Ruminations on RSD

This is spurred by a couple of blog posts at Phil Birnbaum's site ("Noll-Scully doesn't measure anything real") and TangoTiger's site ("Trap of Noll-Scully").  And I also hope it helps with some Twitter discussion that didn't work out there (@BMMillsy, @guymolyneux, @dataandme).

Here's what I understand about the Noll-Scully RSD measure plus the best that I can make out about the issues raised above.

Thought process.  The standard deviation of final winning percent is useful in relation to competitive balance (which also follows from the underlying distribution of talent in the league).  Suppose one version of a perfectly balanced league, given by Pr(win)=0.5 for all teams and all games.  Start this version of a completely competitively balanced league at G₁ games.  Change it to G₂ games.  What happens?

Let ISD bet the standard deviation of this version of a completely balanced league.  Fort and Quirk (Journal of Economic Literature, 1992) show that ISD=0.5/sqrt(G) for the binomial without ties.  Thus, moving from G₁ to G₂, ISD(G₁) will be different than ISD(G₂) because ISD depends on G.  This helps to make clear that in general, the standard deviation of winning percent depends on G as well.  Let ASD be any standard deviation of winning percent from the league.

Now let's think about what to do with this knowledge of the distribution of winning percent.

Here is what I get from the discussion/Twitter noted above.  Suppose we have a statistic Z that measures the outcome of applying the talent distribution in league play.  In a league with season length G₁, we get Z(G₁).  If the league changes to season length G₂, we get Z(G₂).  Define a successful Z to have the following characteristic:  Z(G₁) = Z(G₂) because the underlying talent distribution is the same in either case, just applied in leagues of different season length.

So, how to reconcile Z and ASD?  Even in a league of equal playing strength, so that Pr(win)=0.5, ISD changes with G and so will ASD generally.

Let’s consider three alternatives (there may be more).

Alternative 1:  Z^* = ASD ± dASD/dG.  Now it will be the case that Z^*(G₁) = Z^*(G₂) because the impact of just changing season length will be netted out by calculation of the impact of G on ASD and addition or subtraction.  This is sort of like an “inflation adjustment”.  The distribution of talent didn’t change and our Z^* provides the comforting result that it is the same for either G₁ or G₂.  Of course, this requires knowing dASD/dG.

Alternative 2:  RSD = ASD/ISD.  It is immediately clear that there is no way that RSD can be a successful Z.  It will always change with G and it contains no adjustment that would make it stay the same regardless of G.

Alternative 3:  Just dump the standard deviation as useful because you don’t like either of the above.  There are other measures of final season competitive balance as a reflection of the distribution of talent.  But don’t propose a game-level or playoff access or dynasty alternative since we’re talking about final season competitive balance.  There are other measures of those other aspects of balance as well.

So the point is well made that RSD cannot be a candidate for Z.  But it was never intended to be such (I know Noll quite well and knew Scully well prior to his death).  It really is meant to be the distance comparison measure, 5 steps from my door are farther than 2 steps from your door, so I am farther from my door.

Perhaps there is just a semantics misunderstanding when the literature using RSD states that it "controls" for G?  Surely the Z^* measure does this forcefully.  But RSD does it relatively, so maybe a better way of saying it is that RSD "recognizes" G in its relative comparison.

Some concluding comments...

So far, I haven’t seen anybody take a crack at calculating dASD/dG.  I wonder if the related critics Owen & King (Economic Inquiry, 2015) are actually just simulating dASD/dG in which case they are an ally to those seeking Z^*.  One could just use their simulation results rather than trying to determine the derivate, dASD/dG.  Or perhaps the Pythagorean discussion in the references to blogs at the top of this post handle this problem already?  If not, then there is a ways to go still with Z^* development.

But I’m still not so sure that taking the "inflation adjustment approach" is any more informative than what is done with RSD.  Z^* distills the dASD/dG problem to an absolute level.  RSD just puts the comparison at a relative level.

It does seem to me that the Z* devotees are not really critiquing RSD as a normalization.  They would just prefer to take the direct Z^* approach rather than taking the relativist approach.

And I chose the word “prefer” carefully.  While it is easy to see that Z^* is different than RSD, I still don’t see how Z^* is superior to RSD.  And it is not enough to just say so.  In any event, if Z^* is shown to be better at a later date, future work will be the better for it.

