The Pi-Rate Ratings

January 16, 2022

Refining “Spurtability” In College Basketball

The PiRate Ratings have been big fans of CBS Sports analyst Clark Kellogg, ever since he set records in high school in Cleveland in the 1970’s and earned a prestigious spot on the McDonald’s High School All-America Team.

In the early 1990’s, Kellogg coined the term “spurtability” to signify a basketball team’s ability to experience a big scoring run. He proposed that in the NCAA Tournament, in a game where two teams appeared to be evenly matched, the two that could go on one big scoring run was the team that would almost always win these tense do or die games.

Being old enough to remember the great UCLA Bruins teams coached by John Wooden from 1964 to 1975, his teams routinely enjoyed these scoring runs. Wooden’s first national champion, his 30-0 1964 team with no starter over 6 foot 5 enjoyed a spurt in all 30 games. His vaunted 2-2-1 zone press defense with an occasional 1-2-1-1 thrown in combined with a high post offense that placed offensive rebounders in optimum spots produced these spurts. The ultimate of these famous spurts happened in the National Championship Game against Duke. Even though the Bruins were 29-0 at this point, Duke was the favorite in this game. They had two starters that were 6 foot 10, an all-American wing in Jeff Mullins, and a future NBA star off the bench in Jack Marin.

At the start of the game, Duke routinely beat the UCLA press and with their bigger players, they controlled the boards. They led 30-27 late in the first half when Wooden inserted key reserve Kenny Washington into the game. The spurt started immediately thereafter. Two Duke possessions where the Blue Devils beat the press resulted in hurried shots that missed. UCLA retrieved the rebounds, ran the fast break, and scored quickly. Before Duke Coach Vic Bubas could call a timeout, the Bruins had run off nine quick points in less than a minute to lead 36-30. After the timeout, Duke’s guards felt the screws being tightened. They couldn’t get the ball across the 10-second line, and multiple turnovers led to seven more quick points for the Bruins. The 16-0 run took just two minutes, and the Bruins led comfortably 43-30. The game was never in doubt after that.

Spurtability is rarely that obvious. In a 40-minute basketball game, there are going to be multiple runs, usually by both teams. One team may enjoy a 12-2 run to take a 10-point lead, and then the other team may go on a 14-4 run to tie the score. However, in the Big Dance, the winning team will almost always be the one that had one or more spurts than the losing team.

What makes a spurt happen? In about 5-10% of the cases, it is simply a matter of one team coming down the floor five to seven times and hitting their first shot on the possession, while the other team misses their single shots on their possessions. In other words, this is a rarity. What usually happens to cause a spurt is that one team either controls the glass and gets multiple opportunities to score on their possessions, or one team forces numerous turnovers and scores on the resulting numbers’ advantages, or a combination of both. Just like UCLA in 1964, if the spurting team limits the opponent to one shot on multiple possessions, while they have two, three, and even four opportunities to score at their offensive end, the scoreboard is going to move in their favor. If the other team doesn’t even get to take a shot on their possession, they obviously cannot score. And, when a team commits a turnover, if that turnover is a steal by the other team, they usually give up a lot higher rate of points on that possession, as a steal almost always produces an immediate numbers’ advantage for a fast break score.

The big question for fans watching, maybe with a financial stake in the game of some variety, is how can spurtability be estimated? The simple answer is to look at the teams that do best in the components that create these spurts. If great rebounding and being able to force turnovers, especially by stealing the ball lead to these spurts, then it obviously means that the teams that can best rebound and force these turnovers, especially steals, while avoiding the same are the ones most likely to enjoy these spurts.

The PiRate Ratings first used the stat “R+T” in the early 2000’s. It was a simple formula that attempted to calculate the average number of spurtability points a team had. We then calculated the R+T ratings of all 68 teams in the NCAA Tournament. What we found was that the teams with the highest R+T ratings that were also members of the top conferences, where their schedule strengths were also high, were the teams advancing deep into the tournaments. Additionally, the top two teams by R+T rating and from power conferences continually made the Elite 8. The top overall R+T Power Conference team almost always made the Final Four, and multiple times, they cut the nets down while “One Shining Moment” played on the TV.

