Skip to content

NCAA Seeding — A Rejoinder to Brian Goff

2013 April 3
Comments Off
by Victor Matheson

In his latest post Brian Goff notes that top seeds in this year’s tournament had some trouble putting away their lower ranked opponents and that seed wasn’t a strong predictor of final win margin. While I must agree that I certainly didn’t have FGCU making it to the Sweet Sixteen or Wichita State in the Final Four, at least historically, the people at the NCAA who put together the brackets have done a remarkably good job seeding the teams. Obviously, there is always uncertainty of outcome in any sporting event – that’s quite a bit of the allure of spectator sports – but higher seeds tend to do better than lower ones and beat lower seeds by higher margins. Here’s the data since the men’s tournament went to a fully seeded 64-team tournament (minus the last couple of years I haven’t gotten around to updating yet.)

Seed

Men’s Tournament, 1985-2011

Win Margin

Win %

1

25.84

100.0

2

16.77

96.2

3

11.53

84.6

4

9.49

78.8

5

4.54

66.3

6

3.94

68.3

7

2.20

58.7

8

-0.16

48.1

The margin of victory uniformly falls with the seed and win percentage is fairly uniform as well with the exception of the 5 and 6 seed where there isn’t an observable difference. There is a slight anomaly in the 8 vs. 9 game with 9s winning more often, but call that one a coin flip. Overall, I would give a great deal of credit to the bracket makers for their overall success.

That being said, it is possible that these results are being driven by good seeding in earlier decades, and an examination of whether seeding accuracy or competitive balance has changed over time would be of interest.

NCAA Seeding — A Lot of Noise

2013 April 1
by Brian Goff

The 2013 NCAA tournament has afforded another experiment on whether the seeding process incorporates much more noise than signal.  For those following the tournament, it should come as no surprise that this year offers more confirmation of the suspect seeding process.

I collected scores and seed differences for all 48 games from the first two rounds.  Using a common statistical technique (regression analysis), I examined the relationship between seed differences and score differences.  Seed differences explain a paltry 6 percent of score differences.  When the 1-16 and 2-15 are dropped, this percentage falls to only 3 percent.  Even as a predictor of win-loss rather than score, seed differences fare poorly.  For many of the games, plucking scores from a lottery hopper would come close to providing as much information on outcomes.

Is seeding really intended to predict score differential?  Obviously, not directly.  However, seeding reflects a gauge of team’s in-season performance quality – wins and losses adjusted for quality of competition.  If this gauge has much meaning to it, better seeds indicating higher quality teams, differences between them should show up in scores.  Instead, games with large seed differences wind up close or in upsets and games with narrow seed differences result in some blowouts.

It’s not just the NCAA Selection Committee that struggled with finding meaningful differences between teams.  Vegas point spreads, direct estimates of score differences, predicted only about 17 percent of score differences.  That’s a stunningly low number given that these spreads incorporate information such as injuries and best guesses about team specific matchups.

The 1 and 2 seeds are still very likely to win their opening round games, but even these games have become much more contested.  Based on this year’s tourney, I might amend original suggestion to select seeds 1-4 with everyone else randomly selected to just selecting the 1 seeds and randomizing from there.  Yes, that would lead to some seemingly strong teams facing other strong teams in the first round, and weak versus weak, but, that’s already happening.  The setup just puts a different face on it.

What Are the Falcons Worth to Atlanta?

2013 March 18
by Skip Sauer

“WHEREAS, due to the continuing economic benefits to be derived from the New Stadium Project by the citizens of the City of Atlanta and the State of Georgia…”  Thus begins the latest diversion of revenue from a city’s coffers to a professional sports franchise.

