The Telegraph
Wednesday , February 12 , 2014
CIMA Gallary

Knight choice: tech over ‘gut’

- Calcutta IPL franchise to try ‘predictive’ software at auction
IF HE CAN, I CAN: Shah Rukh Khan’s KKR will bank on computing to form its squad, following in the footsteps of Brad Pitt’s character in the 2011 film Moneyball. Pitt plays baseball manager Billy Beane, who used ‘sabermatic’ analyses of player value to select cash-strapped Oakland Athletics’ 2002 team and led it to success.The Knights will use a software capable of similar analyses to take quick decisions at the IPL auction that starts on Wednesday

Bangalore, Feb. 11: If a Twenty20 game can turn in a matter of seconds, so can an Indian Premier League auction. Quick thinking and back-up plans are key.

Say, you are desperate to have Kevin Pietersen but a rival has bid a price you can’t match, for there’s a cap on how much you can spend on your entire team.

So you face a tricky choice: do you go for another swashbuckler like David Warner or for Kiwi sensation Corey Anderson, who may score fewer runs than either but make up for it by taking wickets?

But then, Anderson is likely to fetch high bids too. So, do you look for a cheaper but almost equally valuable option — someone now flying under the radar but who can become a star, a la Sunil Narine, while leaving you with enough cash to buy that top fast bowler you want?

It’s not a decision to be made at a moment’s thought — but bidding is on and you can’t afford time.

This is the sort of problem that the Kolkata Knight Riders will try to tackle with the help of high technology instead of the gambler’s “gut feeling” at the two-day IPL auction that begins here tomorrow.

The franchise will be using the cloud-based SAP Auction Analytics to pick the best possible side, team CEO Venky Mysore told reporters today.

“From depending largely on basic statistics and gut instincts, we’re now equipped to use high-quality predictive analytical technology at the auction,” he said.

Developed by SAP Labs India, the platform uses “predictive” analytics to advise the user in a split-second on the right call in a fluid auction situation.

“For example, if one major player we had targeted is bought by another team, this platform will tell us who is the next best in (those) terms,” Mysore said.

Value for money

SAP Labs sources clarified that the “predictive” element in the technology does not involve forecasting players’ on-field performance. It merely indicates who is the best replacement for whom, in terms of how much price he is likely to command in the auction.

It’s this “value for money” approach that is the software’s USP in competitive situations where clubs have widely differing buying powers (as in European football) or function under spending caps.

SAP sources said such technology has been used by Formula One and American sports franchises but never before in cricket.

The idea was pioneered by American baseball manager Billy Beane who, with the help of a mathematics whiz, assembled a competitive team for Oakland Athletics in 2002 despite the franchise’s sparse resources. With its less expensive players, Oakland won a record 20 consecutive games thanks to the minute and “objective” player-value analyses allowed by Beane’s “sabermatics” system.

His feat inspired Michael Lewis’s 2003 book Moneyball, which was made into an eponymous film starring Brad Pitt in 2011.


One key feature of such analyses is the stress on “situational statistics” as opposed to the conventional batting averages and strike rates.

For example, a T20 franchise manager may want to know a batsman’s stats in the end-overs, his ability to rotate the strike and his performances when his team-mates flounder. A bowler’s economy rate during power play and his success against top-order batsmen may be key.

In a newspaper article, Lewis has highlighted “selfless” basketball player Shane Battier whose team performs better and the opposition worse when he is on court —the sort of quality that traditional measures such as points scored, rebounds and assists may fail to identify.

“Every player to be auctioned are analysed with 25 data points like runs, balls, dot balls, boundaries, etc. Bowlers are analysed with data points relevant to them,” Mysore said.

“Even Barack Obama’s campaign managers had used such wide-spectrum data analytics to take decisions. International sports uses a lot of it, but analytics at this scale is new to India.”

A mountain of data from independent cricket statisticians and analysts went into the SAP Auction Analytics, said Anirban Dey, managing director of SAP Labs India.

The company said no other IPL team uses anything similar. SAP Labs’ global partners for similar products include America’s National Basketball Association, Formula One team McLaren, baseball giant San Francisco 49ers and American football team New York Yankees.