Data Mining and the Art of Coaching ... How IBM Advanced Scout Put the Spotlight on NBA Sixth Man Award Winner Darrell Armstrong.

ARMONK, N.Y.--(BW SportsWire)--May 14, 1999--

Art or science? Defining the elusive mix of skills a coach uses to spot talent is always difficult. But for Orlando Magic's assistant coach Tom Sterner, there's little doubt about the helping hand lent by IBM's Advanced Scout data mining application in identifying the unique talents of Darrell Armstrong, this year's NBA Sixth Man Award winner.

Developed by IBM specifically for use by NBA coaches, Advanced Scout automatically mines massive amounts of statistical data from league games and condenses it into useful pieces of information, such as identifying player patterns and scenarios that box scores don't show. This information gives coaches new insights about their own teams and their opponents.

According to Sterner, one of the earliest users of IBM Advanced Scout, the tool played a pivotal role in spotlighting Armstrong's talent as the spark plug that could ignite his team during crucial moments.

In the 1997 playoff series Orlando trailed Miami 0 games to 2 having been outscored by 35 and 17 points in these games. Using IBM's Advanced Scout, Orlando coaches discovered weakness in the point guard and center positions. What the Orlando coaches saw by relating Advanced Scout information to the game video was something not likely to be picked up by reading a simple box score. Darrell Armstrong's presence on court in the point guard position meant better looks for Penny Hardaway and Miami's shooting percentage was lower when Armstrong was in the game compared to the other choices for point guard. Orlando also noted that Danny Schayes rebounds well and the Magic's overall shooting percentage was higher when he was matched against Mourning at center.

With this information from Advanced Scout the Orlando Magic decided to play Darrell Armstrong at point guard and Danny Schayes for more minutes. With these changes, Orlando was able to win the next two games at home and even the series, but was unable to return to Miami and win game 5.

The rest is history. Since that turning point, Armstrong's playing time per game increased from an average of 3 minutes to 35 minutes. As recipient of this year's Sixth Man Award, he's been honored as the league's premier player who served as a reserve in more game's than he's started. Armstrong averaged career highs in points (13.8), assists (6.7), rebounds (3.6) and steals (2.16) in 30.0 minutes per game while coming off the bench in 35 of 50 games this season.

Would Armstrong's talent have been spotted without IBM Advanced Scout? "No doubt, it was definitely a contributing factor" says Sterner, "But Advanced Scout gave us the insight to make an on-the-spot decision about a player that we might not have made by simply analyzing box scores. It helped us win two games in the 1997 playoff series, and more importantly, has contributed to Darrell's progress as a premier team player. This is a case where the art of coaching was substantially enhanced by the science of computing."

According to IBM's Tom Burke, senior manager for sports marketing, the company developed Advanced Scout for the NBA specifically to help showcase the competitive advantage that data mining technology can bring to businesses of all kinds. "The tool is now used by 22 NBA coaches," says Burke. "We've used the same data mining principles that underlie Advanced Scout to develop and refine similar solutions for customers in a variety of industries."