Various reports indicate that the negotiations between Atlanta and the NFL’s Falcons over a new stadium are nearing a conclusion. A story in Friday’s Atlanta Journal and Constitution presents a number of basic facts.  The estimated construction cost is an eye-popping one billion dollars.  The public will contribute two hundred million dollars up front through a bond issue, with the Falcons responsible for the rest.  The deal appears to fit the contemporary norm for stadium building and funding.  As in Dallas and New York, a building costing a billion bucks will doubtless be a super spruced-up sports palace.  But along with the higher price tag, the public’s share of the costs is below the norm for the mid to late 20th century.  Raymond Keating’s 1999 survey of stadium costs (“Sports Pork“) estimated the public share of construction costs at about 75%, roughly three times the public share in the present case.  

Neil deMause presents a detailed breakdown of other elements in Atlanta’s “New Stadium Project Financing Proposal,” along with a valuable link to the 200 page proposal itself.  To the construction costs, add $24 million in land costs and $30 million in construction sales tax rebates.  This pushes the subsidized share of full construction costs to about 25%.  That’s not the end of public support however.  In addition, the Falcons will receive  revenues from the city’s 7% Hotel Motel Tax.  This tax, which would sunset under current law in 2020, is extended another 30 years to 2050.  (The extension of this distortionary tax is perhaps the most inefficient aspect of the agreement).  deMause reports that the Falcons’s share of this tax currently runs about $17 million.  These funds are designated for operation and maintenance of the stadium — costs that in the absence of public involvement would be charged to the Falcons (or the Atlanta Falcons Stadium Company, a separate legal entity).  

There are a number of offsets in the various flows of costs and revenue, but deMause’s back of the envelope estimate is that the total public subsidy in this deal is on the order of $554 million.  The costs are spread out through bond issuance, and flows from the Hotel Motel Tax which accrue in the year they are incurred.

This is a significant subsidy, one that the Falcons’ ownership will take happily to the bank.  But it’s standard operating procedure in the monopolized world of North American professional sport.  Monopoly control over the number of franchises creates potential competition from a host city without a team, and through that an exit threat for which citizens in towns with current teams must compensate.  As an economist, with my normative hat I decry the system which generates this diversion of funds between now and 2050 to a sports entity, rather than projects which would truly help develop Atlanta’s economy.

But given the rules of the game, the price tag doesn’t strike me as one that the citizens of Atlanta will get all worked up about.  The subsidy amounts to about $20 million per year — a significant but not massive fraction of ticket revenues.  With about 5 million people living in the Atlanta MSA, the per capita cost is about $4 per year.  I suspect that if the question were asked, “Falcons for $4″ versus “no Falcons”, the people would fork over the $4.  I’m not a fan of this process at all, but must say “well played, Arthur Blank.”

Conference Realignments: (Some) Order out of Chaos

2013 March 8
by Brian Goff

The poaching of past Big East schools by the ACC and Big East and similar events in other conferences has engendered accusations of greed, disloyalty, and nastier things.  Jilted institutions and league officials have sounded a lot like spurned lovers.  While soap operas have appeared as little more of than a mad scramble for cash, a degree of order has emerged.  In fact, the latest headlines pointing toward Xavier and Butler joining  the “Catholic 7” Big East basketball schools demonstrate a sorting out apples with apples and oranges with oranges – call it the “eHarmony” model of relationship matching where “underlying compatibility” matters.

Money, of course, has driven decisions.  That’s not exactly a bombshell, and it didn’t start in the last decade as the initial formation of the Big East in 1979 bears witness.   The landmark 1984 Supreme Court decision striking down NCAA restrictions on televising football started a snowball moving that has gained speed and size over time.  Prior to that decision and for a while after, the formation of conferences like the Big East, Metro, and Sun Belt reflected basketball-centered schools coalescing and seeking market shares.   Ultimately, the unleashed football genie began to dictate alignments and re-alignments.  The 1991 Notre Dame contract with NBC marked a watershed.  It nudged other broadcasters into more lucrative contracts with conferences, enhancing major conferences as key bargaining units and making them more attractive to major independents like Penn State, while encouraging existing conference members to explore new relationships.

