Tourney Time: A Historical Look at the Conferences

March Madness is upon us and I thought it would be a good time to look at what the numbers say over the history of the NCAA tournament since it was expanded to 64 teams in 1985. The first glance I will take is to compare wins and losses by the 6 power conferences and the group of teams belonging to other conferences, sometimes confusingly called mid-majors. The chart compares the winning percentage of each broken down by each particular round of the tournament. The number of appearances for each conference is shown in parantheses
.

Learn About Tableau

A few observations:

1. Clicking on a conference in the color legend will highlight the results for the chosen conference. In general, winning percentage decreases as teams get deeper in the tournament, which is not surprising. There are a couple exceptions to this trend. For example, the Big 10 takes a big jump from the Sweet 16 to the Elite 8 and the SEC jumps to the head of the pack in the Final 4 and Championship rounds.

2. The power conferences have consistently outperformed the others in the First Round of the tournament, but the other conferences catch up to a degree as they get deeper, winning 6 of their 13 appearances in the Final Four.

3. Of the six power conferences, the Pac-12 is the worst performer in terms of both games played(252) and overall winning percentage(.567). A true comparison between the power conferences should probably take into consideration the number of bid opportunities, since there are differences in the number of teams in each conference through the years.

4. The ACC separates itself from the pack when comparing overall winning percentage, but that is mainly due to their high rate of success in the early rounds. They fall behind the Big 12 once they reach the Sweet 16.

5. The Big 12(which also includes teams from the SWC and Big 8 in the early years) takes the biggest jump from one round to the next, leading the pack in the Sweet 16(.604) and falling to the bottom(.394) in the next round.

Note: I ignored results from the Play-In rounds, which were just recently added to the tournament.

I'll take a look at head-to-head matchups in my next post.

The NFL Moneyball Dashboard

With their victory in the Super Bowl XLVII, the Baltimore Ravens moved ahead of the Indianapolis Colts and Pittsburgh Steelers and inched closer to the New Engalnd Patriots on the updated version of the NFL Moneyball Dashboard. 

The Moneyball Dashboard compares the success of each NFL franchise since the year 2000 by measuring their performance in the post-season against their average payroll.  Teams are awarded 1 point for qualifying for the playoffs and 1 point for each post-season win, so a Super Bowl win is currently worth 5 points. Teams landing in the upper left quadrant can be considered over-achievers as they have exceeded the NFL average for playoff success while their average payroll falls below the NFL average. On the flip side, teams in the lower right quadrant could be considered under-achievers. The size of the circle for each team helps to compare their regular season success by measuring their wins per season for each million dollars of payroll.  Use the power of Tableau to compare teams by selecting them from the bubble chart or using the filters to the right.

Powered by Tableau

A few observations:

1. Since the Houston Texans only joined the NFL in 2002, I think it is a little unfair to label them under-achievers.  The Texans have finished .500 or above 5 of the last 6 years and have won a playoff game each of the last 2 years. Even though they have the biggest payroll in the NFL, they are surely headed in the right direction.

2. Due to a lack of franchise payroll data, the dashboard only reaches back to the year 2000.  The New Engalnd Patriots benefit the most from this limitation having started their incredible run of post-season success in 2001. An analysis that includes measures dating back to Super Bowl I would perhaps provide a more fair comparison.

3. What do the post-season numbers say about competitive balance? About half of the teams find themselves clustrered between 6 and 15 playoff points.  Seven teams(NE, BAL, PIT, IND, PHI, GB, NYG) exceed 19 points and could be called high achievers.  Eight teams,excluding Houston(KC, JAX, CIN, MIA, WAS, CLE, DET, BUF) have yet to earn 5 playoff points and only 1 team, Buffalo, has not gained a post-season berth since 2000. Without any deeper statistical analysis, I think we could say these numbers support the idea that the NFL has achieved a fairly good level of competitive balance.