Another surprise came to us when we first started calculating these R+T ratings. Annually, a small handful of teams from power conferences entered the NCAA Tournament with very low R+T ratings or even negative R+T ratings. The real surprise is these teams quickly lost in the first or second round, even if they were #3, 4, or 5 seeds playing as the favorite. Two schools in this era, Georgetown and Vanderbilt, made the Field of 68 as rather high seeds more than once when their R+T ratings were at or near the bottom of the field. Georgetown lost twice as a heavy favorite, including as a #2 seed when #15 seed Florida Gulf Coast ran the Hoyas off the floor with two spurts (in a game where we predicted an FGCU win and possibly by double digits.) In Vanderbilt’s case, they won a lot of games during the regular season by playing smart basketball, winning by getting higher percentage shots than their opponents and beating them at the foul line. It allowed the Commodores to make the field three times in this period where they were a #4 or #5 seed where they lost their first game as a favorite. In all three cases, Vanderbilt had either a negative or barely positive R+T rating, while the Mid-Major underdog enjoyed a much higher R+T rating. In all three upset losses, there was a large discrepancy in at least one part of the R+T rating that swung the game in the underdog teams’ favor. In a 2008 first round blowout loss to #13 seed Siena, the Saints pressed Vandy out of the gym early, and 10 Siena steals led to a 21-point drubbing. In 2010, Murray State dominated the offensive glass, while Vandy’s passive defense played it close to the vest and picked up a couple steals and a minimum number of forced turnovers. In 2011, with three future NBA players on the roster, Vandy lost to Richmond. Even though the Commodores shot 50% from the field and hit 46.2% of their three-point shots, they only forced two turnovers, as Richmond received a dozen extra scoring opportunities.

In the ensuing years, the R+T rating continued to be our secret weapon in picking NCAA Tournament brackets. Although this site began as a football computer rating site, it was March Madness that brought us the recognition when the New York Times noticed our little secret and linked to us one day in the latter part of the first decade of this century. The next year, our R+T rating was publicized throughout the Bracket-picking URLs. When it proved successful, we picked up a couple hundred new subscribers in just a matter of days.

Then, something happened. The onslaught of basketball analytics and the tweaking of the shot clock from 35 seconds to 30 seconds totally changed the game. Just like baseball became strictly a “play for the 3-run homer” sport, basketball became a “dominate the Four Factors” sport. Overnight, almost every team began to play the same exact style of basketball from the Division 3 ranks of college to the NBA. Either get a good 3-point shot or a very high percentage 2-point shot became everybody’s offense and the reverse became everybody’s defense. This tweaking of the game altered the way scoring spurts happen. Now, a team could hit three consecutive 3-pointers and go on a big run. Or, negatively, a team could go ice cold shooting 3-pointers and not score for several possessions.

Because field goal efficiency and defensive field goal efficiency became the be all and end all that decides game outcomes, the game basically broke down into a chess match of getting the most efficiency in shot selection. However, we noticed a similarity between baseball metrics and basketball metrics as it applied to the postseason. Famous baseball metric GM of Money Ball fame, Oakland A’s GM Billy Beane, noted this many years ago, when he said, “My stuff doesn’t work in the playoffs.” We substituted the word “stuff” for the four-letter word he really used, but he hit the nail on the head. In the playoffs, Money Ball stuff did not work. There was a good reason for this. In the regular season, each team in a league plays all the other teams, both the best and the worst teams. They use a five-man pitching rotations. Over the course of 27 weeks of action, teams will face great pitchers, good pitchers, average pitchers, below average pitchers, and weak pitchers. Being highly selective with pitches and playing for 3-run homers works against 70% of the pitchers faced before October. But, once the postseason starts, teams can get by with as little as three starting pitchers and use their top two or three relievers for all the important innings. Now, playing for the three-run homer or waiting for a specific pitch may never develop. When the #4 starter on the 95-game losing team takes the mound, he might groove one pitch per batter faced. When Max Scherzer or Sandy Koufax is on the mound, if the opponent cannot manufacture runs with minimal base runners, they are likely to fail.

The same effect of great pitchers in the playoffs can be linked to great basketball teams in the NCAA Tournament. Among the 32 or so Power Conference teams in the Big Dance, it will require more than shooting efficiency and defensive shooting efficiency to win tournament games. Please read this prior sentence carefully. There is a difference between shooting efficiency and scoring efficiency. Shooting efficiency measures points scored per shot taken. Scoring efficiency measures points scored per possession. If on a possession, a team turns the ball over and doesn’t take a shot, it has not affected shooting efficiency, but it has stopped all chances to score on that possession, thus lowering scoring efficiency. If a team takes a shot and misses, and then gets an offensive rebound and then misses again, and then gets another offensive rebound, and then the defense fouls, and then the foul shooter makes both free throw attempts, the team’s scoring efficiency goes up, while their shooting efficiency goes down.

The R+T Rating perfectly bridges this gap between shooting efficiency and scoring efficiency. However, if we are looking at efficiency, then the R+T Rating has to be alrered. Here’s why. The original R+T Rating is:

(R * 2) + (S * 0.5) + (6 – Opponents’ S) + T

R = rebounding margin

S = steals per game

T = turnover margin

As you can see, the components are counting stats. You count the number of rebounds, steals, and turnovers, and you have the components for the formula. However, the Four Factors are rate stats, where a ratio of stats are used. If Team A outrebounds its opponent 32 to 25, it has 7 more rebounds. If Team B outrebounds its opponent 44 to 36, it has 8 more rebounds. Team A has a rebound margin weaker than Team B, but Team A rebounded 56.1% of the missed shots, while Team B rebounded 55% of the missed shots. Team A actually did a little better. Thus, rates of these stats are more accurate than merely counting the differences.

