Boyd's World-> Breadcrumbs Back to Omaha-> Predictability across Different Sports About the author, Boyd Nation

Predictability across Different Sports

Publication Date: April 8, 2003

Why Are We Here And Not Over There?

There are lots of reasons why we end up as a fan of a specific sport (and, no, I'm not arguing that you're limited to just one, but most of us do tend to specialize at some point). A lot of them are aesthetic -- I loved basketball as a player, and it televises well, but give me a choice between sitting in a gym in December or sitting out in a baseball stadium in May, and it's not a hard choice. Some of them have to do with our own physical characteristics -- baseball and hockey require pretty good eyes at times to follow well. Many times it just comes down to some formative memory that builds a fire. For most folks, they don't even think about why they like a sport, it just clicks with them.

For a lot of these folks, some of this can be traced back to the predictability of a sport -- how likely am I to know the end result when the game starts? Those who like a safe outcome, with just a dash of upset thrown in to keep the mix from getting too bland, tend to become football fans. Those who like to see merit rewarded but like a good bit of unpredictability tend to become baseball fans.

For this study, I've pulled together comprehensive score data for the last five years for a number of sports, and I want to answer three questions for each of them. The first addresses the points above -- how predictable is the sport? For that, I'm looking at this question: Given two teams who differ by a given amount in quality, as measured by the ISR's, which seem to work pretty well with all the sports given here with one caveat, how likely is the weaker team to win? To look at this, I looked at the range of quality within the sport -- how wide is the range between the best team and the worst team? I also looked at the probability functions for each sport, similar to the 2% per ISR point rule of thumb that works pretty well for college baseball (in other words, a team with a 10-point ISR advantage will win 70% of the time, ignoring the home field advantage), but couldn't find a good way to present that in an understandable manner; maybe some other time.

The second question concerns the notion of competitive balance -- how likely is it that a team will be about as good one year as they were the year before? For that, I'm comparing ISR's from year to year with a correlation measure. Finally, I'm curious about how well the postseason is set up for the sport. In other words, how good are their champions? I've done this for eleven sports or variants of sports, and some of the comparisons are interesting. I'd love to add soccer, softball, or volleyball (or any other team sport you can point me to data for), but these are the only ones I've been able to find sufficient scores for yet.

College Baseball

Average Range: 62.1 - 126.5
Competitive Balance: .87
Champions: 1/7/3/4/1/286

We'll start here, with the sport we know best (for those of you who are fans of other sports and got here through Google or whatever, read this anyway) and explain the different measures.

The range is the average of the lows and the highs in absolute ISR measures for the five seasons. I don't discuss absolute ISR values very much (they're designed to balance around 100 and form a nice normal curve over the sport), but they're useful here to show the magnitudes of relative quality in each sport. The tighter the range, the closer the worst and best teams in the sport are, and the more likely an upset is in any given game. College baseball is the most competitive (or most random) of the college sports.

The competitive balance measure listed here is the result of correlating the ISR value from one year to the next for all teams that played in successive seasons in the sport. .87 means that a team that successful in one season is quite likely to be successful in the next.

On the championship line, the first five numbers represent the ISR rank of the national champion. If your goal for the postseason is to find out who the best team is, 1 is good here. If you're a fan of "the excitement of upsets", higher is better, I suppose. For all its faults, the college baseball postseason has produced some fairly good national champions, but it turns out that it's unusual for a team lower than about #3 in its sport to win a title, so Miami's 1999 championship is still off the charts. The last number on the line is the number of teams who competed in 2002.

Major League Baseball

Average Range: 93.3 - 107.9
Competitive Balance: .55
Champions: 1/6/9/4/4/30

For all the complaints about "faith and hope" (and the anti-trust exemption, which means we get to see these lies told directly to Congress), we have here the most highly competitive sport of them all. There's a moderate correlation between success from year to year (which is actually good, since you want well-run teams to continue to succeed and poorly-run teams to be forced to make changes), but in any given game, there's very little difference between the best and the worst. We also have the most poorly-designed postseason I can find.

Men's College Basketball

Average Range: 69.4 - 131.6
Competitive Balance: .79
Champions: 1/2/2/1/2/322

The range is actually closer than I expected, and the competitive balance is at least a little more prone to volatility than college baseball. For all the jokes about being a fan of, say, Vanderbilt football, it appears that there's actually little less likely to be rewarding than being a fan of a bad college baseball team (with the exception of college hockey, to be noted later).