4. How does post-season success corrrelate with payroll? From a quick glance at the distribution, there is little or no correlation. While it is true that most of the poor performers have payrolls below the NFL average, so do the payrolls of the two most successful teams, the Patriots and Ravens. Interesting note: The best-fit line for just the NFC teams actually has a negative slope, suggesting that the cheaper teams actually do better. Either way, there is really no evidence of correlation.

Would love to hear suggestions on further analysis or ways I could improve the dashboard.

Paul

Why Aren't People Tuning in to the MLB Playoffs?

According to the Neilsen Sunday night televsion ratings, the NFL game dominated drawing a 6.9 rating in comparison to a 1.9 rating for the NLCS game on Fox.  This should come as no surprise as the NFL has consistenly clobbered MLB over the years. But there is a bigger issue at play in Major League Baseball, that gets amplified during the playoffs and World Series: the length of the games and the sheer amount of dead-time continue to rise and it makes the games more difficult to watch.

The average length of a regular season MLB game has been on the rise from an average length of 150 minutes in the 1970's to the current average of about 175. Thus far, the average length of a playoff game in 2012 has been 206 minutes. A closer look at the American League numbers show that there were three extra inning games lasting 12, 12 and 13 innings respectively.  Removing them from the calculation completely drops the average down to 199 minutes, so that is still an increase of 24 minutes compared to a regular season game.  This may not seem like a lot but considering the primetime games typically start after 8pm ET, viewers in the Eastern time zone will typically need to stay up past 11:30 to get a result. Not so easy, especially for the early risers.

There may be a bigger reason that viewers may be tuning out: the amount of dead-time during a game.  The first game of the ALCS between the Yankees and the Tigers lasted 294 minutes, 6 minutes shy of 5 hours. In 12 innings, 14 pitchers were sent out to the mound--that's 12 pitching changes with at least one or more coaching visit, a walk in from the bullpen and the requisite number of warmup pitches. An astonishing total of 425 pitches were thrown during the marathon, which brings us to the next source of dead-time: the deliberate approach taken by both pitchers and batters. I don't have stats to support it, but there seems to be agreement that at bats are extended during the postseason because they "matter more."  Batters are said to be grinding it out and are taking pitchers deeper into the count.  If this was the case, we would expect to see the number of pitches per inning to be much higher in the postseason compared to the regular season.  However, as the chart below shows, that is not the case:

	Pitches/Inning
Regular Season(MLB)	32.5
Postseason(NL)	30.1
Postseason(AL)	33.4

So it may be safe to blame the longer games on a combination of the following:

• extended time between innings
• more pitching changes
• more delays added by pitchers stepping off the rubber and hitters being granted time out
• more pickoff attempts

Some of these could certainly be addressed by MLB. Whatever the case, the viewers have spoken:  Some postseason games are painfully long and the lack of action only makes it harder to stay tuned in. 

Experience and Clutch Hitting Become the Difference

MLB added an additional wild card starting this year to give more teams an opportunity to take part in the playoff race. But after the dust cleared and the tears were wiped from the eyes of the distraught in Cincinnati and DC, we are left with four clubs who are familiar with success--together they have appeared in 10 World Series since 2000 and include the winners of the last two. So, if the purpose of the new format was to help with competitive balance, it really didn't accomplish its purpose. 

If the playoffs could be summed up in a couple words it would be experience and clutch hitting. Veterans CC Sabathia and Justin Verlander provided the hammer in the rubber games for the Yanks and Tigers. The Cardinals used their experience and got some unlikely clutch hitting contributions to dig out of a 6-0 hole. The Reds fell behind by the same margin in their Game 5 but just couldn't get that last clutch hit. One consolation for all the losers: they are all relatively young and perhaps another year of experience will prepare them to get over the hump next year...pitchers and catchers report in 4 months.

Take a look at the teams remaining on the Playoff Dashboard.