The affinities in the new relationships reflected underlying similarities in the size of the fan base (or potential base), sport of emphasis, academic standards, and geographic settings.  Arkansas and South Carolina went to the SEC.  The SWC married the Big 8.  These seemed like big shifts at the time, but the matching process was just getting rolling in the 1990s.  Many potentially advantageous relationships went undeveloped.  Texas A&M, for instance, while in many ways fitting much better with the SEC than the Big 12 did not make the move until 2012.

Assorted political obstacles and school-specific interests, such as longstanding rivalries and local interests, restrained movements as with Texas A&M.  In other cases, these obstacles persist and won’t likely vanish anytime soon.  A fortunate few institutions benefit from “grandfathering.”  For example, if the SEC were starting from scratch today, Vanderbilt would be looking for an association with other private schools, and it’s very doubtful that two Mississippi schools would be invited.  Northwestern would not be in the Big Ten.  In other cases, the mix of football and basketball influences decisions.   The basketball-rich tradition of North Carolina has kept it linked to other ACC schools when it could likely have jumped to a more prosperous league.  Turf protection keeps some schools out of conferences where they would seemingly fit well.   Some of the current matches reflect conferences picking from the leftovers to fill out their roster such as the Big 12 picking up West Virginia.  Why the Big 12 hasn’t gone for BYU (or the other way around) or Florida State is hard to understand.

Yes, outcomes in big revenue college sports will always be convoluted.  After all, the structure couples essentially professional sports entertainment enterprises with not-for-profit academic institutions wrapped in the antiquated vocabulary of amateur sports.  Nonetheless, business relationships are most stable when the objectives, revenues, and cost structures are more in line with each other.

The 7 Big East, urban campus, private, basketball-centered schools gravitating toward each other and a couple of other similar schools and away from football-centered, large state universities makes sense.  It makes a lot more sense than these schools adding another dissimilar institution from across the country like Boise State.

NBA Centers Facing Extinction

2013 February 20
by Brian Goff

Russell, Chamberlain, Jabbar, Lanier, Walton, Gilmore, Malone, Olajuwan, Ewing, O’Neal.  Over decades, these and other “centers” dominated the NBA.  The “association” conducted its annual All-Star game on Sunday.   For the first time, the league did not list “Center” as a separate position for fan voting.  Yes, some of the players in the game still carry the label, but only one of the starters, Dwight Howard, possesses the physical characteristics or style of play historically attached to the position.  In altering its All-Star ballot, the NBA merely officially recognized the trend in the game – centers have been trending toward T-Rex’s fate along with “Sky Hook,” “Dream Shake,” and “up and under.”

These trends can be quickly seen through scoring statistics.  Whether looking at yearly leaders or career leaders, only Dwight Howard appears in places where the centers of the sixties, seventies, and eighties did, and there are no emerging players of this sort.  The data support extends far beyond points.  It’s also observable in rebounding and shot blocking statistics.  Whether “center” or “low post” player is the label, the type of players associated with the positions are vanishing.

The devolution of the center position reflects a series of indirect effects and outcomes rather than  direct managerial choices.  The adoption and popularization of the 3-point shot made a big impact.  Ostensibly, the shot would open up play in the middle, but over the long run it devalued size and interior skill.  Prior to the 3-pointer, closer shots tended to be better shots, and highly skilled big players excelled at making close shots.  The 3-point shot slowly changed team offensive strategy as well as the long player development.

Along with the 3-point shot, the ratcheting up of physical play in the middle rendered low post play less effective.  By the mid to late 1980s players such as Rick Mahorn and Kevin McHale wrestled for post position.  In earlier eras, court position was ceded based on first to a spot.  By the 1990s, the low post wrestling morphed into all out body-to-body combat.  Centers with fade-away shots such as Olajuwan or Ewing and O’Neal with his brute force could survive but the die was cast.  The same forces bubbled up at the collegiate level.  I sometimes hear fans and analysts express dismay at the “lack of low post skills,” but low post skill is impossible to employ when defensive players are permitted, literally, to shove offensive players.     With high school and college post player becoming increasingly less effective, coaches at these levels adjusted their offensive schemes.  By the 2000s, few players emerge into the professional ranks with low post skills.

