News Stay informed about the latest enterprise technology news and product updates.

Data-driven decisions win in blackjack -- and in business

Still debating the merits of gut feel vs. data-driven decisions? Take a page from Jeff Ma's book, which tells how he used data to win millions at blackjack.

Winning at roulette takes luck; winning at blackjack takes strategy. "It's the only game in the casino that's subject to something called 'conditional probability,' which means what you see impacts what you're going to see," said Jeff Ma at last week's inaugural Society for Information Management Boston Technology Leadership Summit in Newton, Mass.

Ma knows this better than most. In a past life, he was part of an MIT blackjack team that used math and card counting to win millions, inspiring the book Bringing Down the House by Ben Mezrich and the film 21, starring Kevin Spacey and Laurence Fishburne. More recently, Ma made a name for himself as a different kind of gambler -- a serial entrepreneur. It's safe to say he's winning in that arena as well: In April, Twitter acquired his latest venture, TenXer, a project management platform for engineers, for approximately $50 million, according to TechCrunch.

Rather than Ma's knack for making money, however, it is his unshakable belief in data-driven decisions -- in the face of some daunting setbacks -- that makes his personal story inspiring for IT professionals. During his presentation, Ma described how on two separate occasions while playing blackjack, a card game where players try to accumulate a score that's higher than the dealer's but less than or equal to 21, he lost $50,000 in one fell swoop. The second time it happened, he contemplated quitting before realizing "the decision to quit would have been based on two hands versus the year I'd been playing and done quite well," he said. Instead, he returned to the table and, in a couple of days, not only recouped his losses but walked away $70,000 richer.

Ma sees this as a lesson on how short-term value can lead to bad decision making, and it was one of several examples he shared with IT leaders that highlighted the power of sound strategy and data-driven decisions in blackjack -- and in business. Here are four more:

The fallacy of the gut feeling

Jeff Ma; author and speaker at House Advantage; predicitive analytics expert at ESPN; group product manager and director of business insights at TwitterJeff Ma

Ma jokes that the most dangerous person in the world is author Malcolm Gladwell because "he can [take] what is a faulty concept and make everyone believe it's true," he said. In his book Blink: The Power of Thinking Without Thinking, Gladwell makes the case for "thin-slicing," or relying on one's instincts (honed by experience) to make decisions. Even then, Ma argues, these gut feel decisions aren't data-less; decision makers are simply "using data that's in their brains," he said.

That point aside, when it comes to making decisions, Ma's rule of thumb is simple: "The best way to make decisions is via the data," he said. He guarantees that if he plays blackjack using basic strategy, which relies on math to objectively decide what to do next, against someone using gut feel, "I'll win more than you do," he said.

Don't fall for groupthink

Once, when Ma was sitting with a pair of 10s and the dealer was showing a 6, he decided to make an unconventional decision to split his hand into two. "As I look around the table, all of the people are looking at me like I'm an idiot. The floor person is standing behind the dealer looking at me like I'm an idiot. The people behind me, this crowd now gathering, thought I was an idiot," he said. With everyone around him questioning his decision, he began questioning it himself, falling prey to groupthink, he said.

Normally, Ma would agree because "20 against a 6 is a winning hand a high percentage of the time," he said. But Ma knew something else: He'd been counting cards or keeping tabs on the high cards versus the low cards he'd seen, and he knew the deck was stacked with kings, queens, jacks (all worth 10), tens and aces (worth 1 or 11, whichever makes a better hand). If everyone around Ma had that same information, and "if they had the strategy that I had, would they have wanted to split those 10s? Probably," he said.

In the end, the unpopular move paid off, and he won both hands. "We have to make decisions that are non-traditional, that are going to create conflict if we want to innovate and make change," he said. That can be hard to do, especially when groupthink puts pressure on IT professionals to make emotional rather than data-driven decisions.

The danger of omission bias

Blackjack players (and businesses) can fall prey to "omission bias," which Ma described as inactivity or indecision in the face of uncertainty. At the blackjack table, if a player has 15 and the dealer is showing 9, "the math definitively tells you to hit that," Ma said.

But players know that if they pull a 7, 8, 9, 10, jack, king or queen, they'll exceed 21 and lose fast. When players succumb to omission bias, they'll make the conservative decision to do nothing, despite what the math says. If the dealer wins the hand (likely, in this case, since the dealer won't need much to beat a 15), players often excuse their behavior, believing they did nothing to contribute to the loss.

"The only reason I'm here is because I believe in this whole process of being data driven," Ma said. "You have to believe in this process."

Right decision vs. right outcome

Too often, outcome rather than process is the metric that drives decisions, Ma said. In blackjack, for example, if a player has 15 and the dealer has a 9 showing, Ma would, again, advise the player to hit. "If you get a 6 to make 21, I'm a genius," he said. "If you get a 7 to make 22, I'm a moron."

The declaration of Ma's genius or moronity is determined by the outcome, but in both cases, the strategy -- and the math behind that strategy -- supports the decision to hit. A hand of 15 is a weak hand, especially when it's known the dealer has 9. In other words, process counts. "Everyone loves to do an analysis based on outcome when, in reality, you should do an analysis based on the process," he said.

Welcome to The Data Mill, a weekly column devoted to all things data. Heard something newsy (or gossipy)? Email me or find me on Twitter @TT_Nicole.

Next Steps

Mistakes CEOs make when taking on big data

Data-driven culture helps analytics team generate business value

Dell's chief data officer talks about building a data-driven culture

Dig Deeper on Enterprise business intelligence software and big data