Comparing the 2012 MLB World Series Contenders

Powered by Tableau

The graph above compares the 2012 MLB playoff teams by plotting runs scored versus runs allowed per game. The 2012 Major League average is 4.33 RPG. Also included are the average for all the playoff teams since 1994, when divisional series were added to the playoffs. A few observations:

1. There is a wide contrast between the clubs, with the Yankees scoring almost 5 runs per game(2nd to Texas) and the Reds giving up only 3.63 runs(tops in MLB).
2. The numbers for the Washington Nationals, with the highest run differential of almost 0.9 RPG, support the fact that they finished with best record in baseball.
3. Teams from the American League give up more RPG than their National League counterparts, which reflects the overall difference(4.45 to 4.22) between the AL and NL.
4. The Baltimore Orioles are spot on the Major League average in both runs scored and allowed which is testimony to their gaudy 29-9 record in one-run contests.
5. Comparing the 2012 numbers to the averages for the playoff teams since 1994 is just more evidence that 2012 was indeed the year of the pitcher.

So which type of team will come away with a World Series title? An offense minded team like the Yankees or a team focused on pitching and defense like the Reds or Nationals? Or will it be a more balanced team like Detroit or the defending champion Cardinals? Would love to hear your observations and comments.

The MLB Moneyball Dashboard

Powered by Tableau

The interactive dashboard above contains a plot of the ROI for Major League teams since the year 2000. Instead of focusing on wins per dollar, it measures the post-season success of each franchise by assigning 1 point for a playoff appearance, 2 for a division series win, 3 for a league championship and 4 for winning the World Series. The size of the team's circle is used to indicate the average wins for each million dollars of payroll. Winning percentages and post season results for each team are shown in the bar graphs.

• The St. Louis Cardinals are clearly the biggest over-achievers, with 9 playoff appearances and 2 World Series championships all with a modest payroll half of that of the New York Yankees.
• The three NL clubs with the largest payrolls, the Mets, Dodgers and Cubs are the biggest under-achievers.
• Only 3 clubs--Pirates, Royals and Blue Jays--have not joined the post-season party, with the Jays having the highest payroll of the group.
• The Florida(now Miami) Marlins have the highest number of wins per dollar along with the lowest average payroll of any club. They are somewhat of an anomaly in that they won a World Series in 2003 but had no other playoff appearances.

Competitive Balance: Part One

One of the more discussed issues in sports is the concept of competitive balance or parity.  The traditional argument is that parity is good for sports because it keeps more fans interested knowing that the team they root for has a fair chance of winning it all, or at least qualifying for the post-season.  Given the choice, most fans of MLB, the NFL or NBA would prefer a season where all the teams had the same chance of winning a championship versus one where there were only a few teams that had a realistic chance.  Furthermore, having the same teams winning year after year is not considered good for the sport by most fans except for those who root for these so-called dynasties. 

So how can we measure competitive balance?  One of the ways is to look at the distribution of championships by franchise.  In order to get a good representative sample, let's look at MLB, NFL and NBA champions over the last 40 years.  The graph below shows a  frequency distribution of championships by franchise from 1971 to current.  A sport that was completely balanced would have all franchises winning at least one championship with a few winning two.  Clearly, this is not the case.  In fact, about half of the current NFL and NBA franchises have not won the big prize.  The most useful measure for comparing the "spread" of the numbers is the standard deviation.  Put simply, the standard deviation measures the difference of a set of numbers compared with the mean.  In order to compare two or more sets of numbers, it is sometimes appropriate to divide the standard deviation by the mean.  However, in our case, the difference between the means for each of the sports is negligible. Here are the standard deviations for the three sports; the lower the value, the more the sport is "balanced": 

Major League Baseball: 1.5   

      National Football League: 1.7           

          National Basketball Association: 2.5             

                     

So, comparing the number of championships, we see that MLB and the NFL are similarly balanced and are significantly more balanced than the NBA.  It may be surprising that MLB leads the way, especially since they have not established some of the restrictions(hard salary cap, etc.) that the other sports have to improve parity.

Of course, there are other ways to measure competitive balance that might be more revealing.  We'll look at others next.

Powered by Tableau