Is the extinction of center play a part of NBA brand management?  I can’t say for sure because I have not sat in on NBA executive meetings, but I doubt it.  Brand management in sports is unique in many respects.  Mrs. Smith’s apple pies follow the recipe and packaging decisions of management.  In the film industry segment of entertainment, movies and their marketing display the choices of producers, directors, and screenwriters with a little bit of discretion of actors thrown in.  In contrast, sports entertainment executives exert much less control over their product, at least the on-the-court (or field) aspect.  They can influence it through choice of  rules regarding eligibility, scoring, and fouls (to some extent), but the final combination on display on NBA courts mixes these choices with the unintended consequences of them along with team methods and with the long run evolution of player skills.

This lack of direct control also makes reversing these trends problematic.  In part, the league faces difficult tradeoffs.  The league values the 3-point shot and the the athleticism of  Michael in his era and and LeBron today.  Yet, promoting these aspects of the game may come at the expense of low post play.  Reducing physical play offers an opportunity for positive impacts on both perimeter and low post play, but significantly diminishing aggressive defensive tactics is not an easy task – that’s a topic from my recent Super Bowl post, and one that deserves its own discussion.

Baseball Arbitration Bats 1,000 For the Year

2013 February 18
Comments Off
by Phil Miller

For the first time ever this year, no baseball player went to arbitration.  From ESPN.

Reliever Darren O’Day completed a $5.8 million, two-year contract with the Baltimore Orioleson Monday, becoming the 133rd and final player to settle without a hearing among the 133 who filed for arbitration Jan. 13.

This was the first year since arbitration began in 1974 that no player who filed went to a hearing.

Link via my former student David Dicks.

Baseball uses a form of final offer arbitration in which an arbitrator is restricted to selecting either the final offer submitted by the player or his team (although technically speaking, the offers aren’t final since negotiations can continue even after offers have been submitted).

The economists in the TSE audience may be interested to know that final offer arbitration was originally proposed by the economist Carl Stevens in this 1966 paper.  Up to that time, the main type of arbitration that was used, commonly called “conventional arbitration”, was one where arbitrators could render any decision in a case, and it was not at all uncommon for the judgment to be right between what each party wanted, basically splitting the difference.  Some observers felt that this supposed splitting-the-difference would chill bargaining.  Why go through the trouble of negotiating when the arbitrator will just split the difference?

Stevens wanted a type of arbitration that would lead to more negotiated settlements, believing that settlements arrived at in that manner are superior to those imposed by third parties.  In theory, final offer arbitration should lead to more negotiated settlements than conventional arbitration because FOA makes losing a case more costly to both sides.

This year baseball arbitration batted 1,000.

Here’s a 1990 Chicago Tribune article that has some quotes by Stevens on what he originally called “either-or arbitration”.

Fans Care about PED Use — Sometimes

2013 February 14
by Brian Goff

What spurred the public’s negative view of the use of PEDs?  Why have MLB and pro cycling received much more attention than the NFL?  Some would say fans don’t care, even in the MLB and cycing cases, that it’s really just a media-generated series of events.  No doubt, media writers and talking heads can get worked up over things that create yawns among most fans.  In the case of the steroid stories, their long-lasting legs suggests that they resonate with the sporting public to some extent.

