The War on Stats (Part One)

Derek Zumsteg

[Note: This article is a substantial re-working of "Sabermetrics and +175," which originally appeared on Rob Neyer's website as a Guest Shot at http://www.robneyer.com/guests.htm, and has been corrected, revised, and shortened for publication here]

Rotisserie baseball has gone from being viewed as something not far from pen-and-paper role-playing games like Dungeons and Dragons to being a valuable demographic that gets its own segment, however brief, on SportsCenter. Fantasy owners have created a huge market for deep, accurate player and team information, delivered in timely fashion, giving rise to some of the best coverage the sport has ever seen.

There's a contingent of baseball media figures and others who regard fantasy baseball as something not only useless, but bad. They say that fantasy players see only numbers, that they don't know about the intangibles, the little things that win, that their view of the sport is limited by their obsession with their ranking in the RBI category. The distaste for fantasy baseball has become a club with which to beat serious statistical analysts, in an attempt to discrediting not only their results, but also their very existence, as being counter-productive to the enjoyment and tradition of baseball.

However, this sniping and bitterness is petty compared to what's to come. There is a huge market out there, far larger than fantasy baseball will ever be, that is going to cause a tremendous, bitter battle for the future of the sport between baseball, sports media, and the intelligent fans ... and the fans will lose.

The $380 Billion Gorilla

Sports gambling is so large and powerful that even defining it is pointless, because it has seeped into everything. Professional sports go to great lengths to keep gambling at a distance, but in the end the war cannot be won, only fought.

I've read estimates that the gambling industry takes in anywhere between $30-60 billion per year in revenue. Now that's for all forms of legal gambling, from my brother hustling people at Texas Hold 'Em to slot zombies, but it doesn't count the guys taking bets in local sports bars, claiming that a 'push' is actually a house win, and the growing internet gambling market.

No one really knows how large sports gambling is, but Nevada Gaming Commission Chairman Steve DuCharme said last year that "legal gambling brings in less than $2 billion per year, but illegal betting dwarfs that, with upwards of $380 billion per year." (Digital Collegian, April 8, 2000). In 1998, it was estimated that about $2.3 billion dollars were bet on sports in the Las Vegas sports books alone. (Las Vegas Review-Journal, Jan 05, 2001).

Of that, not much is bet on baseball. Twenty-five or thirty percent is bet on college sports (LVRJ, ibid). Outside of that, most sports gambling is on football, followed by basketball, with baseball a distant third among pro sports. But it's the only game to bet on for much of the year, so let's figure that it's only 10 percent of all bets, which is quite likely wildly inaccurate, but bear with me.

Let's figure the good Chairman was right, and there's a total of at least $382B wagered total a year. Even if you believe that baseball betting is only one percent of all the money wagered, there's $3.8 billion wagered a year. At 10 percent, $38 billion dollars. At 20 percent, $7.6 billion.

Major League Baseball says that total team revenue from all sources comes to about $3 billion dollars (Blue Ribbon Panel report), though of course that's just what they fessed up to. I've been unable to separate ESPN's revenues from ABC, Disney's cable unit, and other Aaron Carter-related business (though Disney's cable holdings in total booked $9 billion in total revenue [Walt Disney Company Fact Book 2000]). Then there's FOX Sports, one of Rupert Murdoch's vast cable holdings which netted him $2 billion in total revenue (NewsCorp 2000 Annual Report), so there's obviously a great deal of money in traditional sports media and in traditional sports coverage.

Even one of the internet's worst sports sites, CBS Sportsline.com, booked $95 million in revenues last year. This defines where the next battle is going to be fought: on television and in the papers, in smoky sports-book rooms, in the offices of billion-dollar casinos, anonymous servers on racks in the Dutch Antilles, and in Congress. All of the players are amazingly well funded and employ many of the best lawyers in the country, leaving the rest to file frivolous class-action suits against cell-phone companies.

The Line

Handicapping is the art and science of setting the line, which in its most basic form looks like this (from the 8/14 lines on Yahoo! Sports):

FAVORITE	LINE	UNDERDOG	LINE
at Houston	-165	Chicago	+155
Reynolds		Bere

Or, to read that in English, you have to bet $1.65 on the Astros to get a dollar back, while a dollar on the Cubbies would get you $1.55 in return. You'd think it'd be written more clearly for the newbie, but it's not.

Let's say that two people, Joe and Pete, decide to place bets on their favorite teams, and decide to be cooperative and place uneven bets for my convenience. Joe bets $165 on Houston, and Pete bets $100 on Chicago.

If Houston wins, Joe gets his $165 back and picks up a cool $100 on top of that. Pete gets nothing and has to have a wrestler sit on his face during that week's WWF pay-per-view to keep the lights on. The bookie keeps Pete's $100, pays Joe $100 out of that, and keeps $0.

If the Cubs win, Pete gets his $100 bet and picks up a cool $155 on top of that. Joe loses his $165. The bookie takes Joe's $165, stashes $10, and pays off Pete. Pete then hits the horse track, loses his $255, and has to appear on QVC so he can go to Supercuts. Again.

And so the line is split -- the bookie makes $5 every time this particular game is played, if it's played over and over, as long as the money is pretty much even on both sides.

The line can move -- if everyone in New York has left their hellish city for a vacation in Vegas and they all placed bets on the home team to win, New York will become more of a favorite until there was equal money on each side. In this way, the market ensures that there's no risk to the casino.

