Meanwhile, I'll take note of an interesting post from Brendan I. Koerner, guesting over at TNC's place, about sports betting. Koerner's relative is, apparently, a successful professional gambler, and he recommends baseball and horses as the two best investment opportunities.
I'll get back to baseball...the first question I have is about this odd claim (my emphasis):
Horse racing, meanwhile, is a great way to make money if you're willing to put in time at the track. As my relative notes, 98 percent of the people who bet on horses have no idea what they're doing; they plunk down money based on gut feelings, past performance, or cuteness of a competitor's name. You can run rings around those folks if you're willing to attend morning workouts with stopwatch in hand, as well as understand when it's appropriate to take a risk on a parlay. (My relative's share in a successful Pick Six gambit is what put home ownership within his reach.)Well, first of all, yes, I suppose that accounts for 98% of action -- but one of those things isn't like the others. I'm confident that my past performance based wagers have an enormous advantage over bets based on hunches, names, lucky numbers, and the like. What I'm a lot less convinced of is that personally collecting workout info beyond what the Form has is going to be a major source of advantage. And, alas, there isn't really enough money wasted on hunch betting to overcome the large track take-out. That doesn't mean that you can't beat the horses, but I don't think there's any huge advantage available to anyone. My main feeling about racing over the time I've played the horses is that over time, the percentage of all possible useful information that is easily available to casual players (which certainly includes me, even when I've been to the track a lot more often than I go these days) has increased dramatically.
Betting on baseball? I don't know. Koerner says "if you have a contrarian notion that a certain ace is about to have an off game, you can really clean up." Well, sure, but that's not exactly a formula for success, is it? Well, unless you can actually predict ups and downs of front-line pitchers, but my sense is that we're talking random variation, not anything that's systematic (and therefore usable for betting).
On the other hand, as a long-time roto player, I've always believed that, over time, my sense of which pitchers were overrated in the league and which were underrated was pretty good. Now, granted, I've never been in an ultra-competitive league, and presumably sports betting is dominated by relatively smart money. Here's what I would do if I were thinking of starting to bet baseball: I'd make season-length, not day-to-day, decisions on teams and pitchers, and then play those decisions throughout the season. In other words, if I thought that Matt Cain was overrated and Jonathan Sanchez was underrated, I'd make a standard bet for the Giants on all Sanchez starts and against them on all Cain starts.
Not that I'd actually do that; I'm not going to bet against or root against the Giants, regardless. I've only played in AL-only roto leagues for that reason; it's bad enough that interleague play has left me with divided loyalties from time to time.
Anyway, I'm not at all convinced. I guess I'd be interested to hear from those who do bet on baseball (outside of fantasy types of betting): do you think it's theoretically beatable?
Posts
Multi years ago, I chaired the Statistical Analysis (research) Committee of the Society for American Baseball Research. We had a newsletter (that continues under other editorship) called By the Numbers, and in it we ran a couple of articles based on simulating a couple of betting strategies. One, as I recall, tested the strategy of betting on teams that had lost 3 or more games in a row and on teams that had won 3 or more games in a row (which struck me, anyway, as an internally contradictory system). The finding was that it was a good way to lose money; the outcomes tended to be random, and, given the vig, the expected loss was about 10%. For baseball, I suspect that would be a typical finding.
ReplyDeleteI'd be shocked if any brute force type of system worked; presumably, the smart money would move to it until the odds moved and it no longer worked.
ReplyDeleteI knew a guy (now deceased) who was a retired statistician from the IRS. He retired to S. Florida and lived on his boat. I knew him from sailing on the Chesapeake and kept up with him from time to time over almost 30 years. I believe what he told me.
ReplyDeleteHe used to hang in local bars and once met a guy who after a suitable period of acquaintance asked my friend if he would mind putting down a football bet from time for him with local bookies. They were substantial but not attention attracting amounts and the guy didn't mind if my friend put down some of his own money too. It happened only once or twice a season. What amazed my friend as a professional statistician was that the guy never lost despite sometimes betting against the prevailing odds.
So apparently there is a way to win more often than you should at football betting but unfortunately the knowledge resides among shadowy folks with networks of anonymous friends to place the money about and collect the winnings.
Jonathan, I quite agree that no "brute force" system can work. But I'd extend that. I don't think any system can work, for precisely the reason you point out. Once the way in which the system has worked has been discovered (or reverse-engineered), then the gambling market will arbitrage away its effectiveness.
ReplyDeleteSo while it's possible for someone to have an effective system in the short-run (ala Anonymous's comment, or what shows up in one of Dick Francis's books--I'm blanking on the title), that system will work only so long as its existence as a system is unknown.
doc,
ReplyDeleteYup. What presumably can work is: find the stuff that is currently undervalued, and play that up until it's no longer undervalued, and repeat forever. Is there always something undervalued enough to beat the house take-out? Hard to tell.
Pick up APBA
ReplyDeletebetter than roto though more work.
Having Jeff kent during his heyday in Roto was a no-brainer.
having him as a 2b6 where the ratings go from 5 (Larry Walker at second ) to 9 (legendary i.e. a Frank White) where 7 is average and 8 is a Boone caliber defender ...
well all those miscues start adding up.
It also gets you understanding under-rated.
Cain rocks in APBA.
Oh and give me a high OBP and team where everyone hits near or above 40 doubles even though they are all below 20 HRs ...and I will outscore almost any other team.
One thing I have seen translated between APBA and roto.
Building up the middle always works.
Pitching staff, C, 2B, SS, CF.
Then fill in the rest with whatever tier you can find.
Yea in roto CF is same as any other OF ... but you get the point.
Oh and you damn well better not root against the giants!!!