Can de­ci­sion sci­ences help teams win the IPL?

Let’s see how an­a­lyt­ics and de­ci­sion sci­ences have the po­ten­tial to in­flu­ence suc­cess in the IPL by en­abling ap­pro­pri­ate player se­lec­tion, as well as help­ing the teams to de­velop smarter strate­gies and take bet­ter match de­ci­sions

InformationWeek - - Front Page - The ar­ti­cle has been au­thored by Aditya K As­so­ciate Di­rec­tor, Mu Sigma) and Ke­shav Athreya As­so­ciate Man­ager, Mu Sigma

oney­ball, the base­ball flick star­ring Brad Pitt, in­tro­duced In­dian au­di­ences to the con­cept of Saber­met­rics, de­fined as the art and sci­ence of ap­ply­ing find­ings from sys­tem­atic anal­y­sis of player and match data. The film nar­rated the story of how a team, small in fi­nances but big in heart and smarts, im­ple­mented Saber­met­rics to win more of­ten than big­ger teams.

The film led sports fans in the coun­try to won­der if sim­i­lar meth­ods could be ap­plied to achieve sig­nif­i­cant crick­et­ing suc­cess. This ques­tion could be an­swered by con­sid­er­ing the re­quire­ments for cre­at­ing such a sys­tem, in light of the sim­i­lar­i­ties and dif­fer­ences be­tween the sports and their re­spec­tive en­vi­ron­ments. These re­quire­ments may be clas­si­fied into three broad head­ings: need for prob­lem iden­ti­fi­ca­tion, rel­e­vant data avail­abil­ity and eco-sys­tem de­sign.


Saber­met­rics took a while to es­tab­lish it­self amongst base­ball ad­min­is­tra­tors be­cause of the need to iden­tify wellde­fined prob­lems that it could an­swer com­pre­hen­sively. Billy Beane helped set­tle this to an ex­tent by com­ing up with a prob­lem — the story of which was nar­rated in Money­ball. He had to put to­gether a team that could hold its own against top teams but with a budget that was a frac­tion of other teams’.

This prob­lem of max­i­miz­ing rev­enues with con­straints on re­sources has al­ready made it­self known to cricket au­di­ences in the form of IPL player auc­tions. This re­sulted in cer­tain teams be­ing bet­ter heeled than the oth­ers. Though cer­tain safe­guards and mea­sures have been placed to en­sure un­fair com­pe­ti­tion.

Apart from player se­lec­tion, teams could also look to de­velop smarter strate­gies and match de­ci­sions based on match data. Cur­rently, much time, en­ergy and com­pu­ta­tional re­sources are ex­pended on the study of player ‘mech­a­nisms’ and ‘tech­niques’, en­deav­ors which are bet­ter left to the play­ers them­selves.

Iden­ti­fi­ca­tion of the right prob­lem ar­eas could help set an en­vi­ron­ment which sys­tem­at­i­cally iden­ti­fies the rel­e­vant data from the avail­able data­base, thus gen­er­at­ing the arms and am­mu­ni­tions re­quired to win matches con­sis­tently.


Base­ball has a long his­tory of doc­u­ment­ing match data, go­ing as far back as 1870. This even­tu­ally helped cre­ate the

space for data driven an­a­lyt­ics in the sport, per­fected and im­ple­mented with re­sound­ing suc­cess by Billy Beane and the Oak­land Ath­let­ics. Cricket like­wise has a long and proud tra­di­tion of pre­serv­ing match data- there are de­tailed score­cards for matches played in the 17th century! Cricket’s mas­sive data­base of match data could be used to solve com­plex and sig­nif­i­cant prob­lem, if set up in a re­spon­si­ble eco-sys­tem.


This eco-sys­tem would need to be a unique or­gan­ism, in­her­it­ing the best qual­i­ties of math, busi­ness un­der­stand­ing and tech­no­log­i­cal knowhow, to meet the chal­lenges posed by sport’s evo­lu­tion.

Math: Math pro­vides the abil­ity to sum­ma­rize vast amounts of data into small nuggets of in­for­ma­tion. Player and team per­for­mance could be eas­ily con­densed into a few mean­ing­ful met­rics which could be stan­dard­ized and com­mu­ni­cated eas­ily. Math could also play a vi­tal role in sci­en­tif­i­cally dis­sect­ing the role played by var­i­ous fac­tors in in­flu­enc­ing even­tual match out­come. Ul­ti­mately, assig­na­tion of causes to each ef­fect could be made more sci­en­tific.

Busi­ness Un­der­stand­ing: Busi­ness un­der­stand­ing could play a role in iden­ti­fy­ing prob­lems which pro­vide the most wins per buck — those prob­lems which could pro­vide the great­est ben­e­fits for the in­vest­ments made.

Tech­nol­ogy: Match data is gen­er­ated ev­ery day. The best-in- class tech­no­log­i­cal know-how must be lever­aged in or­der to ag­gre­gate all this in­for­ma­tion. This would re­quire up to date knowl­edge of all the ad­vances made in the fields of Big Data — data ag­gre­ga­tion, sum­ma­riza­tion and vi­su­al­iza­tion are ex­pected to be key build­ing blocks of the eco-sys­tem. Com­bin­ing these el­e­ments into a co­her­ent whole would be a wor­thy chal­lenge for any­one in­ter­ested in the ap­pli­ca­tion of de­ci­sion sci­ences to crick­et­ing prob­lems. How­ever, as Billy Beane and his team at the Oak­land Ath­let­ics have demon­strated, the ben­e­fits for any­one will­ing to meet these chal­lenges head- on are sub­stan­tial. Per­haps as im­por­tant as on-field suc­cess, it is the deeper un­der­stand­ing of sport that de­ci­sion sci­ences pro­vides that could be the real re­ward of this chal­lenge.

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.