The MLB Hall of Fame voting  supplies the explanation — not the details but its very existence.  When PED-fueled performances start encroaching on cherished legacies, fans start to pull back.  In the mid and late 1990s, the baseball-watching public shrugged at  whether Mark McGwire or Sammy Sosa were bashing on higher octane additives.  McGwire’s 57 homers in 1997, just a pinch below Maris’ 61, promoted fan interest rather than probing questions.  By 2001, McGwire, Sosa, and Barry Bonds had surpassed the landmark six times.  Players from the mid 1990s to early 2000s climbed  the 50 homer plateau 23 times, more than in all of the years prior combined.  Beyond the yearly marks, Bonds surpassing Mays, Ruth, and even Aaron raised more fan eyebrows along with eight players from this era who wound up among the top 15 home run hitters of all time.  Former players whose achievements were being devalued raised voices in protest as Hall of Fame voters and fan views soured.  As a fan myself, when Bonds passed the 660 home runs of one of my first sports heroes, Willie Mays, it sickened me.

Of course, the attacks on legacies also invited more media attention and investigation, leading to more revelations and negative publicity to foment fan distaste.  This avenue seems especially apparent in the Lance Armstrong case.  Had Armstrong won 2, 3, or even 4 Tours, he may well have escaped with his long run reputation somewhat in question but relatively intact.  That’s how it worked out for Miguel Indurain, who, ironically, was still publicly touting belief in Armstrong’s innocence as late as October in this Guardian article.  Rising up to the hallowed 5 wins and then blowing past it put Armstrong in the bulls-eye of many European writers as well as international doping regulators.  While his return to the sport, by his own estimation, may have led to his successful prosecution, his smashing of legacies of riders like Indurain and Hinault made him both popular and a popular target.

The other side of legacy effect on fan backlash is the lack of it in sports where legacies are not impacted so much, such as the NFL.  At first glace, one might conclude that football fans don’t care as much as baseball fans, when the underlying difference may be the lack of legacy effects.  In part, the lack of testing may, ironically, play a role.  Cycling has been stung the most, in part, because of its rigorous attempts to get rid of PED use.  MLB has been less diligent, and the NFL has lagged behind even MLB.  In addition, the team nature of these sports and their outcomes has shielded players from the kind of scrutiny given in individual sports or in a “individual sport with a team score” as baseball is described.  In recent years, quarterbacks and receivers have run roughshod over many passing achievements of the past, but this can be attributed to changes in rules and enforcement that make passing easier and promote more sophistication in passing, not to QB or wide receiver PED use.

Widening out the lens from sports and sports legacies, similar outcomes can be observed with companies.  Extraordinary success invites extraordinary scrutiny, whether justified or not, and whether it applies to athletes like Lance Armstrong, or companies like IBM in the sixties and seventies or Microsoft, WalMart, and Google in more recent years.  How a limited supply of scrutiny is doled out or, more to my point, gets traction among the public is more puzzling.

Games at Work on Super Bowl Officials

2013 February 6
by Brian Goff

The Super Bowl officiating has generated a loud buzz of activity, much of it centered on the “no-call” of holding against the Ravens on the 4th down from the 5 yard line.  From my seat, 49er fans probably have some legitimate complaints over that call along with a no-call on Ed Reed for offside on a player near the goal line in the first half, which was particularly glaring in  that an offside call on the 49ers setup the first Ravens’ TD.

Beyond the advantage gained or lost by either team, however, the big 4th down play summed up the entire game — officials, whether on their own or by directives by supervisors — kept their flags tucked away on many discretionary calls penalized during regular season games.   This kind of reluctance could be seen throughout the playoffs, but especially in the Super Bowl.  From the outset, post-play activities that could be penalized as personal fouls were overlooked.   When these antics reached their peak with the mini fight/scrum in which at least one player had his helmet ripped off  and one shoved an official (normally automatic ejections), both teams received an inconsequential offsetting penalty.  Even though both teams’ secondary players aggressively used their hands all game in ways normally penalized, I don’t remember a single defensive holding call.

The NFL playoffs and Super Bowl in particular resembled the shift that annually occurs in the NBA between (most) regular season games and playoff games (and some bigger regular season ones).   Players ratchet up the physical play, knowing that the league doesn’t want the “game decided by the officials.”  Some analysts laud this “the them play” approach, not seeming to realize that it doesn’t permit many other players to “play their game.”

