Both Dubner and the folks at Party Poker think so.
Although supposedly illegal for Americans, online poker and sports betting is a huge business. At any given time, as many as 100,000 people are playing poker at Party Poker. More than half of these are Americans.
Party Poker is successful for much the same reason that eBay is. Most internet activities like searching, shopping, or viewing porn, the user doesn’t care whether other people are using the same site. All the user cares about is when it comes to auctions and poker, the big poker web sites have a built in advantage (what economists call network externalities). The more players at a site, the easier it is to find just the sort of game you are looking for, just as if you are trying to sell something you want to be at the site with the most buyers. By taking a small cut of each pot, Party Poker and the other poker web sites have been minting money, making the founders among the richest people in the entire world.
Now, there is PartyGammon. The Party Poker people are hoping that they can do the same to backgammon. I would not bet against them given their past success.
For years, Dubner has been bragging to me that he is a fantastic backgammon player and bemoaning the fact that he can play online, but not for money. Now he has his chance to prove how good he is.
My hope is that he really isn’t that good after all. Otherwise, all the hours he is supposed to be working on the sequel to Freakonomics, instead he will be getting rich playing backgammon.
10 Comments
One big issue with that is that Chess and Backgammon have solution spaces that a computer can manage. There is a fixed information space that the computer can calculate or brute force ahead of time and map.
Poker on the other hand is a game of incomplete information, so it is much harder to write a computer program that will play as optimally as a human.
There are already programs out there that play chess and backgammon pretty well - so the risk of playing online is that someone is using a computer to play optimally. While in poker, that doesn’t really matter since the computer player would only be as good as the person/people who write it - the bhess/backgammon program can be completely optimal.
As a result, it is “dangerous” to play online if there is someone using a computer on the other side.
— stenzSome theorize that this is why Party is currently limiting the backgammon games to fairly low limits.
Hmm it would seem so. Such that poker is taking up a lot of prime time television. Not a huge surprise most online poker sites/players want a piece of action. I’d say get out your wallets and have some fun.
— SonyYes stenz is right. Way back in 1979 a computer program beat the then human world champion (although a closer analysis revealed that the human played better but had bad luck with rolls). I’m sure in the last 27 years the computer has gotten at least 36 times better/faster(Moore’s Law) so that it’s not even close to a competition now and that anybody with a decent computer at home can calculate the entire tree of backgrammon in a reasonable amount of time.
— sophistryI don’t know if there are better programs, but a couple hundred bucks gets you Jellyfish:
http://www.jellyfish-backgammon.com/
“JellyFish is a neural net based backgammon program that plays at a very high level. On the highest playing level it matches the best humans in the world, and on the very fast level 5 a top human will hardly win more than 55% of the time. Also, its use of the doubling cube is outstanding. JellyFish is able to play matches of any length, or ‘money games’ where each point is equally valuable.”
I’m only a backgammon dabbler, and the freeware version crushes me like a bug in short order. That alone (not to mention decent awareness of my own limitations) is enough to keep me away from sites where I might have to face it for money.
— Jim BiancoloMy understanding is that the top backgammon programs are based on neural nets and not on composing a whole game tree. For chess programs it’s the other way around. This means that when computing power increases, backgammon programs don’t actually get all that much better.
— Dr. FunkMy understanding of neural network computing would suggest exactly the opposite. The possible plays and outcomes of chess and backgammon are set. So, as computing power increases, the computer would just come to the same conclusion faster. However, the very nature of neural nets is such that as computing power increases, the possible number of virtual neurons and virtual synapses (the connections between neurons) increases. Its like your brain getting bigger and bigger. Kasparov could probably beat the pants off of a cro-magnon any day of the week. Brains are as powerful as they are precisely because they dont opperate on the very simple logic of charting all the outcome possibilities. Additionally, neural nets can learn and just like people (or any other animal) there is no optimal, maximum level of learning. Neural nets will only make better and better plays, where as game tree computers will only make faster and faster plays.
P.S. I’ve done some work with neural nets, but I’m by no means an expert.
— edwardmkingI keep waiting for poker’s 15 minutes of fame to be up, unfortunately the clock seems to have stopped. Don’t get me wrong, poker’s probably a decent game, BUT IS IT NOT A SPORT AND THEREFORE SHOULD NOT BE HOGGING SO MUCH AIRTIME ON ESPN! Sorry for the shouting.
— prosaIf backgammon becomes increasingly popular, no doubt we’ll also see it on ESPN all the time. Meanwhile, those of us who want to watch REAL sporting events are out of luck.
8. Poker has a little bit more of that dangerous aura or mystique than backgammon–I don’t think you’ll be seeing backgammon on ESPN at all.
— zenofguitarI found this blog due to this entry, and had to comment as an expert backgammon player, backgammon theorist, and the maker of a coming backgammon program, Zbot.
Backgammon programs (bots) are extremely strong, on the level of the top human players if not beyond. They rely on a good ability to evaluate positions rather than an exhaustive search. All of the top players use them as coaches and sparring partners. However, the fear of encountering an opponent consulting a bot (which is considered cheating in backgammon) is grossly overrated by most people who do not play backgammon.
It is difficult to cheat effectively, and it is relatively easy for observant opponents to determine whether you are cheating by analyzing their play with a bot, as most serious players do. A player who is cheating enough to gain an advantage will typically play annoyingly slowly, and will have an extremely low error rate. While some people have been caught cheating, I have encountered very little evidence of cheating in thousands of online games. The vast majority of my opponents are obviously not cheating, since they make plays no bot would make, and have high error rates. Backgammon servers may have an even easier time determining whether players are cheating than individuals do.
One of the biggest obstacles that PartyGammon faces is the mostly groundless fear that opponents are cheating.
— DouglasZareIt’s interesting the gambleing angle that backgammon seems to have. growing up I never thought of it as a gambling/ casino game. But even google now won’t sell ads for backgammon is so apperantly offenceive.
Another great NON gambleing backgammon site for folks that just want to have some wonderful gameplay, is hardwood backgammon at http://www.hardwoodbackgammon.com
Who knows if Backgammon is the next big thing, and does it need to be a gambleing site to work?
— Jonas