Inside the Shadowy World of High-Speed Tennis Betting | Hacker News Reader

Inside the Shadowy World of High-Speed Tennis Betting

fivethirtyeight.com

43 pointswallflower25 comments

"Tennis bans courtsiders to protect the sport’s integrity, U.S. Open tournament director David Brewer said."

"In 2011, the men’s and women’s tours made a deal to sell their scores through a company called Enetpulse, majority owned by IMG. Many of the buyers are sports-gambling websites that provide the scores for in-play betting — wagering on the match after play has begun."

So when someone goes to a match and transmits the points, it ruins the integrity of the game, and when the tournament operators do it, it's totally fine. I wish there were a law against blatant hypocrisy.

Counterpoint: allowing those watching the game to place bets on an event after it happened (or to pass information to others that enables those others to do so) is dishonest towards those who do not have access to that information. In some jurisdictions it also is against the law.

If the tournament organizer is the only channel through which information flow as towards the betting office, by definition, it is impossible for the better to know the outcome of something he bets on at the time of the bet.

Of course, there are other ways to prevent this. For example, one can forbid bets on events that are expected to happen within five or ten minutes. I guess that isn't done because it decreases revenue.

5-10 minutes isn't what they're betting on. They're betting on matches. You predict outcome of the match based on current information. If you have more of the match behind you, you can make more accurate bets, even if the outcome is still an hour away. The information asymmetry means you can find winning bets even if the market is otherwise efficient.

Interesting parallels between this story and the prologue of the excellent book Fortune's Formula. In the early 1900's a former Western Union employee became rich by setting up a wire service for bookies. He hired men to watch the horse races and flash the winner in code to a nearby building using a mirror . Someone in that building would wire the results to bookies who paid for the service. This prevented bookies from taking bets on completed races. Of course if someone tried to bet on a losing horse, the bookies would take that bet anyway.

Funny how similar it is to high frequency trading. They both invest a bunch of capital to get better/earlier information, then leverage it to make profitable bets and trades. Atleast in the professional gambling world they don't deny it's a zero sum game.

Aren't they both obviously a zero sum game? There's a winner and a loser to a bet. If you buy a stock and it then rises, you profit and the seller has missed out. There's really no magical insight gained by the phrase 'zero sum game'.

No. The stock market isn't a closed system. The movements of capital and prices in the stock market impact capital allocation overall. Hence the arguments that HFTs that provide liquidity benefit the market because liquidity enables a larger pool of firms to seek listings or issue equity or debt.

Yes thats basically true but if the stock market goes up overall it looks like net gains for everyone to many observers. It is not of course as the companies who issue the equity are technically in the short position.

Even the basketball league in New Zealand is introducing courtsiding policy. It strikes me as completely ridiculous in all sports.

If you want to offer real-time betting on sporting minutiae and for some reason you can't do it such that some able to bet or gain information in real-time could screw you over, tough luck.

What do we ultimately lose without the policy? An inability to bet on the next point right up to it happening? Big deal.

At the center of this is a Markov model that gives the courtsider's employer a small betting edge with each additional point.

http://summerofjeff.wordpress.com/2011/01/13/python-code-for...

Much better statistical models are probably possible. That is basically the simplest possible model of a match.

There are services you can subscribe to for live (in-play) statistical information for football. This is exactly the same thing. You position yourself on betfair, pay a bundle of money for a "courtside" feed, because all media broadcasts are delayed, and you have a slight advantage again the average punter who doesn't have access to the feed, oh and the data feeds aren't cheap either.

Usually it's the large internet based bookies who subscribe to these feeds so as not to over-expose themselves to the public

This is a hugely interesting article. I used to prosecute corrupt betting practices in horseracing - before I decided to join the startup ecosystem - and this is a huge problem. A combination of human refereeing delay and TV transmission, combined with the betting public's desire to have in-running better on all points/outcomes, ensures that the race is on to get the result of any point up to the exchanges before the market settles. Some say that the advent of Betfair caused this problem, this is of course not true, it simply gave the normal punter easier access to a greater range of options (i.e. lay betting).

Also back in the olden days of ring side bookies at the race tracks, it was illegal to transmit information "off-track".

It didn't stop people inventing systems to do it though.

Clever use of very old mobile phones which used acoustics to transmit bookies markets off-track, and then used for wagering purposes. It's been around for a while.

These day's it's just not necessary. You have the William Hills, the Ladbrokes, the Tom Waterhouses, the various on-line bookies, all providing easily scrape-able data sources.

Sorry there's something I'm missing. Are talking about games with no coverage at all?

When plays Djokovic vs Murray at US open, how many seconds before the TV/iPhone/Android applications transmits the signal/point/data?

I don't get why do they need a courtsider. Can someone explain what I'm obviously missing here???

The television delay for "live" sports seems to be surprisingly high (and has got worse since the advent of digital TV); we're talking 5-10 seconds here. Probably even more for streaming video.

Would be interested to see a breakdown of how all the additional latency is added. I'm sure part of it is from the increased of overlaid graphics. The recently-posted article about the virtual American football firstdown line mentioned that they added a 2/3 seconds delay to accommodate the technology (at least, in the original implementation).

Having dabbled in sports betting this sounds like a business with terrible ROC and no scalability.

These aren't guys dabbling in sports betting, they are betting $50k+ per game and often time it's multiple games a day. Having dealt with many of such people I can tell you that the ROC can be ridiculous. Not only that, it doesn't have to scale well because there are only so many sporting matches each day. Universities already do this with recruiters (sending individuals all over to get info on every kid in some assigned area)

I'm surprised they don't have custom hand-held score clickers that would allow them to send out scores more discreetly (e.g. by forwarding over a phone hidden in their pocket).

In the article it says that he does it with his phone in his pocket.

While I have some ideas, a bit surprised so little explanation on how info used to improve bets. The article took forever to not say very much.

"The FFT bans courtsiding because it “owns the right to the data.”

Wat. Especially given that courtsiders are usually making their point determinations before the official results. Has this argument ever been tested in court?

It never will be. Tennis tournaments take place on private property and you agree to certain terms that ban you from "transmitting match data" by purchasing a ticket and attending. The tournament can throw anyone out that they please at their discretion.

Is this real and not just intended to make a point about high frequency trading?

Reply on news.ycombinator.com