In the meantime, I have competitive balance to compare, within a league where season length changes and across leagues with different season lengths.  And I haven't yet been dissuaded on RSD as one useful measure.

The "Upshot" of Portraying Sports Success Incorrectly

TheUpshot is at it again (David Leonhardt, NYT, 6/4/15), tweeted numerous times since).

The object this time is "city success" at pro sports.  The definition:

Success = Championships won, as a percent of possible chances across all of a city's teams, over the last 50 years.

Seems straightforward enough--higher percentage, successful sports city; lowest percentage, unsuccessful sports city.  Only two problems.  First, the measure, itself, gives ambiguous results.  Second, the definition of success, championships only, doesn't look like it coincides with a fan definition of success reflected by their observed behavior.

First, the measure.  Suppose one city has teams that won recently in the 50-year span, while another city has teams with the same percentage, but enjoyed 50 years ago.  Yet another city has teams with the same percentage spread over the entire 50 years.  All of these are equally successful cities by the definition.  But fans do not think so; dated success is just that.  If a city has only one team and another city has two teams, but both cities have the same percentage, both cities are equally successful again.  So two teams, say, half as successful are equivalent to a a possible dynasty.  Finally, and this is just the way the ball bounces, completely equal success in a given league would have 20 teams winning twice and 10 teams once over 50 years.  10 cities are then losers even as the distribution is trying to reveal itself as completely equal over time.  Finally, lumping leagues together in a given city means that a championship is a championship is a championship... say, NHL fans view an NBA championship the same as they would a hockey championship.  But does a championship by any other name really smell as sweet?  Ambiguity abounds.

Moving on to the definition:  Why just championships? [Appearances in semi-finals are charted but not labeled by team and the rest of the article clearly only focuses on winning championships.]  Again, there isn't much room for many to be successful at the championship level.  Instead, let's ask fans for their definition of success since it is their definition that matters.  If Bob Costas is correct, it's a legitimate belief that the post-season is possible.  But there are other indicators of fan preferences as well--the tens of thousands of fans driving hundreds of thousands (NFL) or millions (MLB) in total attendance, even without a championship.   Let alone the TV audience.  I guess this "success" is just irrelevant.  After all, no championships, right?  The alternative, of course, is that the assessment of "success" based on championships alone simply misses the mark.

Really, Mr. Leonhardt just points out championship concentration in major league pro sports that is news only to newcomers.  And they would be better served if Mr. Leonhardt pointed them to all the work already out there on that topic.  But that just doesn't "sell papers".

"Giving Up Raises" = Reallocation from Coaches to Players

In a recently re-worked contract, MTSU's head football coach Rick Stockstill is "foregoing" raises ($100,000 per year, next 4 years) that were in his previous contract (Will Borthick, DNJ 6/4/15).

He and the AD at MTSU claim it was Coach Stockstill's idea. And it's to make sure the MTSU athletic department can catch the wave of offering full cost of attendance (FCOA) to football recruits.   And the coach hopes maybe some of the $400,000 can go toward facilities (unspecified, but presumably football).

The tone of the article is that this is a sacrifice on the part of Coach Stockstill, but the result also has another economic explanation tied to the reallocation away from coaching salaries (and AD salaries, facilities) due to gains by players.  The gains are both real, e.g., the Power 5 move toward FCOA, and more probabilistic, e.g., the pending decision in O'Bannon.

Now, just how would markets reallocate in the face of the Power 5 choice to move to FCOA, and at "Group of 5" departments like MTSU that are not in the Power 5 but choose (and I stress the word) to move to FCOA anyway?

Since all of the money is already being collected by very effective ADs, and taking as given that MTSU has already determined its optimal athletic department size, the answer is: Just what you see happening at MTSU. Whether Coach Stockstill sees the handwriting on the wall, or whether it would have happened a bit later at the AD's insistence, this is the expected outcome across all departments choosing FCOA.  Reallocation away from coaches toward players.

[Should O'Bannon go the players' way, the same thing will happen at departments that choose to pay up to the court's "lid" on image rights payments.  The decision under appeal did not require any department to do so, remember.  But those departments that do choose to pay for image rights will behave the same as MTSU's athletic department.]

I find it simply fabulous that it is happening at MTSU, a "Group of 5" member rather than a Power 5 member, and it is happening as many of us have predicted--with coach's pay, not cutting sports as predicted by doom and gloom forecasters at places like the Knight Commission.

