“The Man Who Broke Atlantic City,” Don Johnson, professional blackjack player

High-level cardplayers and gamblers interest me because one of their main jobs is to quickly assess the value of the data they have in front of them; they must take account of the situation and make a decision in a split second. Making executive decisions is less interesting to me. I like it better when others do that, and I go along for the ride, observing.

Anyway, here is an interesting account of how a professional gambler, Don Johnson, recently ran off a string of big wins playing blackjack. This part in particular hit home:

Johnson is not, as he puts it, “naive in math.”

He began playing cards seriously about 10 years ago, calculating his odds versus the house’s.

Compared with horse racing, the odds in blackjack are fairly straightforward to calculate. Many casinos sell laminated charts in their guest shops that reveal the optimal strategy for any situation the game presents. But these odds are calculated by simulating millions of hands, and as Johnson says, “I will never see 400 million hands.”

More useful, for his purposes, is running a smaller number of hands and paying attention to variation. The way averages work, the larger the sample, the narrower the range of variation. A session of, say, 600 hands will display wider swings, with steeper winning and losing streaks, than the standard casino charts. That insight becomes important when the betting terms and special ground rules for the game are set—and Don Johnson’s skill at establishing these terms is what sets him apart from your average casino visitor.

What Johnson’s suggesting is this: that in the interests of a singular actor (or singular set of actors) aggregative data need to be reflexively and immediately contextualized. Each of 400 million hands does not share the same logic, even if a singular logic does pervade the entire 400 million. The same is true with social data: each person of 300 million persons — a part of the same society, population, and/or political-economic system — does not follow the same rationality, even if, on average, a typical rationality can be deduced within the society. The context immediately surrounding the human behavior is where you go to understand the human behavior, like in cards when the preceding hands are where you go to make sense of the current hand.

Johnson’s lesson for data analysts was stated in another way by Jurgen Habermas, an almost unreadable German philosopher of capitalism and democracy. He wrote in The Theory of Communicative Action:

We may not assume, even in the cases of the capitalist economic organization and the modern state organization, any linear dependency of organizational rationality on the rationality of members’ actions.

This entry was posted in contextualized vs aggregative data. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s