Must being data-driven as an organization mean abandoning human ways of making decisions? The history of computers in chess suggests that these two seemingly incompatible approaches actually strengthen each other.
Can we elevate analytics and algorithms without crushing the necessary human insight or forgetting where we are trying to go?
The Rise of analytics in chess
For over a millennium, the game of chess has challenged players’ strategic skill, cunning, intelligence, planning, and even emotional insight as players bluff or attempt to read their opponent’s façade. That 64-grid board seemed uniquely the domain of humans.
That perception cracked 25 years ago, when the chess grand master Gary Kasparov faced off against a purpose-built supercomputer in this very game. Deep Blue was designed to play chess – it could compute 200 million moves per second, which according to its builder, IBM, was 199,999,997 more than Kasparov could. But winning chess takes far more than just being able to compute all possible moves. You must choose the best move to make, against this particular opponent at this particular time with this particular situation.
Could a computer do that? Kasparov didn’t think so. And yet after beating Deep Blue in the first game, the computer program rallied to be the first ever machine to be a Grand Master. While Kasparov went on to win the six-game match, the loss established that computers were capable of beating the best humans in the world, and since then they’ve only gotten better. A year later, Kasparov lost an entire match to an upgraded Deep Blue after it foiled a trap the Grand Master laid for it.
People + Data
Why should we care now about this history a quarter of a century later, when we all have that kind of power in our pockets? It’s obvious to everyone today that computers can do many things better than humans, including many tasks far more important than a board game. But that’s not the whole story. Kasparov didn’t just hand the game over to the computers. Instead, he started his own series of chess competitions with humans and computers, and discovered something extraordinary. A mediocre program combined with a decent human player could soundly defeat those same programs and grand masters alike.
“Human strategic guidance combined with the tactical acuity of a computer was overwhelming,” Kasparov noted.
In our lives and our businesses, many of us unfortunately fail to integrate digital analytics as Kasparov and his chess colleagues have. On one side, we may feel like the early chess masters, confident that the intuition, insight, wisdom, and skill needed to succeed is uniquely and sublimely the domain of humans. How could a dataset or algorithm possibly know more than we do? Other times, we cede all authority to algorithms, forgetting that machines can’t decide what matters most or set priorities or understand context.
I started my own analytics company because I want to help people and organizations learn how to achieve this integration of human and analytic strengths. It starts with recognizing that we have to choose, as humans, where we are trying to end up. What goals are we trying to achieve? What does “success” look like? No dataset or algorithm can decide this for you or your company. Only once we’ve made those human value judgments, then analytics can help us determine the most effective, fastest, or most efficient path to get there, and tell us when we take a wrong turn along the way.
Harnessing the Power of Analytics
In order to actually achieve such an amalgamation, though, three things have happen.
First, our strategies must include analytics as a core element, including what outcomes we care about and how to measure progress towards them. Second, we must commit to using data within all our navigation decisions once the objectives are set. Third, we need to be comfortable asking for, interpreting, and using analytics.
For organizations, step one requires having a clear data strategy that speaks to how analytics will be resourced, generated, and distributed, as well as determining at an organization level what key performance indicators will track each of the organizational goals. Then, front-line workers should have point-of-decision dashboards or reports directly tied to these team, departmental, and organizational goals. Finally, those decision-makers will only be able to use the analytics if they have the necessary training to interpret the analytics appropriately.
Data in everyday life
Many of us don’t run companies, so what does all this mean for the rest of us? Imagine you’re having knee pain, and you go to see your doctor. Your doctor informs you that you have bone-on-bone arthritis. From a strategic point of view, you have to decide what matters most to you. Reduce pain as much as possible? Achieve a certain level of mobility? How will you measure what counts as “pain” or “mobility”? If you’ve committed to using analytics, once you’ve determined that mobility is your primary goal, you would speak to your doctor about the metrics around the possible interventions, and if your doctor is committed to analytics, he or she would provide evidence-based options. And you’d have to make sure you had the data literacy needed to understand measures like “number needed to treat” and likelihood ratios of risk.
Are you ready to soar on the wind of data?
Twenty-five years on from Deep Blue, computer assistance in chess is now common-place. Those players who resisted analytics in the game either “got tired of losing and quit, or got tired of losing and adapted,” according to chess grandmaster Sam Shankland. Whether as individuals or as companies, are we ready to embrace digital analytics as our partner?
Doing so does not mean that we relinquish our human strengths, but that we are compensating for our human weaknesses.
One without the other is like a one-winged bird, incapable of flying anywhere. But together, we can soar to great heights.