Thus, a new R+T Rating using rates is called for. We actually devised this new formula in February of 2020 with the plan to release this new data in our annual “Bracketnomics” report the day after Selection Sunday. But, the entire tournament was wiped out by Covid-19, and the release and the new data was mothballed. Last year, we were trying to release our tabletop baseball strategy game, “Sabertooth Baseball,” and in a hurry to gave the game ready for sale before the 2021 MLB season, we didn’t devote as much time to March Madness as in the past. This year, we are well out in front of the action and have the time.

Here is the new and improved R+T Rating using rates instead of counting stats.

((R * 8) + ((S + T) * 4)) / 3.5

This formula now refers to Rate Stats.  The “R” in the formula now stands for Rebounding Rate.  This is a combination of both offensive and defensive rebounding rate and it is a deviation from the norm and not just a percentage.  The norm in our experiment will vary some from year to year. At the current time, it is 28.3%.  If a team has an offensive rebounding rate above this number, it is above average, and if it is below this number, it is below average.  Thus, the norm for defensive rebounding rate is the opposite of the above number, or 71.7%.  We then calculate our R part of the formula by taking each team’s offensive rate minus 28.3 plus their defensive rate minus 71.7 and then add the two results and divide by 2. The result goes in the “R” spot in the above formula.

Let’s look at a real team as an example. Kentucky is killing it on the boards thanks to a once in a generation glass-cleaner in Oscar Tshiebwe. Kentucky’s current offensive rebound rate is 39.3%, and their defensive rebound rate is 20.5%. The differences from the norms (28.3% O and 71.7% D) are 11.0 (off) and 7.8 (def). Adding the two and dividing by two for the average, gets you a result of 9.4. 9.4 would go into the “R” part of the formula.

The current constant for Steals is 9.7, which is the same for offensive and defensive steal rates. The current constant for Turnovers is 16.3, which also is the same for offensive and defensive turnover rates.

Using the same method we used to calculate R, let’s look at Kentucky’s S and T rates. The Wildcats currently have a steal rate of 9.8%, while their opponents’ steal rate is 9.0%. Kentucky doesn’t excel in this rate, getting just 0.1 for its steals and 0.7 for avoiding steals. Add the two and divide by 2, and you get 0.4 for S, not much.

For the T part of the formula, Kentucky’s offensive turnover rate is 13.9, and their defensive turnover rate is 16.6. Remember on Turnover rates that on offense, the lower the number the better, so a lower number than the constant is positive and a higher number than the constant is negative. The constant as of today is 16.3. Kentucky’s offensive turnover rate is 2.4%, and their defensive turnover rate is 0.3%. Add the two and divide by two, and the result is 1.35.

We now have all the numbers we need to plug into the new R+T formula.

R = 9.4

S = 0.4

T = 1.35

((9.4 * 8) + ((0.4 + 1.35) * 4)) / 3.5

Result: 23.5

23.5 is a high number. Kentucky will enjoy nice spurts against just about any opponent. We can also use this criteria to see where they are vulnerable. If a team is strong in the S & T parts of the equation, they could exploit the Wildcats, and erase the rebounding advantage Kentucky will have.

Here’s another example. Houston made the Final Four last year by dominating in all phases of the game. Kelvin Sampson’s teams have typically been great at the “hustle stats”, and that correlates to a high R+T rate. Here are UH’s stats as of today.

R = 5.5

S = 2.65

T = 3.75

((5.5 * 8) + ((2.65 + 3.75) * 4)) / 3.5

Result: 19.9

Once again, Houston has an excellent R+T rate. They excel in all three components of the formula and while not as strong overall as Kentucky, they have no weakness here.

Now, let’s look at a pretender. Loyola of Chicago was the darling long shot when they snuck into the Final Four a few years ago. This season, the Ramblers look like a stronger team playing in a tougher conference. At 13-2, they are nationally ranked. However, their R+T rate is well below the threshold to be a serious contender again.

R = 1.3

S = 0.95

T = 1.6

((1.3 * 8) + ((0.95 + 1.6) * 4)) / 3.5

Result: 5.9

An R+T of 5.9 might be just enough to beat a weaker team in the Round of 64, but if that opponent has a higher R+T rate, Loyola will be in trouble. If, by chance, Abilene Chrisitan was to gain the WAC’s automatic bid with their 94 feet of pressure defense and unorthodox playing style, ACU’s superior R+T rate (16.6) might more than make up for Loyola’s stronger schedule strength and their patient style of play.

Don’t fret if you don’t totally understand all the stats. We won’t leave it to you to compile these numbers on Selection Sunday. The PiRate captain will huddle in his quarters and calculate all of this data as well as many other factors used to pick the bracket.

Blog at WordPress.com.