Commentary

Having fun with blind résumés!

Originally Published: March 6, 2012
By Eamonn Brennan | ESPN.com

It's easy to mistake numbers for science.

The NCAA tournament selection process -- in particular, this final week of agony before the conference tournaments crown their champions and Selection Sunday makes its glorious return -- is inherently awash in numbers. The Ratings Percentage Index is the big one, the figure that structures and underwrites everything the committee does in the room between now and Sunday (whether the NCAA always likes to admit it or not).

But as the RPI is sliced and diced to create other categories -- from strength of schedule to top-50 wins to bad losses to conference strength to, well, you name it -- it can give a casual fan the general impression that tournament selection is a matter of rigorous data collection and analysis. Add in the selection committee's usual excellence in selecting the field, the typical predictive success of bracketologists like our own Joe Lunardi (and countless others) and the forgiving nature of the 68-team bracket, and it's easy to think the committee constructs the tournament in academic fashion. Team X meets A criteria? They're in. Team Y doesn't? They're out!

To the contrary: Tournament selection is much more like art. There is no golden ratio or Fibonacci Sequence to apply here, no one set of obvious and elegant rules that governs what the committee chooses to do when it gets down to the last three or four of the 37 at-large selections. This is partly because the committee uses a flawed metric in its deliberations (the RPI is outdated and mostly dumb but, well, let's not go down that road again). But it's mostly because the committee comprises living, breathing humans, all with their own preferences and tendencies and post-lunch food comas and specious arguments and powers of persuasion. The RPI can make the committee room seem like a distant, calculating being unto itself, a hoops HAL 9000. In reality, it's more like "12 Angry Men."

And each committee's members are different. Some may value their own "eye test" more than they should. Some may rely solely on the RPI. Some may be inherently inclined to give a middling high-major team -- as opposed to a shaky mid-major with a gaudy record -- the benefit of the doubt. Or vice versa.

In the big scheme, this doesn't add up to much: With 68 teams in the field, and only one champion to crown, the chances the committee's varied whims will change the course of the tournament are always slim. But as VCU showed us in 2011 -- and George Mason proved in 2006 -- the difference between an historic Final Four run and a trip to the NIT can, every once in a while, come down to some seemingly indistinguishable and confusing differences.

How will the committee weigh such factors this year? Which criteria, if any, will come to define the teams that either make or miss the 2012 NCAA tournament? We don't know. But for fun -- and to highlight some of the contrasts the committee will be examining as it convenes in Indianapolis this week -- we figured we'd have a little fun with a timeless bubble endeavor. You know what it is: blind résumé time!

In each of the three charts below, we compare a handful of current bubble teams side-by-side based on their RPIs, strength of schedule numbers, records in road/neutral court environments, and records against teams ranked in the top 50 and top 100. These are, for better or worse, the big ones -- the data points the committee most frequently cites as its most important. As you'll see, distinguishing between the "best" and "worst" résumés can be a frustratingly opaque, inherently subjective, ordeal.


In case you needed proof of why the RPI can be so deceptive, well, here you go. To look at either Team A or Team D's résumés, you'd assume that not only were those teams in the tournament, but that they'd long since sewn up their bids and were playing these past few weeks for nothing more than conference titles, pride and/or a bump or two in projected seed placement.

Not so much, especially in Team D's case. Frequent bubble watchers (or anyone who spoiled the fun by mousing over the team names already) can probably guess who the teams are, for they are the best examples of how, thanks to smart scheduling and statistical quirks, a team's RPI numbers can become inflated and gaudy despite otherwise mediocre overall profiles.

Meanwhile, Teams B and C have RPI figures that typically bode well for selection in the tournament: In 2011, the first year in the new 68-team format, only three teams (Saint Mary's, Missouri State and Cleveland State) with an RPI better than 50 missed out on the tournament. In fact, that number was the same in 2010, before the mini-expansion went into effect, back when we were all still freaking out about the chance the NCAA would disastrously expand the tournament field to 96.

The point is, a raw RPI number goes only so far. Teams B and C have played 15 top-50 games between them, earning just one win for their combined efforts. That's the kind of tally that might give any committee member pause about placing either team in the field. Comparatively speaking, Team B's win totals are drastically bad. One win in 11 games against the top 50? Just five wins in 17 tries against the RPI top 100? The strength of schedule figure helps out -- per the computer's calculations, this is clearly a team that faced a rigorous 2011-12 season -- but what else on that profile is there to feel good about, exactly?


How much does the committee value strength of schedule? We know this much: In recent years, the committee has consistently placed an emphasis on nonconference strength of schedule. Sometimes, it seems to matter as much that a team makes the effort -- that it goes out of its own building and its own backyard and plays quality teams in November and December -- as whether it wins or loses said nonconference games, or even whether it clearly improves down the stretch in conference play.