And make no mistake.  If it were truly charity, Coach Stockstill could have taken the raise and given it to the athletic department as a gift.

Instead, revenues being generated by players, being spent elsewhere in the department, will be diverted from those other spending areas to players at those departments moving to FCOA.

Of course, the behavior of one school does not make a data set.  But this is the shape of things to come. And don't be distracted when you see raises and FCOA going hand in hand at some departments. The raises would have been higher in the absence of the new FCOA policy.

Joe Nocera's Dog Just Won't Hunt

Joe Nocera, eminently readable as always, starts from a fundamental premise that powerful college sports-oriented alumni always win against the academic side ("Books vs. Games", NYT, May 12).  Of course there are other elements in the sports constituency than just alumni, but I digress.

Nocera reels off a number of truths that he holds to be self-evident:

"... the athletic department's out-of-control costs..."
"...Rutger's athletic department has consistently run large deficits."
"...money that might have gone to professors’ salaries or other academic needs."
"...took it out of the hide of the students themselves."

Hammering the point home on books vs. games:

"It [Rutgers] is responsible for educating 65,000 students. Why isn’t that more important that (sic) competing in the Big Ten?"

As the final nail in the coffin:  "Why does the tail always wag the dog?"

But as Jason Winfree and I point out in our 15 Sports Myths (Stanford U. Press, 2013), there is a much more productive explanation of the relationship between university administrators and their athletic departments.  Universities actually do both books and games and carefully weigh value and cost margins in making their spending decisions on each.  The choice by university administrators to put relatively small amounts of money into their athletic departments pays off in relatively small ways, but typically with a believable and legitimate return if you look in the right "spot" (only mild apologies for the doggy pun).

Compared to this alternative economic and policy description, the "tail wags the dog" belief is a dog that won't hunt.

Rather than just spending what it wants with university administrators close at hand with the poop bag, actual budget hearings place the athletic department in the same setting with the rest of the programs at the university.  The allowed spending decision comes first, then the spending.  And that budget is then overseen by the Board of Regents, the Governor, and then on to the legislature.  The athletic department is as controlled in its spending as any other program on campus.  [This may cause chuckles among those dissatisfied with the spending control over universities, generally, but that doesn't distinguish the athletic department on this dimension in any way.]

In an assertion begging justification, Nocera simply names the difference between generated revenues and actual expenditures a "deficit".  But this "deficit" actually represents the administrators' determination of the added amount of money they wish to spend in order to have their athletic program at its desired size and scope.  As it goes, the current level of administrative spending at Rutgers is around 25%  of the total (it has not always been that way, and last year's level is higher, which is why Rutgers is interesting).  However, especially during a conference transition, this is not uncommon and administrators expect to see a return on that investment.  If things go at Rutgers as they have elsewhere, they will, as Winfree and I have documented for all of college sports.

Next up for Nocera:  The administration's portion represented "...money that might have gone to professors’ salaries or other academic needs."  Having already ignored the budgeting process within the university, might as well ignore the rest of the budgeting process beyond the university!  Administrators decide the right level of spending at their university, tote up research revenues, ask their legislature and other funding sources to cover the rest, and spend the proceeds according to their request.  If $20 million or so for athletics is granted, and then administrators spend it on something else, they will have some explaining to do in the next budget round.  If university administrators tried to spend the allocation to athletics on academics, as Nocera suggests, it is likely that the next go-round would see $20 million or so less in the Rutgers budget.  Budgeted funds are simply not always as fungible across different purposes as many believe (desire?) them to be.

And taking it out of the hides of students is also an assertion outside reality.  Student fees go to fee-backed student activities, and the distribution of that budget is determined by student government.  So, actually, student government spends the money on athletics because that is the (admittedly never perfect) reflection of the self-assessed welfare of student government leaders.  Usually, this is also the source of the lower-price ticket quid pro quo.  Yes, I recognize the imperfect nature of the reflection of the preferences of some students.  But that is hardly unique to this little democracy.

Besides, we're really just not talking about all that much money from the overall Rutgers budgeting perspective.  Nocera reports $183 million through 2022.  Taking that to mean 7 years, that's $26 million on average.  Nocera also reports Rutgers' annual budget is $3.6 billion.  So, that's .026/3.6 = 0.7% on athletics (yes, less than 1%).