For all the CBS chest-thumping about the unpredictability of March Madness, the postseason does a remarkable job of picking a team that's at least very close to being the best. The only team outside the top 10 in the last five years to reach the title game is Indiana in 2002; two years of the last five have featured #1 and #2 playing for the title.

Women's College Basketball

Average Range: 65.3 - 139.9
Competitive Balance: .84
Champions: 1/2/1/3/1/324

As the budgets shrink a bit from the men and the limelight fades a little, things get a bit less competitive; as much as there are fabled programs in men's basketball, there's really nothing there that compares with Tennessee or Connecticut women's basketball.

NBA

Average Range: 83.9 - 113.3
Competitive Balance: .69
Champions: 1/1/1/1/1/29

The less random nature of basketball relative to baseball helps out the NBA tremendously in the PR wars against MLB (although less so than a willingness not to repeatedly denigrate their own product). The competitive balance correlation is higher than in baseball by quite a bit, but nobody's complaining in August that certain NBA teams have no chance to make the playoffs, because they can let over half the league into the postseason and still manage to quite predictably have the league's best team win the title. Side note: Nobody does NBA pools, and gambling's illegal in this state, but take San Antonio this year, although it's closer than usual.

WNBA

Average Range: 86.4 - 117.2
Competitive Balance: .54
Champions: 1/1/1/1/1/16

Even if the nature of the sport is to encourage teams to stay good once they're good, new leagues are not a great proving ground for it, so I'm not sure there's much to be concluded here.

College Football

Average Range: 47.4 - 132.7
Competitive Balance: .64
Champions: 1/1/3/1/1/152

As a minor side note, I'm not convinced that the ISR's are the best measuring tool for college football, since there are so few observation points in a season. The wide range is due to the fact that I don't have a source for scores that excludes 1AA teams. I'm actually somewhat surprised that the top end is as low as it is; my impression (although I haven't followed football in over a decade) was that the best team almost always won. I guess the relative rarity of undefeated teams would argue against that -- relatively speaking due to the number of games, the best women's college basketball teams are actually more dominant than the best college football teams.

For all the grousing about the lack of a real playoff, it's worth noting that the best team almost always ends up as national champion; something that's not as likely under any proposed playoff system.

NFL

Average Range: 81.6 - 116.4
Competitive Balance: .29
Champions: 1/4/1/4/1/32

Here we have an interesting paradox -- a sport where, by professional standards, there's a fairly wide range between the best and worst teams in any given year, but being good one year is only a slight indicator that a team will be good the next year. It will be interesting to see if the change in the way of producing the schedule will change these numbers.

Men's College Hockey

Average Range: 42.4 - 140.4
Competitive Balance: .93
Champions: 2/2/2/2/132

I'm on thin ice here, so to speak, since I know almost nothing about hockey, but this looks like a sport with a huge gap between the haves and the havenots that's unlikely to change. There's nothing inherent in the postseason structure that I can see that would keep the #2 team winning, so that may just be one of those weird coincidences. I don't have data for 1998.

Women's College Hockey

Average Range: 54.9 - 141.6
Competitive Balance: .92
Champions: NA/NA/NA/2/2/68

Remember how little I knew about hockey? Subtract some of that for women's hockey. As far as I can tell, there's only been a national championship for two years now, so there's not even much history to look at.

NHL

Average Range: 88.4 - 109.3
Competitive Balance: .66
Champions: 2/1/4/1/1/30

We've got very competitive individual games, and a small but acceptable amount of turnover from year to year in the good teams, along with a postseason that produces a surprisingly accurate result, given the amount of randomness in individual games. So why are these guys going broke?

Pitch Count Watch

Rather than keep returning to the subject of pitch counts and pitcher usage in general too often for my main theme, I'm just going to run a standard feature down here where I point out potential problems; feel free to stop reading above this if the subject doesn't interest you. This will just be a quick listing of questionable starts that have caught my eye -- the general threshold for listing is 120 actual pitches or 130 estimated, although short rest will also get a pitcher listed if I catch it. Don't blame me; I'm just the messenger.