We saw as much in 2011, when Colorado and Alabama, both of which scheduled (and played) horribly in the nonconference, saw their solid wins in league play (Bama finished 12-4 in the SEC, no less) go to waste on Selection Sunday.

That's why Team A, which has looked uninspired and feckless and mostly just ugly for the past two months of the season, has never really flirted with true bubble trouble: because the RPI formula (somewhat inexplicably) loved its nonconference schedule. Team B, meanwhile, went 12-5 on the road and is 27-6 overall, winning its conference at 16-2. In a vacuum, or at least at first glance, that's a tournament team, right? Maybe not: Team B didn't really play (or beat) anyone of note before conference play commenced, and its league rivals unhelpfully followed suit, leaving it with strength of schedule numbers that cast doubt on the reasons that gaudy record exists in the first place.

Team D has a worst-of-both-worlds feel to it: It went 22-7 and 12-4 in its league, and it played (and beat) a couple of top-50 nonconference opponents, but its league wasn't good enough overall to prop up its strength of schedule figure, and the schedule itself isn't good enough to justify such a middling mark against the top 100.

But at least Teams B and D have showed themselves capable, albeit in small doses, of keeping pace among the RPI top 50 and top 100 opponents. Team C has not. The schedule is good, but that 4-10 mark against the top 100 is truly dispiriting. What good is a difficult schedule if you don't actually beat anyone on it?


You often hear amateur (and professional!) basketball commentators say that NCAA tournament selection should be about nothing more than "who you beat." (Note to said commentators: It's "whom," not "who." You're slowly driving my English-teaching mother insane. Please stop.) As you can see here, when you get down to the bubble cut line, the better question to ask is whether you actually beat anyone. For example, see Team D.

The RPI is prohibitively high, and the nonconference SOS figure sure isn't pretty. Neither is the 3-10 road/neutral split, for that matter. But Team D has the most top-50 wins of any team in this group, all four of which came against teams ranked No. 34 or higher in the RPI. Team D has also won eight of its past nine games.

Is that enough to warrant inclusion? What about the other dings on the profile? And what about Team A? That team -- and I'm giving a major hint away here, but whatever, you already scrolled your mouse over the name anyway -- just so happened to go 12-6 in one of the three best leagues in the country. Does that count for something? The committee says no, insisting that it eschews conference record (and conference RPI) in its process. But Team A also got five wins among the top 100. Clearly, this team can beat decent opponents. But if it can't take down any of the 50 "best" in the country, should it get a spot in the Dance?

And, lastly, does strength of schedule matter in how the committee views various top-50/top-100 records? Probably not. A top-50 team is a top-50 team, after all. But not all top-50 teams are created equal. Some of those games are on the road. Some are at home. Some are against the nation's elite. Some are against mediocre teams with inflated RPIs.

Should Team C's brutal mark in top-50 games be hedged against its schedule strength? Should Team B's solid-but-unspectacular computer numbers be considered in the context of its schedule? And what about when those wins happen? If they came all the way back in November, before a certain season-changing event (cough), do they still reflect this team's current ability?

So many questions. So few definitive answers.

Such is the curse of NCAA tournament selection. Three-fourths of the tournament is simple (or, in the case of 31 automatic qualifiers, mindless) to select. Most of the 2012 field's teams have had their names written in NCAA-sanctioned ink for weeks now. That's the easy part.

But after the committee gets its locks in order, after it adjusts for the impending rush of automatic qualifiers, it must get down to the business of distinguishing between 12-15 résumés that are frequently indistinguishable in their mediocrity. This season's bubble is no different. Even worse, the who-gets-in and who-gets-snubbed portions of the bracket are both the most difficult and disproportionately scrutinized portions of the process. If you dig around long enough, you can find a reason to include any one of the teams listed above. Just as easily, you can find a reason to leave them behind.

At this level, there is no magic selection bullet, no obvious criteria to be fulfilled, no threshold that must be crossed. At this level, there is no science to NCAA tournament selection. For as much as we try to boil these teams down to groups of numbers, whether they end up in the tournament is a matter of open and varied interpretation.

Your mileage may vary. But the season is long. If after four months you still leave your fate in the hands of the committee -- if you leave your tournament credentials open to the interpretation of a small group of human beings stuffed in a hotel conference room in Indianapolis -- then you can't complain when things don't go your way.

In the end, the most we can do is pray that 2012's team of destiny -- 2011 VCU or 2006 George Mason -- isn't cast aside, just one more casualty to bubble subjectivity. To hope for anything more would be a request to the committee that it divine the future.

As today's exercise hopefully pointed out: Those people have enough on their plates already.

Eamonn Brennan covers college basketball for ESPN.com.