many games between U.S. based teams cannot be viewed in Canada without purchasing the leagues Center Ice subscription package. That will all change with the new Rogers deal where virtually every NHL game will be made available at various tiers of subscription prices. The much loved Hockey Night in Canada broadcasts every Saturday night by the CBC (broadcast since 1952) will be made available to the CBC by Rogers gratis for a period of five years, however Rogers will retain all revenues from the broadcasts. It is unclear how and if Rogers plans to sell broadcasting rights to games to other networks. Rogers already digitally streams some games to its cell-phone subscribers and this could be expanded to computers and third-party streaming services.
Economists understand the strategy by Bettman and the NHL to sell all of the rights in an “all or nothing” type auction. An all or nothing package extracts some of the monopoly surplus from Rogers that the NHL could not with a larger number of smaller bidders. Rogers earns a larger surplus (including profit) than it would if it were not the sole provider of NHL games in Canada, so it still has plenty of surplus left. The NFL follows a different business model, selling its television rights to 3 major networks (ESPN, CBS, NBC) and broadcasting some games itself on its NFL Network. However the sheer size of the NFL television market insures very competitive bidding – not the case in the NHL with its much smaller U.S. market. One could say that the Canadian market is subsidizing the U.S. market in the case of hockey giving Rogers a major toehold in NHL business decisions.
The loser will be the Canadian hockey fan who will pay a hefty price to watch games on exclusive Rogers SportsNet channels. And Don Cherry.
When the most recent print version of the Sports Business Journal arrived at Chez Miller, an above the fold headline read: “NFL: We Have Final Say in L.A.” with a subhead reading “Teams warned not to do own stadium deal.” Here’s the online version of the article.
The NFL has reiterated to its 32 clubs that Los Angeles is the league’s market and that any franchise seeking to negotiate its own stadium deal in the city could threaten the best economic result for the sport, according to team and other sources familiar with the matter.
The league outlined its points in a memo sent to clubs last month. In that memo, the NFL also cautioned that a team buying real estate in Los Angeles would not preclude the league from moving forward on its own stadium deal. There has been some concern in league circles that a team might squat on Los Angeles through buying land for a potential stadium.
The reason given in the article?
The NFL has shepherded the effort to return the league to Los Angeles almost since the Rams and Raiders departed the city after the 1994 season. Owners have sat through countless updates at meetings on Los Angeles and on the league’s perspective on what it will take to get a deal done — so a wild-card club moving in on its own is enough to spark some unease. The potential revenue of the Los Angeles market is another key element, because if a team were to strike a poor economic deal, it would mean less shared money for the rest of the league.
But we can’t ignore the elephant in the room: Los Angeles is a great, credible threat point that individual teams can use for leverage in stadium and lease negotiations with their local and state governments. This threat was effectively played, at least through the local press, when the Vikings negotiated for their new stadium.
Full collusion among NFL teams gives the best outcome for the league as a whole (monopoly is preferred to competition), but each team has an incentive to strike out on its own and negotiate its own deal. This should surprise none of our readers here at TSE, but sometimes the members of the cartel have to reminded of this.
The Nippon Professional Baseball players’ union has agreed to some changes in the Japanese Posting system that allows some NPB players to negotiate with MLB clubs. The posting system is on hiatus for this year while changes are hashed out.
The posting system allows NPB players not yet eligible for free agency a way to negotiate with MLB clubs. It was developed after Hideo Nomo and Alfonso Soriano exploited loopholes that allowed them to leave their NPB club to play in the states.
As it is currently structured, a NPB club may unilaterally post one of its players or a MLB can inquire about a particular player between Nov. 1 and March 1. Once the player is posted, MLB teams may make sealed bids which are given to the MLB commissioner. If no one bids, the player returns to his NPB club. If teams submit bids, then the MLB commissioner notifies the NPB commissioner of the amount of the winning bid, but not the team that made the bid. The NPB team has four days to accept or reject the bid. If the bid is rejected, the player returns to his NPB team. If the bid is accepted, the winner gets a 30 day window to negotiate with the player. If no agreement is reached, the player returns to his NPB club, but the MLB team does not have to pay its bid. If an agreement is reached, the player goes to the states and the MLB team pays its bid as compensation to the NPB team.
Under the proposed changes, if negotiation rights are awarded and the two parties can’t reach an agreement, the MLB team would pay a fine, although I haven’t seen how high the fine would be. However, if an agreement is reached, then the MLB team would only pay the average of the top two bids. I presume that if there were only one bidder, that team would still have to pay its bid if negotiations were ultimately successful. The union has agreed to these changes for only two years.
There are a couple of potential reasons for the changes. One is they will keep MLB teams from setting ridiculously-high bids to block other teams from getting players. As currently structured, the posting system does not penalize bad faith bids.
A second reason suggested at Baseball Reference is that the changes will lower the costs for MLB teams, especially when highly-regarded players are posted. In 2011, the Texas Rangers bid $51.7 million for the right to bargain with Yu Darvish and the Dodgers bid $25.7 million to bargain with Ryu Hyun Jin. This year, highly-regarded pitcher Masahiro Tanaka is expected to be posted (if the revisions are accepted by the NPB).
However, I’m skeptical of this second reason because a lowering of the fee would just leave more money on the table to sign the player. It’s not at all clear that this would lower the overall cost (fee + salary) of the winning MLB team. Come to think of it, perhaps this is why the NPB players’ union signed off on the changes.
In any case, the NPB still has to sign off on the changes and until it does so, the posting system will be on hiatus until further notice.