By ensuring there's always enough losers to pay the winners, the casino makes their spread regardless of the outcome, which is a low-risk, nice, consistent income. The line doesn't care about who's going to win and what the actual chances are except as a means to get the line pretty close to even right off the bat. They would rather balance the money and make their 5% percent on every event than be right.

Think of the line, and the bookie side of sports gaming in general, more like a stock-market hedge fund or automated market-trading programs, something that involves things you're familiar with -- like Coca-Cola and IBM -- but once removed, where the actual businesses are less important than the movement of the stock price.

The line, like the stock market, is an amazingly efficient method of valuation. If a claim, like "Team A will win the AL West," is too far out of line, money will drive it towards a better number. And to beat the house, you need to win enough to beat the house's overhead, which means you'll need to win at least 30 percent of the time to break even in theory, and much more in practice.

A great bettor who only picks the best situations and who shops their bets for the best odds might win 60 percent of their bets in a good year. Anyone who claims he can pick 70 percent of games over the long term is lying. Many a man has gone to Vegas with a system, a dream, and seed money, and walked away with none of them.

But what if?

Let's say that Pete Gambler comes up with a system that picks a certain kind of underdog, which gives Pete a good chance to beat the line 60 percent of the time. His return on a $1,000 bet on a +150 underdog will net him $1,500 if he wins, and cost him $1,000 if he loses, so he'll make on average of $500 per bet placed. To make $50,000 a year, then, he has to find 100 games a season suitable for him to bet on, and place a bet of $1,000 on each one -- requiring a bankroll of $100,000 over the course of a season.

Of course, he doesn't have to come up with all the money up front, and you can fund later bets out of earlier wins, but that also assumes Pete doesn't have to endure a losing streak. $50,000 on $100,000 invested over a season -- that ain't bad money. That said, there are only going to be 2-6 potential +150 or better potential games a day, which means his system would have to turn up one in five of those suitable for wagering, which is, off the top of my head, not likely. But it's a hypothetical system anyway, so we'll move on.

Pete decides that those kind of returns aren't enough, so he starts a sports-betting service, and through whatever means at hand he builds up a big, wealthy clientele. These clients start to bet huge sums of money on these underdogs. The line moves, ticking these underdogs towards their true value, and their returns diminish. If Pete has enough clients and money, the line will move too far the other way, and a truly sophisticated industry watcher can bet against Pete and eke out profits for a while.

This is the first piece: a sophisticated risk-mitigation system that is extremely difficult to beat, but extremely lucrative if beating it can be done, and with a huge and well-funded base of people eager to do so (remember, there's 380B out there).

The State of the Masses

The current state of the people trying to make money off of the casinos is not pretty. If anyone out there is using sophisticated methods, they're wisely keeping quiet about it. The extent of statistical analysis seems laughable; it would seem, from cruising the Web, that it extends to figuring out what teams are on 'hot' streaks and thus good bets.

This is in the grand tradition of gambling advice, which endorses the streak view. Avery Cardoza's How to Win at Gambling (1991) essentially boils down to two pieces of advice: minimize your losses by only taking those bets where the house has minimal advantage, and bet more during your winning streaks (and, conversely, walk away from your losing streaks).

The mind of the sports bettor is a strange thing. It is edge-hungry, calling 900 numbers offering impossible hit rates, biting on baited hooks, supporting an entire industry rife with the sort of misrepresentation that would make an old-school carny barker proud. Honest 900 ads might tout their current hot streak. Dishonest ones might claim an 80 or even a 90-percent success rate, and I'm surprised that I wasn't able to find examples of FCC actions against fraudulent services.

Sports bettors are greatly influenced by baseball coverage at large. Statistics once intended to offer information on what happened in a game (for instance, saves and RBI) are used now as a measure of ability, and we get closers and clutch hitters, and these stats are incorporated into betting strategies and lines.

This makes sense: in large part, the people who are betting are fans of the sports they bet on, and the vast majority of content available to the sports fan, from Baseball Weekly to the beat writers across the country, take more or less the same viewpoint. This shared perception of what wins games determines where many bettors put their money. If a touted pitcher just called up is given a profile on ESPN the night before a start, people will see an opportunity to take advantage of a line which penalizes the player for being an unknown -- and thus move the line.

The Opening Shots

Sports gambling is going to become more and more stats-oriented because, unlike astrology, strong analysis offers a long-term way to find an edge on the line, to pick spots where conventional wisdom is clearly wrong and make money from it. There is too much money and too many smart people interested in that money for this not to become obvious to them.

And winning ideas are contagious. Word will get out if teams with weak middle relief are found to hurt their starters enough to throw the line off, or if any other consistent exploitable hole appears in the lines. Over time sports bettors who want to win and are good enough to be impartial will look at the odds sheets armed with park-independent data and quality team evaluations, and those bettors will smoke Joe Tourist betting on his local team to win.

The line will move, slowly, as the money does, but as long as the majority of the betting public remains relatively uninformed, it won't move much, keeping the edge profitable. Eventually though, this will become the great story of the year, probably in some rag like Time: "Beating the House: How Statistics and Gamblers Came Together" will be a June cover story a couple years away. Once the story breaks, we can easily predict what's going to happen in each corner.

[Next: If you think MLB is irked by complex, highly useful stats now, wait until they finally have a good reason to strap on the jackboots. In Part Two, appearing next Thursday here on Strikethree.com.]

Custom Search