It’s a very difficult strategic game for leagues.  In pro basketball playoffs as well as in college basketball in the late 1980s and 1990s, coaches and players figured out that leagues and officials prefer not to foul out more than one or two players in a game.  If a team commits 25 acts that could be whistled for a foul, maybe 20 would be called.  However, if a team commits 250 punishable acts, then only 25 would be called.  That’s a small price for a big increase in defensive aggression. Now, NBA playoffs, particularly the deeper rounds, and NCAA games devolve into wrestling matches as ESPN’s Jay Bilas, a former Duke player who enjoys defensive play, observes in this Sportstalk Chat.

The NFL faces the same issues with pass defense.   As passing offenses of the 1980s and 1990s became more effective, the best NFL defenses adopted, essentially, a holding everywhere and all the time policy.  The league responded with a crackdown in the early 2000s, increasing the number of holding penalties.  On occasion, one can observe holding penalties on back-to-back plays during regular season games.  However, the Super Bowl illustrates the difficulty of maintaining a hard line.  Players step up their aggressive tactics which gives officials two choices — call a load of holding penalties early or permit a teams to engage in tactics that are usually penalized.   Jim Nantz and Phil Simms, CBS’ announcing crew, seemed oblivious all night to this, with even the 4th down play with the two-handed bear hug by the Ravens’ defender only drawing a “close call” comment from Simms.

Ironically, in spite of the defensive wiggle room permitted, both teams scored over 30 points.  This reflects the development of offensive skills and exploitation of offense-spurring rules, particularly in the passing game.  At one time a 28-6 deficit in the second half would have been a yawner but not in this era.

Also Posted at Forbes SportsMoney

A not so Super Bowl for NOLA

2013 February 4
by Victor Matheson

While last night’s Super Bowl was certainly filled with excitement on the field, the game may ultimately be best remembered for the the excitement in the utility room as the game was marred by a 34-minute delay due to a power outage at the stadium.

Economists have long taken issue with the NFL and organizing committees’ claims that the Super Bowl brings hundreds of millions of dollars to the host city. While New Orleans claimed a $434 million impact from the game, academic work looking back at past Super Bowls seems to find an identifiable increase of only about one-fifth of that. For example, my 2006 study with Rob Baade of Super Bowls from 1970-2001 found an average increase of just $92 million and was able to reject gains of $300 million or more at a 5% significance level.

When boosters are pressed on the economic data, they will inevitably turn to the claim that we are missing all of the great intangible benefits of the game. The game brings great publicity to the host city, it is said. I often counter that publicity doesn’t always have to be positive, and last night’s events provide one more story to add to the list. Instead of focusing on New Orleans’ fantastic cuisine, culture, nightlife, and music scenes, viewers around the nation were treated to the spectacle of a city that couldn’t even keep its lights on during the biggest night of the year. Not exactly the thing that gets captains of industry to relocate their businesses to the area.

Indeed, rather than a moment in the spot light, NOLA received a moment (34 minutes worth, in fact) in the no lights.

Forget ‘Fab Four’, Men’s Tennis Still More Competitive

2013 January 25
by Liam Lenten

As the Australian Open approaches the business end, this year’s instalment has developed a palpable sense of familiarity of recent years about it. Each of the top four seeded men made the semi-finals, and it’s hard to escape the feeling that fourth-ranked Rafael Nadal would have been in the place of his compatriot David Ferrer if he were present. And so we have the familiar trio of Andy Murray, Roger Federer and Novak Djokovic left standing.

The story in the women’s singles has been notably different, and not just to this stage. There were a sprinkling of upsets in the first week – earlier in the tournament only 3 of the top 16 men’s seeds were eliminated before the fourth round, compared to 7 casualties from the top 16 seeded women.

In comparing the sexes, it is worth reflecting how the modern era of tennis is arguably an unusual one in terms of the relative competitiveness in both the men’s and women’s games.