While Yankees v. Dodgers would been MLB’s dream television matchup for the World Series, the Red Sox v. Cardinals setup is likely better than most. When these same two teams met in the 2004 fall classic, they averaged the second most viewers of any Series since 1986. Unfortunately for overall viewership, the Red Sox swept the series in 4 games. However, so far, viewership is well below that of 2004, in the 15 million viewer range versus 22-25 million per game in 2004. Even at 2004 WS viewership levels, baseball’s premier event would draw only about the equivalent of an average Sunday Night Football regular season matchup. In years with lower profile teams playing, the viewership for MBL’s showcase looks more like an audience for each week’s Monday night game on ESPN. In contrast, the Super Bowl draws over 100 million viewers.
For many years, this kind of disparity between national television audiences for football and baseball led to concerns over the health of baseball along with discussions as to how baseball might “fix” the problem. Over time, however, it has become clearer that baseball’s future in terms of television revenues lay along a different path.
Before the late 1960s, there was no Super Bowl, and it took a decade for it to evolve into a major, primetime television event. Monday Night Football did not emerge until the early 1970s. College Football televised on or two games per Saturday. In this era of American sports history, the World Series stood at the pinnacle of sports consumer interest. By the 1980s, pro football had fully gained its television legs, and after a key 1984 Supreme Court decision, college football’s television footprint began expanding.
Baseball appeared relegated to the minor leagues. As the figures above indicate, in terms of national television audiences, the World Series barely matches up a run-of-the-mill NFL prime time game. The lesson MLB has learned is that a pie doesn’t have to come all in one pan. The big for numbers for MLB are not found in national audiences for a single game or series, but, instead are tied to regional telecasts. Beyond the Yankees, the Dodgers, Angels, Rangers, and teams in much smaller markets draw thousands of television fans per game over the course of the 162 game season. This Forbes article on Baseball’s Biggest Deals provides details. The overall pie is split into a lot of smaller pans. The aggregate numbers for TV revenues are still below the NFL but are sizable.
While NFL fans have regional team interests, they also watch games involving other teams. Consumers care about the NFL’s product at both a team-specific and league-wide level. In contrast, baseball consumers exhibit much more limited league-wide interest than NFL consumers and much less than MLB consumers of 60 years ago. The bulk of baseball consumer interests center on a specific team. It took baseball quite a while to fully grasp that evolution, so even though the World Series may not be what it once was to sports consumers, it really isn’t an accurate indicator of baseball’s health anymore.
The WSJ reports that Governor Phil Bryant is allocating $15 million from Mississippi’s share of BP oil spill damages to be invested in a minor league ballpark in Biloxi. The city itself has voted to issue $21 million in bonds backed by stadium revenues, but the deal apparently hinged on the state’s investment of BP funds.
Biloxi’s population has declined 10% (to 45,000) since Hurricane Katrina in 2005, and the 2010 oil spill must have added to the economic destruction. But BP’s oil spill didn’t ruin baseball in Biloxi — it damaged the beaches, the livelihoods of fishermen, and so on. It’s a stretch to take BP’s settlement money and pour it into new stadium construction.
A counter-argument might be that building a baseball stadium is the most effective way to stimulate economic development in Biloxi, and generates a higher return on public investment than alternative projects. But it is well known that stadium building is an effect, not a cause, of economic development — see Coates and Humphreys (2000) for example. Nevertheless, the city council’s 5-2 vote approving the stadium project was cheered by “most of the 150 residents and business owners” attending a recent meeting.
No doubt there are real benefits to be realized from this project, although the fact that outside funding is required suggests they may total less than the costs. I’d wager that the Beau Rivage Resort and Casino, on whose land the stadium will be built, is the primary beneficiary, and an active political player in the deal to approve it.
Six games into the season, Denver Broncos quarterback Peyton Manning has thrown for 22 touchdowns and only 2 interceptions. If he continues at or near his present pace, he will generate the best statistical season of his career and one of the best by any quarterback ever, while at the age of 37. How long can he sustain this level of performance?
Sometimes, sports commentators let their sentiment run wild when it comes to long-lived and well-liked players like Manning or the Yankees’ Derek Jeter and make statements to the effect that they can continue as long as they have desire. A less enthusiastic consideration from sports and life suggests otherwise. Time and age inevitably diminish athletic performance. Yale economist, Ray Fair, supplies extensive data on running and field event performances by age along with swimming and even chess. Masters running records indicate that by age 50 runners, whether in the 100 meters or mile, drop off by about 10 percent relative to world records. Of course, the incentives driving these records are not nearly as substantial as the open records and may not attract the most talented performers. During the height of PED use, particularly in baseball, 40 became the new 28, but even the modest enforcement policies of MLB have curtailed the drug-enhanced career lengths.
How long can Manning keep it up? After all, it’s the skill in Manning’s head and arm more so than his ability to run 100 meters that makes him great. Evidence collected specifically for NFL quarterbacks by Advanced NFL Stats offers clues. QBs over 35 tend to drop off with another precipitous drop after 37. These data indicate that the drop-off does not happen gradually. Instead, it’s usually quite sudden. As their analysis points out, however, the sudden drop-off could be an artifact of the data. All QBs experience up and downs in their careers, only to bounce back toward the mean in subsequent years. With a QB in his late 30s, the down year frequently leads to retirement and may accompany an injury, which imposes a selection bias on the data. Another selection issue with the data is that most (starting) quarterbacks who continue playing into their late 30s are among the league’s best, which makes for a small sample.