As to the faculty response, let's turn to just the 27.6% of budget, or $994 million, that goes to instruction (and ignore all of the other spending on students like housing, dining, aid, services, share of library and plant operations, etc.; see Rutgers budget description).  First, again, that's 26/994 = 2.6% and relatively speaking just not very large.  But, second, $26 million is $26 million and we are a jealous lot.  Any dollar going to anything except what we cherish (usually our own program and our own salary) is mis-spent.  And all of us would like $20 million added to our favorite.  Alas, third, as noted above, that $20 million or so a year isn't really ever going to go to our heart's desire in the first place.  Benefactors agreed to spend it on athletics, not on us.

So, university administrators balance books and games, just as they balance economics books against sports management books against everything else that the university is expected to do.  All of the dogs line up to be fed and they are, each according to their contribution to the goals of university administrators--research, teaching, and service.  They each have also already wagged their tails and their ration is the result of that demonstration.  The academic side does mostly research and teaching, and the sports side does mostly service.  But all are valuable to university administrators which is why they "feed" them all in the first place.

Pitting one against the other misses the point of just what it is that universities are up to in the first place, fomenting conflict along the way.  And to what end?

Before We Get Too Carried Away with the 8-year Project Royals...

Announcers are not sports economists, so I don't hold this too much against Rex Hudler (guest announcer on the Tigers game the other day).

Hudler went on a bit about how the KC Royals should not be such a surprise.  According to Hudler, the team was eight years in the making with careful strategy/moves etc. by sharp management.

No doubt; Kendrys Morales is really coming through for example, and some low-salary pitchers are going great guns so far.

But let's not forget this:

2011- Royals ranked last (30th) in MLB payroll at $36.1M.  They also received $29M in revenue sharing to take into...

2012 - Royals move up to 27th in MLB payroll at $60.9M.  If the increase of $24.8M all came out of previous revenue sharing, that would still leave $4.2 M.  The Royals received $16M in revenue sharing at the end of 2012, so their revenue sharing account now contained $20.2M.  And we're off to...

2013 - Royals move up to 22nd in MLB payroll at $80.5M.  If the increase of $19.6M all came out of previous revenue sharing, that would still leave $600,000.  The Royals received $36M in revenue sharing at the end of 2013, so their revenue sharing account now contained $36.6M.  And we're off to...

2014 - Royals move up again to 19th in MLB payroll at $92M. If the increase of $11.5M all came out of previous revenue sharing, that would still leave $21.1MM.  I could not find any published report of revenue sharing for 2014 (and a quick check with guru Maury Brown - @BizBallMaury - informed me that even devout followers cannot find the data anymore).  But don't let that us stop us...

2015 - Royals move up to 16th in MLB payroll at $113.6M.  The increase was $21.6M.  Given the $21.1M left last year, and given that net sharing to the Royals was surely positive in 2014, it appears a truly trivial amount came out of the 2014 share to get to this payroll level.

So, net revenue sharing proceeds for the Royals more than covered the 33% per annum nominal growth in payroll observed over the last few seasons.

Indeed, nearly all of 2014 net sharing must have gone into other player development expense, since that is required under the CBA and monitored by the Commissioner's office, and reviewable by the players.  At least in the case of the Royals, it looks like revenue sharing has done its job.  [I haven't looked at any other teams and this need not be true in all cases, as many others have argued in terms of "revenue sharing freeloaders" or some such.]

It takes sharp management to get where the Royals are today with pretty close to the median payroll.  But let's not forget that their payroll has dramatically increased, as tempting as it is to attribute it all to outstanding management on and off the field.

Sports Econmics: JEL Recognition and a Reminisce

So, sports economics has its own JEL codes.  In addition to other paths I have followed to sports economics publication (e.g., the “L-path” in Industrial Organization), there is now a “Z-path”.

I join the current leadership in sports economics—Peter von Allmen and Dave Berri— hailing the new JEL codes as overdue recognition of the area.  But I’m also sure the codes are a mixed blessing (note that sports economics is under "Specialty Topics" as Z2, that is, the second specialty topic after Cultural Economics, probably with more to come).  The Z-path does, indeed, recognize sports economics and that is nice.  But the codes also duplicate other paths (e.g., the L-path has an industry studies sub-specialty that includes sports).  The Z-path also codifies the “sub-specialty” path many journal editors tell us to take anyway, back to our specialty journals.

