Unlike many casino games, such as roulette or craps, poker is as much about skill as it is about good fortune. And that skill has much in common with big data. The very best players will learn as much as they can about their hands, their chances, and their opponents so that when it comes to making that key strategic decision, they have as much information as possible to help them make the right choice.
Some players have even become famous for their statistical analysis of hole cards in Texas Hold ‘em. Mathematician Bill Chen came up with the Chen formula, which can be tricky to learn; but once mastered, it will help you to assign a value to any pair of cards you are dealt. David Sklansky took a comparable approach, but he grouped similar value hole cards onto an easy quick-reference sheet. The website Sharkscope takes things even further by analysing data from millions of poker hands across the globe so that players can track different strategies in real games in real time.
Testing the big data approach to poker to the limit, a computer named Libratus recently beat four of the world’s leading poker aces over a twenty-day tournament, winning $1.7m in the process. This was seen as a huge step forward for AI and big data, since there are far more random variables in poker than in a game like chess. Not only did the machine accurately predict the chances of success for its hands, but it also processed endless information about its opponents’ strategies and playing styles, and this improved its play with each day of the tournament.
No industry uses big data more than the gaming industry, as it works hard to not only stay ahead of the game and the odds, but also to keep customers interested and loyal. Data has been used from the start to work out the odds of different game outcomes, but today it is being used for much more subtle decisions, such as when it is worthwhile to tip the odds in favour of the player to encourage them to play. This would have seemed counterintuitive back in the day, but big data has helped these businesses see the bigger picture.
With over 100 million registered users, PokerStars certainly has big data at its fingertips and seeks out the best young talent for its business intelligence graduate program to help them make the most of that information for both the business and its customers.
Of course, in business—just like in poker—data isn’t everything, no matter how much of it you have. As soon as you introduce human beings into the equation, you’ll find that they can all too often base their decisions on hubris rather than analysis, especially when in the midst of a run of good fortune or what appears to be a ‘bull’ market. Conversely, one could argue that without a human’s flexibility to both use big data predictions and adapt to unexpected circumstances, even massive volumes of data will be flawed.
How often have we seen a sports team analyse its opponents to the nth degree to create a winning strategy, only to end up losing when events take unexpected twists? Without the ability to adapt and think on their feet, players can seem lost when plan A fails and there is no plan B.
So what lessons can business learn from the poker table? It is clear that in both, the more information you can gather, the better your chances of success—even if random, unpredictable events occasionally throw you off course. However, just like being at the poker table, it’s important to remember that your business will never have all the information it needs, and even if you get close, nothing is guaranteed. For all its computing power (and ultimate success), Libratus still lost many hands along the way. Yet sticking to a well-researched, data-driven formula will usually win out in the end. As Phil Simon says in his blog Poker, Predictive Analysis and Big Data, “Organisations that use Big Data well are tantamount to the (top poker players) of the world. No, they won’t always win, but I’ll bet on these progressive companies every day of the week.”