Date   Team   Pitcher   Opponent   IP   H   R   ER   BB   SO   AB   BF   Pitches
Mar 28 North Carolina-Charlotte Zachary Treadway St. Louis 9.0 9 3 2 5 8 34 42 129
Apr 4 Campbell Josh Blades Central Florida 7.2 12 4 4 5 4 32 38 147
Apr 4 Mercer Brandon Davidson Jacksonville 7.0 5 3 3 4 7 23 28 124
Apr 4 Samford Stephen Artz Troy State 8.0 13 5 5 3 6 35 39 147 (*)
Apr 4 Pacific Matthew Pena Cal State Fullerton 7.0 14 10 9 4 2 33 39 134
Apr 4 South Florida Jon Uhl East Carolina 8.2 8 4 4 2 9 33 36 130 (*)
Apr 4 North Carolina-Charlotte Zachary Treadway Tulane 8.1 9 6 5 3 7 33 38 122
Apr 4 Houston Brad Sullivan Alabama-Birmingham 5.2 4 3 3 6 5 19 27 124
Apr 4 Ohio Chris Bova Marshall 9.0 5 1 1 1 12 30 35 141
Apr 4 Evansville Tom Oldham Creighton 9.0 11 6 4 0 6 37 41 130 (*)
Apr 4 Alabama Taylor Tankersley Auburn 4.1 9 7 7 3 7 23 28 125
Apr 4 College of Charleston Matt Soale Davidson 9.0 4 2 2 2 8 31 33 121
Apr 4 Rice Philip Humber Hawaii 8.0 6 3 3 1 11 29 31 125
Apr 5 Missouri Garrett Broshuis Texas Tech 7.2 8 4 1 2 5 31 34 126
Apr 5 Cincinnati B. J. Borsa Southern Mississippi 9.0 7 3 2 2 8 34 38 138
Apr 5 Eastern Michigan Anthony Tomey Ball State 7.0 4 1 1 5 9 26 31 129
Apr 5 Eastern Michigan Trevor Carpenter Ball State 6.0 8 5 4 1 7 27 28 128
Apr 5 Miami, Ohio Graham Taylor Arizona 9.0 10 4 4 4 6 35 39 155
Apr 5 Ohio Novosel Marshall 8.2 8 4 4 4 10 32 38 146 (*)
Apr 5 Evansville Mitch Prout Creighton 9.0 6 3 3 5 7 30 40 144 (*)
Apr 5 Austin Peay State D. Smith Tennessee-Martin 9.0 5 3 2 3 10 31 34 130 (*)
Apr 5 Murray State Kyle Perry Eastern Kentucky 9.0 11 8 5 2 3 36 42 132
Apr 5 Alabama Johnson Auburn 7.0 5 1 1 2 9 25 27 120
Apr 5 Kentucky Heath Castle Mississippi State 9.0 7 2 2 3 5 31 34 132
Apr 6 Texas Justin Simmons Baylor 8.0 3 3 3 2 7 27 30 129
Apr 6 Cal State Northridge Leo Rosales Long Beach State 9.0 6 2 1 1 9 32 35 121
Apr 6 William and Mary Chris Shaver James Madison 8.2 9 3 3 4 5 29 35 122
Apr 6 Texas-Pan American Travis Parker Texas A&M-Corpus Christi 10.0 9 2 2 4 4 37 41 145 (*)
Apr 6 Texas A&M-Corpus Christi Jimmy Hamon Texas-Pan American 9.0 6 2 1 4 11 32 38 148 (*)
Apr 6 Yale Mike Elias Princeton 10.1 7 7 7 5 4 36 43 144 (*)
Apr 6 Evansville Trevor Stocking Creighton 8.2 8 2 2 5 6 30 38 139 (*)
Apr 6 College of Charleston Brett Harker Davidson 8.0 9 3 0 0 5 34 34 126
Apr 6 Gonzaga E. Clelland Portland 11.0 7 3 2 3 5 37 43 138
Apr 6 Hawaii Chris George Rice 6.2 7 6 6 6 5 26 32 123
Apr 7 Washington State Aaron MacKenzie Stanford 10.0 12 5 4 2 4 38 42 138 (*)
Apr 8 James Madison Leatherwood Richmond 7.1 7 3 3 5 6 28 35 130 (*)
Apr 8 Georgetown Salvitti Maryland 9.0 7 4 3 5 8 34 39 149 (*)
Apr 8 Texas-Pan American Aaron Guerra Texas 6.1 8 7 4 7 5 26 37 139
Apr 9 Southern Mississippi Cliff Russum Mississippi 9.0 4 2 2 2 10 30 32 129
Apr 10 Sacred Heart Chuck Ristano Long Island 9.0 8 0 0 0 10 33 35 122

(*) Pitch count is estimated.

The Treadway line from the 28th is a correction based on an actual pitch count.

If you're interested in reprinting this or any other Boyd's World material for your publication or Web site, please read the reprint policy and contact me

Google

Boyd's World-> Breadcrumbs Back to Omaha-> Predictability across Different Sports About the author, Boyd Nation