The current prevailing wisdom among occasional observers of the sport is that men’s tennis is highly predictable – at least until the semis where the fab four seem to meet more often than not, and that now, it is the women’s draw that offers a far more balanced and interesting proposition.

Historically, this is unusual – but long-passed are the successive eras of Margaret Court, Chris Evert/Martina Navratilova and Steffi Graf/Monica Seles. Back then, you could almost bet your vital bodily organs that any Grand Slam trophy would end up safely in one of only two or three possible pairs of female hands, even before a ball was served in anger. Meanwhile, the four majors seemed to be far more contestable on the male-side of the draw.

Things began to change in the early-to-mid 1990s – the rise of emerging women, such as Arantxa Sánchez Vicario, Mary Pierce, Martina Hingis and Lindsay Davenport, renewed interest in the women’s game. Concurrently, the Sampras/Agassi duopoly began to hold sway, as it would for the better part of the following decade.

A short period followed in which Jennifer Capriati and the Williams sisters dominated the majors, at a time when the male duopoly loosened its grip and other men, including Lleyton Hewitt, had comparable claims to slam titles.

In the decade since, however, other women such as Justine Henin and Maria Sharapova have once again challenged at the top to produce a more open field, while the ascension of Roger Federer, since joined by the other three members of the quartet, has made the men’s game look like a closed shop.

However, this prevailing wisdom is ill-informed. Many of us will only watch a handful of games on television during the Open. Locals, even certified tennis tragics, who go to Melbourne Park will not see many televised matches in the other months of the year, let alone live overseas.

In fact, the ATP and WTA tours each involve roughly 2,500 matches each year in the top-four tournament tiers (and the season-ending championships), and only an analysis of all of these matches can tell us the complete story of the overall tour, as opposed to merely anecdotal evidence at the very apex of the talent pool.

I was motivated to find whether this conventional wisdom was indeed true over the entire tours, and so having obtained data on all matches (excluding walkovers, retirements and disqualifications), I set out to find the answer. I went back as far as 2007 – the last year in which the fab four did not quadrella the end-of year rankings (Andy Murray would join them in 2008). During this period, Serena Williams found a way to win a further eight slam titles during her post-injury renaissance, but she hasn’t had it all her own way – she has only occupied the top end-of-year ranking spot once (twice finishing outside the top five).

One simple statistic would be the percentage of matches that go to a third set. For this, we would have to exclude best-of-five set men’s matches (mostly grand slam) to circumvent an unclean comparison.

On this score, more men’s matches are competitive, with 33.8% going to a deciding set compared to 31.2% of women’s matches (31.6% if grand slam matches are excluded for closer comparability).

Another sensible statistic is the proportion of matches where the favourite was beaten, as a measure of propensity for upsets. Only matches in which all (at least three) bookmakers odds were unanimous on the favourite were analysed.

Here, slightly more women’s matches were upsets (26.8% to 26.6%), but again, best-of-three and best-of-five matches are not comparable – statistically, the underdog is less likely to beat the favourite when the match is longer.

When grand slam matches are excluded for both sexes (plus a handful of other best-of-five tour event finals), the scales again tip in favour of the men (28.1 to 27.7%). The story was similar with two other metrics I investigated, specifically the median of the log-ratio of the betting odds and the average number of games won by the loser.

It is worth noting that these differences are not statistically significant, but it dispels the prevailing wisdom that elite women’s tennis has been more competitive in recent times.

I emphasise that this result should not be taken as an endorsement of men’s tennis over women’s; rather, it warns of the ever-present danger of various stakeholder groups in sport (or any other industry for that matter) making false inferences due to extrapolating on a small number of matches (or observations) to make crude generalisations without looking at the data in its entirety.

Ultimately, however, the title odds at Melbourne Park were more open on the women’s side two weeks ago. Therefore, next time you have to decide between watching a men’s and women’s match, there are numerous factors that will determine whether or not it will be an exciting and close match, so I would advise to simply use your own discretion.