'Orac' is not happy

[bennythedip2]

Decades later, poker history repeats itself - By Mike Caro

(Note: Not at the old Poker1 site. A version of this entry was originally published (2008) in Bluff magazine.)

You’ve heard about Orac, right? Of course not! And if I don’t tell you about Orac, apparently nobody will.

I’m pissed. True, I’m supposed to be this mellow poker ambassador who politely shares his research and plays the clown as the so-called “Mad Genius of Poker” and never gets upset. But yesterday I was gloomier than ever. So I asked myself this unspoken question: “What’s wrong, Mad Genius?” And in the time it would take you to roll over a straight flush to win a $100,000 showdown, I had blurted out, “I’m really pissed.”

Let’s consider this self-therapy, and I’d be honored if you’d share this session with me.

Orac is “Caro” spelled backwards. Developed in 1982 and 1983, it made its first public appearance at the World Series of Poker in 1984. Orac also appeared on ABC television that year on Ripley’s Believe It or Not in a half-million dollar challenge at Vegas World, now known as the Stratosphere.

Orac was written up lavishly in over a dozen publications, including scientific and computer magazines in plain English (as well as in several foreign languages). It was a big deal for poker almost a quarter century ago. Browsing the web, I encountered headline stories reporting a series of heads-up, limit poker matches against some of the best online players. Polaris II, programmed by a team of computer science experts at the University of Alberta, headed by Michael Bowling, was able to compete competently against humans, winning three, losing two, tying once. This happened at the Rio during the 2008 World Series of Poker.

This is a remarkable achievement, and I know from experience how hard it is to program the complexities of poker. So, I begin to type a congratulatory notice as a lead-in to this column when suddenly, from out of nowhere, appears my director of operations, Diane McHaffie. She is perplexed, breathing warm air on my neck. Scanning the words on my computer screen, she says, “You’re too nice to everyone. You need to tell the truth about Orac.” Then, in what I interpreted as a failed dramatic attempt to win an award, she whooshes from the room. She isn’t done though. Fifteen minutes later, she returns with a stack of publicity about Orac and leaves it gently by my side, exiting again, this time without comment.

Doyle

Coincidentally, Doyle Brunson called a few minutes later and when I asked him what he knew about the computer challenge, he responded with, “Whatever happened to Orac.” Final straw. Brooding time for the Mad Genius. Minutes lost in memory. After 10 minutes, anger builds. After 15 minutes, truly pissed — which is where we began this column.

I programmed Orac in the early 1980s, spending much of two years in isolation. It ran on a glorified Apple II computer, tricked out with an additional monitor, a scanner, and a special ramp that allowed bar-coded cards to be read facedown by the computer. It began its life by playing limit until I decided to bite the bullet and deal with the complexities of no-limit play. As an artificially intelligent heads-up hold ’em player, it rolled over professional-level opponents in private tests in 1983 and even won most of its matches against me.

Then 1984 marked its public unveiling. It played in WSOP-sponsored exhibition matches against then world champion Tom McEvoy and twice world champion Doyle Brunson. It lost the match to McEvoy, although he eliminated it by drawing out with an inferior hand. And it won the second of two matches against Brunson, after losing the initial round when faced with an all-in decision on the first hand. That was my fault. Knowing Doyle Brunson as it did, I thought there was a good chance he would test the computer on the first hand by bluffing, and I had instructed Orac to make an exception and call with any pair or with ace-jack or higher. Result: Ace-jack wasn’t good enough. Match two had Brunson graciously conceding after an hour due to time constraints with Orac holding two-thirds of the chips.

Then came the major publicity match against Bob Stupak, owner of Vegas World on national TV. Orac may be the first computer with a bad-beat story to tell. The contract specified that if the computer froze or was unable to complete a hand, that hand would be ignored and be edited out of the video. Guess what? In the first of the two-of-three matches chasing $500,000, Orac flopped trips against Stupak’s top two pair. Orac moved all-in and Stupak called. Mysteriously a cable got kicked and the computer needed to be rebooted. The hand was deleted from history. Stupak went on to win.

History repeats

Anyway, when I scan the hundreds of online reports of the Polaris II achievement at the 2008 World Series of Poker, stating that a computer was finally able to compete on an equal level with humans, forgive me for thinking, “Been there, done that.” Every 24 years, history repeats itself.

In none of the stories did I find a single mention of Orac. To me, it’s like reporting that in 1927 an airplane successfully left the ground and landed safely, proving that powered flight is possible, without mentioning that the Wright brothers did it in 1903. I’m sure I sound egocentric in this rant, but I’m agitated. I guess I’m learning that many events that predate widespread use of the Internet remain unreported.

While programming Orac I learned a great deal about heads-up poker. Maybe you’ve heard me brag that if the aliens land and demand to play heads-up poker for control of Earth, I should be our chosen contestant. I’m serious about that — and I got that confidence while programming Orac.

To summarize: In reading this year’s coverage of the computer challenge at the WSOP three letters come to mind: WTF?

I feel better now. Shortly I’ll return to my more diplomatic self.

Maybe I’ll regret every word I just wrote, but Bluff’s deadline is right now and it’s too late to take it back. — MC