March 08, 2006

Good post, yerfatma. I'm a frequent visitor to The Sports Economist and was fascinated by the stuff in 'Freakonomics' (which my 10 year old is now reading - including the chapter on the economics of crack dealing in Chicago). I wouldn't have believed it was that common if not for the fact that (a) there are cases in the public domain where people have admitted to fixing games and (b) the stats back it up. I think it was Damon Runyon who said never to gamble on anything that could talk.

posted by owlhouse at 02:49 PM on March 08, 2006

Steven Levitt apparently provided "feedback" for the paper. My first instinct was that higher spread games are harder to handicap, but the paper pretty demonstrated a bias towards the favorites failing to cover by just a few points. Now I'm wondering if the betting public is overestimating the the favorites in these lopsided games. But I'm skeptical that such a market inefficiency would still exist. Basically, I just don't want to think that 5% of high spread games are being fixed. I wonder if Wolfers would advocate systematically betting the underdogs for big spread games. I think the juice is too high for this to work, though.

posted by mbd1 at 03:10 PM on March 08, 2006

Wow. Excellent post.

posted by rocketman at 05:00 PM on March 08, 2006

mbd1, I went through the same reactions as you, but I don't think they make sense (if I can remember my reasoning now). My versions of your theses were as follows: High-spread games may feature teams that do not play each other often, so comparative information about the teams is harder to come by, more likely to be biased. But the line is only initially set by Vegas. The efficiency of The Line is in the thousands of votes cast by bettors. There's no inherent reason why that should be less effective at different data points (unless there was a lack of data). Damnit, I forgot everything I was thinking about this earlier. The interesting thing would be to see what happens to the raw data as you redefine "heavy spreads". Which may be available in the paper. I should go check before I run my mouth.

posted by yerfatma at 06:46 PM on March 08, 2006

I don't think any sports are immune to this type of thing. "Sports," yes, but my (perhaps naive) guess is that this type of thing is pretty much restricted to those on the collegiate level, where the profit margin for gamblers is much greater. The cost of getting even a peripheral player at the professional level to jeopardize his career for more money than they are already making seems to me to be so much less a profitable enterprise. Nobody here, including the article, indicated that this was going on at anywhere other than the college level, but it seems to be a point worth making. This is a good post, but I'm curious to know how the Spo-Fi sample of the sports community views this in the big picture of good/bad. Sure, the notion of college players taking money from gamblers can make your skin crawl, but this scenario on paper shows that the better teams are always beating the lesser teams legitimately. Gamblers are coming ahead, but only at the expense of other gamblers, so the imbalance is between "crooked" gamblers and "honest" gamblers. Couldn't one just write this off as caveat emptor on betting slips, and shrug that the integrity of the sport is scarcely nicked, wins and losses weighing the heaviest in the measure of "integrity?"

posted by BullpenPro at 07:01 PM on March 08, 2006

this type of thing is pretty much restricted to those on the collegiate level, where the profit margin for gamblers is much greater The NYT article addresses this directly. The profit margin for gamblers shouldn't be any different (am I missing something?), but basketball is easier for one player to fix and, as you suggest, you'd have to offer a big incentive to get pro athletes to throw a game. Couldn't one just write this off as caveat emptor on betting slips I think one could; I can't. First of all, I think I can speak for the Spofi Econ contingent when I say the skewing of real outcomes is distasteful in and of itself. Beyond that, I suppose it's a question of perception. Shorting the spread may lead to more dramatic moments, but I prefer my dramatic moments born of honesty. I watch sports to see the result when two people or groups give 100%. I want to know who wins when everybody is trying their best.

posted by yerfatma at 07:09 PM on March 08, 2006

Yerfatma, I liked your post so much I read it twice. The profit margin for gamblers shouldn't be any different (am I missing something?) My implication was that, for the gamblers involved for the fix -- assuming they would bet the same in every instance -- since the cost of a fix would be higher at the pro level, their overall profit from the game would be lower. When I said "good/bad" I didn't mean to imply that some people would consider this scenario "good" at all -- since the basic scenario is the difference between a 15 point game and a 12 point game, I don't think "dramatic moments" plays into this too much. I just wonder how "bad" this is to the average fan. Is it DefCon 5 bad? I doubt it. There is a sizable thread on the front page that addresses another "bad" issue (I don't want to say what for fear of leaving blood in the water) which I would guess rates a higher alert than this one (maybe not, but that's what I'm asking). You kind of make my point, though, when you say, "I want to know who wins when everybody is trying their best." You DO know, don't you, if the winning team is the only one that's letting up -- again, assuming that wins and losses constitute a high percentage of a game's "integrity." On review: Owlhouse, I see your point, but I think a large enough percentage of the population fail to see gambling as a legitimate "market." (Some people, I'm sure, feel the same about the stock market, but my belief is that that population is much smaller.) Gambling is borderline taboo -- straight-up taboo in some regions -- so the notion of upsetting that market may not be looked at with so much disdain.

