PlayerTrackingintheNBANicolasMonOverview.StatisticshavebeenrecordedabouteverybasketballgameeverplayedintheNBA,datingbackthe league’s first game in 1946.1 As is true formost professional sport leagues, these statsdescribe importanteventsthattookplaceduringagame– intheNBA’scase,theyarethingslike shot attempts, points scored, assists, etc. Traditionally, these are recorded by a humanwith a tally sheet, and only provide a surface view of the game – only certain events arerecorded. New player tracking technology now allows for much more granular resourceselectiontooccur,automatingandtransformingthebasketballstatistic-takingprocesstoallowinteractionsthatwerenotpreviouslypossibletoaidteamstrategyandilluminatethegameasawhole.Whatisbeingorganized?Thegameofbasketballisitsownorganizingsysteminwhichtheteamsorganizetheirplayersinpursuitofacommongoal–namelytoscoreandpreventtheotherteamfromscoring.Duringagame,many events occur as these playersmove and interact in differentways, yetmost oftheseeventsleavenotrace,andmostaretooburdensomeorsubtleforhumanstatisticianstorecord.Player tracking technology is able to outperform human agents by creating real-time digitalrepresentationsofthephysicalresourcesofabasketballgame,namelythe10playersandballonthecourt2.Fromthere,interactionresourcesareselectedbasedonthemovementpatternsofthesedigitalrepresentations,enablingthedetectionofthemoresubtleaspectsofthegame.Thissystemisabletoviewagameasacoachmight,recognizingthings likescreens,dribbles,differentdefensivestructures,andmore.3From the plethora of interaction resources that this system selects, proprietary softwareorganizesthemtocreatedescriptionresourcesthatdescribethegameonalmostanydesiredlevelofgranularity.Usersofthesoftwarecangeneratestatisticalreportsforawholeteamorindividual player, and easily create visualizations to illustrate a point. During a game, thesoftwarecanshowateamthingslike:whereagivenplayerhasshottheballinthelastquarter,whichof theirplayershashad thegreatestdefensive impact in thegame,orevenhow faraplayerhas traveledon thecourtover theentire season.4 Thus, thesystemenablesusers toquantifyaspectsofthegamethatwerepreviouslyunquantifiable.Whyisitbeingorganized?Basketballisahighlycompetitivesport,andprofessionalteamswithalotofmoneyatstakewillgotogreatlengthstogetanupperhandontheiropponents.Therefore,adeepunderstandingofthegametoenablesmartstrategyisparamount.Playertrackingtechnologyprovidesawayfor the complexities of the game to be better captured and quantified, enabling teams toemploynewdata-drivenstrategyapproaches.

Allcoachesknowtheconceptof‘good’and‘bad’shots–it’sobviouslyeasiertomakeanopenlayup than a contested 3-point shot. By capturing a great deal of information about shots,playertrackingenablesadeeperunderstandingofplayerperformancethroughmorerelevantstatistics,onesuchbeingQuantifiedShotQuality(qSQ).Thismetricdescribesthedifficultyofagivenshot,expressedasapercentage tellinghowoften theaverageNBAplayerwouldmakethat shot under those particular conditions. Thismetric helps distinguish between shooterswhoget easy looks and those thatmake tough shots, even though theymayhave the samefield goal percentage. This is highly relevant to teams considering offering a player amultimillion-dollardeal.5In addition tobetter understandingplayer performance, this newdata also illuminates teamstrategy as a whole. Coaches now have access to data that quantify the effectiveness ofdifferentdefensivestrategiesagainstagiventeamorplayer.Teamshaveusedthisdata-drivenapproachtodiscovernewstrategiesthathavehelpedthemwinplayoffseries.6Howmuchisitbeingorganized?The scopeof thisorganizing systemcoverseverygameplayed in theNBA, andat its core, ittracksonlythelocationoftheballand10playersonthecourtatarateof25timespersecond.800,000datapointsarecollectedpergame,totalingaround1billionoveraseason.7Theinteractionresourcescreatedareorganizedtocreateawiderangeofdescriptionresourcesatdifferentlevelsoffocus,describingindividualplayers,games,orentireteams.Forexample,traditional field goal attempts can now be automatically categorized at higher granularity,distinguishingcatch-and-shootshots,toshotstakenoffthedribble,todrivesandmore.Whenisitbeingorganized?Player tracking delivers multiple options for data streams in real time, organizing theinteraction resources on a play-by-play basis. Proprietary software allows teams andmediaoutletstointeractwiththedataduringorafteragametoaccessdesireddescriptionresourcesasneeded.8Howorbywhomisitbeingorganized?Six cameras installed in all NBA arenas track the positions of the 10 players and ball on thecourt in real time, to collect anabundanceofdata about the interactions in a game. Fromthere,proprietarysoftwareutilizesspatiotemporalpatternmatchingandmachine learningtorecognize and select the interaction resources from the game, collecting data about manyaspects of each event on the court. After these interaction resources have beencomputationally classified, the software allows for the creation of description resources todescribevirtuallyanyaspectofthegameinquestion.Whereisitbeingorganized?Asdiscussedabove,thePlayerTrackingsystemprovidesmultipleoptionsofdatastreams.Formostusers, the interactionresourcesareorganizedbytheproprietarysoftwareof theplayertrackingsystem itself, thus the information isorganizedby thesystem in theNBAarena that

