1 files changed, 49 insertions, 0 deletions
diff --git a/wired/old/published/poker3.txt b/wired/old/published/poker3.txt
new file mode 100644
index 0000000..144b299
--- /dev/null
+++ b/wired/old/published/poker3.txt
@@ -0,0 +1,49 @@
+
+
+The Luck of the Draw Still Has Computers Stumped or Poker-Playing Computers Still Can't Seduce Lady Luck
+
+by Scott Gilbertson
+
+
+If the rise of the machines hypothesized in The Matrix plays out like a game of chess, we're totally doomed. However, if it plays out more like a game of Texas Hold 'Em, we could still have a chance.
+
+Computers have long been able to outsmart humans at a number of games of strategy. Most famously, chess Grandmaster Garry Kasparov squared off against IBM's Deep Blue 10 years ago and lost. In April of 2007, computer scientists at the University of Alberta completed a program capable of beating human players at checkers.
+
+But when the same team put its poker-playing application to the test against human opponents last month, it came up bust.
+
+The biggest challenge facing programmers trying to solve poker is one of uncertainty -- the players don't have all the information about the game. Unlike Chess, where everything you need to know is on the board, poker players must grapple with not knowing their opponent's hand, something computer programmers call imperfect information.
+
+"Almost every problem you'd want to address in the real world is one of imperfect information," says Michael Bowling, Associate Professor at the University of Alberta, where much of today's cutting edge game-playing software is being developed.
+
+Last month, the Alberta team pitted their software in two heads-up limit games against Poker champion Phil Laak and fellow pro Ali Eslami. While Laak and Eslami were able to eke out a 2-1 win, the Alberta team sees the recent match as a victory.
+
+"We think this was a great success," says Bowling, "I think you need to go no further than the players to know how close we are to humans -- they were definitely scared of the program."
+
+Perhaps our best advantage over software is our adaptability -- we make do with imperfect information. While computers need massive amounts of processing power to sort through information and make decisions, the best poker players can reach the same conclusions with far less data.
+
+Professional poker players rely on a variety of information sources to figure out what cards an opponent holds. In both live and online games, betting patterns can be a good way to guess at the missing information. But so far the software is unable to learn from a player's betting habits.
+
+The Alberta team has yet to tackle the more subtle aspects of poker like so-called "tells" -- facial tics, vocal abnormalities and other changes in a human's behavior.
+
+Tells are often less important than many players believe, in fact they mean little if you aren't good at the other aspects of the game. Still poker pro Greg Raymer says, "tells are one of the biggest differentiators between a good player and a truly world class player."
+
+So long as computers are unable to learn, they will be just that -- good, but not great poker players.
+
+According to Jonathan Schaeffer, founder of the Alberta research program, that's the most difficult task for computers to replicate -- learning. "The program will play strong poker, but it won't adapt to your playing," he says.
+
+Raymer isn't worried about losing to poker software. Raymer, who won the World Series of Poker in 2004, says, "I don't think computer software will ever outperform humans in a live game because of it's inability to take advantage of tells."
+
+Indeed the human players in last month's match recognized the computer's inability to learn and were able to adapt their playing strategies to exploit the computer. If the program was capable of learning, however, it might soon have humans on the run.
+
+"Nash equilibrium programs these are the strongest that we have right now, but they don't learn," says Schaeffer.
+
+Schaeffer is referring to programs based on the work of John Nash, the mathematician whose life inspired the movie <em>A Brilliant Mind</em>. Nash helped develop Game Theory, which hypothesizes that a set of strategies exist in certain games where every player's return is maximized.
+
+For instance, in the game Rock, Paper, Scissors, the best strategy is randomness. To win, you should select each of the options an equal proportion of the time in no particular order or pattern. This is known as equilibrium -- statistically, each player should win one-third of the time, lose on third of the time and tie one third of the time.
+
+But Nash's theories aren't going to make you a World Champion Poker player. An equilibrium program isn't designed to win, it's designed to avoid losing. And, since it can't learn, it can't adapt to gain an advantage.
+
+"It's a very hard problem to solve. We're still looking for the magic recipes," says Bowling.
+
+In the mean time, the team is refining its equilibrium program. And Bowling promises a rematch is in works.
+