Whatever.  The new codes won’t change my approach to sports economics at all and I wonder if they will for anybody else?  For me, the new codes will just make it easier to identify some of my work for the editors at the JSE and IJSF, for example.

The JEL accomplishment does do one thing for me.  It reminds me of the continual progression of sports economics from some pretty humble origins.  My thanks go to Joel Maxcy and Bruce Johnson for their help on the following recollections.

Larry Hadley (I miss him) and Elizabeth Gustafson organized the first WEAI sessions at the San Diego Sheraton on Harbor Island in 1995.  They hosted a cocktail party in a suite at the Sheraton and we remember it being pretty crowded, but joint recollection dredges up only Larry, Elizabeth, Dan Marburger, Bruce Johnson, Bill Kaempfer, and yours truly.

Please let me know if you were there and I will update the recollection.  Thanks in advance.

The 1995 version was successful and Larry and Elizabeth continued on organizing sports economics sessions-- next at the Westerns in San Francisco in 1996 and Seattle in 1997.  By 1998 (Lake Tahoe), the sessions ran most of a day.  I know, pretty small compared to the all day/all conference sessions of today.

And they were very fruitful sessions that followed.  I remember especially the 2001 San Francisco meetings where I first met my long-time collaborator, Young Hoon Lee.  Our first paper was in, of course, EI (2005).  We now have nine pieces together, and a co-edited volume just out (The Sports Business in the Pacific Rim, Springer Verlag, 2015).  Like I said, fruitful indeed.

It is also worth noting the first sports economics professional association which still struggles on today, namely, the International Association of Sports Economists.  I recall completely the IASE founders, Paul Staudohar and Vladimir Andreff.  The early IASE activity brought together those who would later sort themselves out first into NAASE and, most recently, into EASE.

It continues to be an interesting progression.

Why Definitions Matter, Part 2: Demand or Quantity Demanded?

Here is my second installment on economic/business descriptive definitions and why they matter.  The first was on "bubbles".  Arguing for that distinction earned me the "pedantic" accusation on twitter from @patrick_hruby but he is just wrong, both in context and in label-slinging.  These types of distinctions are critical to communication; without them, we must always begin a conversation with, "What do you mean by that?"  And that seems inefficient and potentially quite confusing for no real reason.

Demand or Quantity Demanded?

This is an old favorite of economics professors and, judging from treatment in the press, always will be.  Just look at the two carefully:  "Demand" versus "Quantity Demanded".  They do not look the same and for good reason:  They do not mean the same thing!

Let's think about "demand" and "quantity demanded" in terms of what might be a current problem for college sports ADs, the recent decline in student attendance at football games.

If attendance falls off due to bad scheduling or ever-more-enticing TV viewing, that's a decrease in Demand.  The language is reserved for a change in attendance even at the same price, where it follows that attendance has changed at all prices that could be charged.  Responses to this type of decrease are aimed at changing Demand back to its previous levels:  If an AD brings in more air shows, flying rocket man for the pre-game, more blaring video boards, more extravagant half-time programs, or alters a non-market imposition on rights to student tickets, that is an attempt to change the product so that attendance will rebound even at the same price.  Again, if fans respond, that would be an increase in Demand.

[My friend and colleague Stef Szymanski recently turns this distinction to extremely good use by pointing out that rising EPL ticket prices will not drive fans away...when they are rising because fan demand is increasing in the first place!]

On the other hand, an increase in Quantity Demanded would follow a ticket price reduction to bring fans back, more people will come through the gate (it may take a dramatic drop as this type of consumption goes, but return they will).  The language is reserved for only the response of ticket holders to a change in the price of their ticket.  As price falls, more tickets are purchased.

The distinction is important because confusing the two descriptive definitions leads to less clear assessment of AD choices.  For example, different ideas about what to do in the case of a decrease in demand are made to appear in conflict when they are not.  One group might argue, "Lower ticket prices"; another might argue "Make the fan experience better".  Both will get more people through the gate, but with different consequences.

The Quantity Demanded for college football attendance appears quite unresponsive to price.  In such a situation, reducing price will lead to an increase in Quantity Demanded of tickets, but will also decrease revenue.  Enhancing the fan experience will require increased spending, but at the same time will increase revenue (if successful), that is, attendance will increase even at the same ticket price, an increase in Demand.

Again, to the press and casual fellow travelers, please do everybody a favor and keep this straight.  The difference between Demand and Quantity Demanded matters for judging AD choices.