the game is played in. Player tracking information is also stored in a historical database,enablinguserstoviewinformationaboutpastgamesandseasons.9Otherconsiderations.WhilethiscasestudyfocusesontheNBA,playertrackingisbeingusedinmanydifferentsports,andprovidesanexampleofhowautomation istransformingmanypartsofhumanlife. Withmachinesbeingbuilt that canunderstand the gameat a deep level, it stands to reason thatautomationmaysoonseepintootheraspectsofsports.Automatedrefereescouldbecoming,making‘missedcalls’athingofthepast.Beyondsports, thespatiotemporalpatternmatching thatenables thisorganizingsystemmaysoonbeapplied inotherareas. At itscore,playertrackingtechnologytracksandcategorizesmovement, thus many of the same techniques could be applied elsewhere to situations inwhich understanding the movement of things is important. As Rajiv Maheswaran, CEO ofSecond Spectrum10, explains, “I believe thatwith the development of the science ofmovingdots,wewillmovebetter.Wewillmovesmarter.Wewillmoveforward.”111"FirstGameinNBAHistory."FirstGameinNBAHistory.Web.05Dec.2016.<http://nbahoopsonline.com/Articles/firstgame.html>2"NBA.com/Stats."AWholeNewView.Web.05Dec.2016.<http://stats.nba.com/featured/whole_new_view_2013_10_29.html>3"TheMathbehindBasketball'sWildestMoves."TED.com.RajivMaheswaran.Web.05Dec.2016.<https://www.ted.com/talks/rajiv_maheswaran_the_math_behind_basketball_s_wildest_moves>4Manahan,Matt.“ALookAtTheNBAPlayoffsThroughSportVUStatistics.”SportTechie,Apr.2015.<www.sporttechie.com/2015/04/30/sports/nba/a-look-at-the-nba-playoffs-through-sportvu-statistics/>5"WhatAdvancedTrackingDataRevealsaboutShooters."ESPN.com.Web.05Dec.2016.<http://www.espn.com/nba/playoffs/2016/story/_/id/15530589/what-advanced-tracking-data-reveals-nba-shooters>6"SecondSpectrumShootsforArtificialIntelligenceinNBAFinals."TheWallStreetJournal.DowJones&Company,2016.Web.05Dec.2016.<http://www.wsj.com/articles/second-spectrum-shoots-for-artificial-intelligence-in-nba-finals-1465471801>7Partnow,Seth."NylonCalculus101:IntrotoSportVU."NylonCalculus.2015.Web.05Dec.2016.<http://nyloncalculus.com/2015/08/13/nylon-calculus-101-intro-to-sportvu/>8“SportsDataServices.”STATS,<www.stats.com/data-feeds-packages/>9"BasketballDataFeed."STATS.Web.05Dec.2016.<http://www.stats.com/sportvu-basketball-media/>10Leung,Diamond."NBAPartnersWithSportradar,SecondSpectrumInBettingData,Player-TrackingDeal."SportTechie.N.p.,2016.Web.05Dec.2016.<http://www.sporttechie.com/2016/09/22/sports/nba/nba-partners-sportradar-second-spectrum-betting-data-player-tracking-deal/>11Bloom,Taylor."HowTrackingBasketballPlayersCanHelpSocietyMoveForward."SportTechie.N.p.,2015.Web.05Dec.2016.<http://www.sporttechie.com/2015/07/07/trending/tracking-basketball-players-can-help-society-move-forward/>

PlayerTrackingintheNBAArtifact:HowdoesplayertrackingenableadeeperunderstandingofNBAplayers?

Description:Inaddressingthe‘why’questionofmycasestudy,Iclaimthatthissystemallowsformorerelevantmetricstobecomputedthanwerepreviouslypossible,givingteamsabetterunderstandingofthegameanditsplayers.Toillustratethispoint,IbrieflydescribeQuantifiedShotQuality,anewmetricthatdescribesthedegreeofdifficultyofaparticularshot.Theinfo-graphicaboveshowsthemultitudeinformationthatistakenintoaccounttocomputethisstatistic,includingthedistanceandspeedofdefenders,numberofdribbles,shotangle,andmore.

Description:TheabovetableshowshowanaggregationoftheqSQmetricoveraseasoncanprovideanewwayinwhichtocompareplayerperformance.Qualitatively,viewersofbasketballknowthatDeAndreJordanshootshighpercentageshots–infact,mostofthemaredunks.TheqSQmetricnowprovidesaquantitativemeasureforthisaspectofthegame,andsureenoughhehadbyfarthehighestqSQofthe2015-2016season.Onthesamenote,IcannowquantitativelysaythatKobeBryant’sshotsoverthatseasonwereonaveragethehighestdegreeofdifficultyintheleague.

VideoLink:https://vimeo.com/152740765Aboveisascreenshotfromavideo(linkabove)thatdemonstrateshowtheqSQmetriccanbeusedtobetterunderstandspecificplays.ThescreenshotshowsRussellWestbrooktakingahighlycontestedshot,withtheqSQofthatparticularshotoverlaidonthevideo.BoththevideofootageofthatshotandthecomputedqSQvaluebothdescribeaspecificeventthathappenedinthegame,thusthevideofootageaugmentedwiththeqSQdescriptionresourceisaneffectivewayofshowinghowpreviouslyunquantifiablethings(i.e.thequalityofashot)cannowbequantifiedwithplayertrackingtechnology.