While examples like these represent the primary connections between gaming and data science though, there have also been some instances over the years of aspiring data scientists learning through games. And in fact, one of the games that can teach some fundamentals of how to use and analyze data is the same one that ultimately produces the stiffest challenge for AI: poker.
So how exactly can an aspiring data scientist learn fundamental practices through poker? It’s actually a fairly straightforward idea, if an intricate one.
Learn The Rules
First and foremost, it is necessary for anyone considering this sort of effort to learn the basics of poker inside and out. This means understanding the rules of the game (and its different varieties); knowing what different hands are worth; understanding betting sequences; and learning basic strategies. It sounds like a lot, and it certainly takes time to nail it all down. But the core fundamentals are relatively easy to pick up, even if you’re new to poker.
Choose Your Platforms
Next, you’ll need to choose platforms, both for playing games and for tracking results. When you play online, whether for real or fake money, you’ll have your pick of different websites and tournaments, and you’ll want to find one you feel you can rely on — and where, preferably, you’ll see the same opponents relatively often. Next, you’ll want to find a way to track hands played. There are software programs designed to do this automatically, such that the details of each hand played are recorded for you to reference after the fact.
Determine & Track Parameters
Once you understand how to play, and you’ve found a platform and a means of tracking results, you’ll need to determine what it is you want to track in order to begin assessing your opponents and how they play. You might measure something like how often they bet at a certain stage of a given hand, for instance — or how frequently they fold at the first round of betting. Over time, measuring tendencies like these alongside the basic math of the game (such as knowing how many different combinations of cards can be dealt), you can begin to glean invaluable information. This in turn can help you to begin predicting opponents’ moves and behavior with a degree of accuracy (whereas most players are largely guessing).
In terms of your poker game, gathering this kind of information over time — and over hundreds if not thousands of hands — will help to show you which types of players you have the most success against, and how you manage that success. But beyond poker, this sort of data gathering and analysis is also a means of learning segmentation. This is basically the practice of sorting people (typically consumers or clients, in this case poker opponents) into sensible categories that make them easier to deal with in a targeted manner. In poker terms, you might group regular opponents according to their tendencies and behavior, based on data that tells you with relative certainty where those opponents belong.
Ultimately, it really is a straightforward concept! But it’s one that some working data scientists will spend a great deal of time and effort on, because the results can be incredibly valuable. Being able to divide, segment, and cluster consumers according to known patterns and recognizable behavior means being able to approach those consumers more directly and tactfully. And the general process behind this kind of data science can actually be learned through little bit of time at the poker tables.