posted by BullpenPro at 08:07 PM on March 08, 2006

This is a great piece of work. The author suggests that the problem could be alleviated if against-the-spread betting went away, to be replaced by some kind of odds-based system a la horse racing. Would this significantly decrease, or increase the appeal of betting on basketball? For honest gamblers, I mean. TSE also mentions that the author is working on a similar analysis for pro ball, which should be interesting.

posted by Amateur at 07:18 AM on March 09, 2006

The information currently encoded in a game's spread would be moved to the payoff system, a la baseball betting Sure, but in MLB the odds never get too far out of whack. When North Carolina plays East Bumfuck State University, EBSU has very close to zero probability of winning. I understand that that would lead to enormous payoffs, and in that sense the two systems are equivalent ... but that doesn't mean that the "market appeal" is the same to the bettors.

posted by Amateur at 09:06 AM on March 09, 2006

i can think of another reason why heavy favorites don't cover. vegas will stretch the point spread on "public" teams . . . You realize the point of The Line is to set it to split the betting evenly between the two sides so Vegas makes money, right? No, you don't, otherwise you wouldn't suugest they regularly screw themselves on purpose. Let me suggest to you responding to an in-depth economic analysis with "this article is complete bullshit" and not backing up your hypothesis (such as it is) with evidence is not the way to foster healthy debate. When North Carolina plays East Bumfuck State University, EBSU has very close to zero probability of winning. Right, but betting has little to do with straight-up winning of contests. I take your point though; most likely those kinds of games would not be eligible for betting because they wouldn't draw enough action to even things out. No one is going to put down $10,000 to win $100. Well, not enough degenerates to make it risk-free for Vegas.

posted by yerfatma at 09:38 AM on March 09, 2006

Bet against a lot of teams favored by more than 12 points in college basketball games. </humorless>

posted by yerfatma at 11:32 AM on March 09, 2006

would an odds-based system become an incentive for gamblers to actually influence the outcome of the game rather than just shaving points The incentive to influence outcomes is there for the gamblers regardless; but the odds-based system changes the net incentive for the players / coaches. Players and coaches are highly motivated (for various reasons) to win the game; they are not highly motivated to cover the spread, so a financial incentive can easily provide motivation not to.

posted by Amateur at 12:45 PM on March 09, 2006

if you remove the gamblers' ability to influence their outcome without influencing the outcome of the game, doesn't that give them additional incentive to rig game outcomes Yes, probably, but it's immaterial: if you change the betting to straight-up, you remove the incentive asymmetry* (he said, trying to justify the money his parents spent), which would require much larger offers to actors involved in the game to affect outcomes. The change in betting rules would not affect gamblers' incentives in any way, but it would drastically affect the actors' willingness to ever throw a game. *If you're a shady gambler and I'm a star player on a team favored by 12 points, your desires and mine match up in the space where my team wins by 1-11 points. Take away spread betting and they no longer match up.

posted by yerfatma at 03:02 PM on March 09, 2006

While I was initially upset by the idea that 5% of the games involve point shaving, I've come to a conclusion that I just don't care. And here is why: 1) The only people being hurt are those that gamble on sporting events. The players aren't hurt and the fans aren't hurt (assuming they are of the "hope my team wins" type and not the "hope my team covers the spread so I can make $50" type). 2) The only games that seem to suggest being fixed (big spread) are games where the fundamental outcome (win/loss) aren't affected. If you win by 14 or win by 11, it is still a win. Now, if games were being fixed that affected the outcome (win/loss), then I'd be upset. Okay, more upset, as this still bothers me a little bit. I still would have no problem with all of the players involved (if proven to shave points) to be banned from organized sports (professional or collegiate) for life. And I'd like to add my voice to those that encourage people to read "Freakonomics". As a "stats" guy, I loved it, but I think even those that are averse to statistics would find it an interesting read. And then follow it up with "The Tipping Point", another excellent book about trends and sociology.

posted by grum@work at 05:22 PM on March 09, 2006

And Collapse, Jared Diamond's latest. Shouldn't we form a book club?

posted by owlhouse at 06:31 PM on March 09, 2006