SlideShare ist ein Scribd-Unternehmen logo
1 von 13
1
Cheating Detection in Call of Duty®
Arthur Von Eschen, PhD
Sr. Director, Game Analytics
Activision
Predictive Analytics World
Berlin, 2014
2
High Level… then Real Stuff
» Have you heard of Call of Duty?
» Do you play Call of Duty?
» High Level
» Provide context
» Real Stuff
» Actual analytics work
3
Console Gaming New to Analytics
Game Analytics – Our Place @ Activision
Independent Studio Model (autonomy)
4
Game Analytics – Our Place @ Activision
Central Studios
Started Aug 2012
5
Game Analytics
» No customers
» Thought reporting was analytics
» Completely dismissed that ‘math’ or
a ‘computer’ could do something
better than a human
» No desire for analytics
» Needed an internal case study
» Didn’t try to sell accuracy, better
decision making, etc
» I didn’t talk about analytics
» Sold as ‘automation of manual work’
» They focus on more valuable work
» Built analytic models as a
comparison to heuristic rules
» Cheating (or Boosting) detection
» Was this first project / case study
6
Call of Duty®
» A first person shooter
» Over 30 million people play
» For Black Ops 2
» 4.6 billion hours played
» 6.5 trillion shots fired
» 227 billion grenades thrown
» 386 billion kills made
» Multiplayer is the most
popular mode
» Video
» Bad player behavior can ruin
the experience
7
Boosting (Cheating) Detection
Boosting
» Most frequent of the
unwanted behaviors
» Most diverse patterns
The problem
» Integrity of leaderboards
and competitions
» Players 20% more likely to
quit playing for day
» Players 40% more likely to
drop out of match early
We built an analytic service
» Similar to fraud detection
» Multiple models
» Decision centric
» Runs at scale
Boosting
» /api/boosting/get_top_offenders/<…>
» /api/boosting/get_offender_stats/<…>
» /api/boosting/get_counts/<…>
Other unwanted behavior
» /api/cheating/<method>
» /api/spawn_trapping/<method>
8
Boosting Detection Service - Analytics
» Modeling - Classification
» Objective is to reduce false-positives
» Trained against known behaviors
» Tried logistic regression, decision trees, neural nets, rnd forest
» Settled on gradient boosting machine (GBM), AKA boosted trees
» Ensemble modeling technique
» Over 30 models in production
gbm.perf(m.deaths.gbm.final,plot.it = TRUE,oobag.curve = FALSE,overlay = TRUE,method="cv")
system.time(
test.pred_prob <- predict(m.deaths.gbm.final,data.deaths.test,type="response",n.trees=200,n.cores=4)
)
test.pred <- ifelse(test.pred_prob > 0.5,1,0)
confusionMatrix(data=test.pred,reference=data.deaths.test$boosted,positive = "1")
m.deaths.gbm.final <- gbm(boosted ~ death_space_killer_anomaly_x + death_space_victim_anomaly_y +
death_space_victim_anomaly_x + death_space_killer_anomaly_y + nemesis_deaths_ratio +
nemesis_deaths_total + hd_ratio + death_freq_anomaly + nemesis_distance_avg + bitch_kills_ratio,
data=rbind(data.deaths.train, data.deaths.test), shrinkage = 0.06, n.trees=200, cv.folds=10,
interaction.depth=29,distribution="bernoulli")
gbm.perf(m.deaths.gbm.final,plot.it = TRUE,oobag.curve = FALSE,overlay = TRUE,method="cv")
test.pred_prob <- predict(m.deaths.gbm,rbind(data.deaths.train, data.deaths.test), type="response",n.cores=4)
test.pred <- ifelse(test.pred_prob > 0.5,1,0)
confusionMatrix(data=test.pred,reference=rbind(data.deaths.train, data.deaths.test)$boosted,positive = "1")
9
Boosting Detection Service - Analytics
» Modeling - Identify new methods
» Over 10 models in production
» Modeling work wasn’t the hardest part
» Each model processes over
100 billion records each day
» Applying data transformation
» Applying sophisticated algorithms
» Single model took 12 – 14 hours
to score our 30 million players
» On a large analytics cluster
» We needed to score over 40 models
every day
10
Boosting Detection Service - Scaling
» Scaling was the most difficult step
» Database optimization
» Break apart queries, reuse datasets, caching,
flat-file import, chunk data
» Break process into multiple jobs and split
across server nodes
» Out of memory on DB, CPUs can’t handle
» Jobs compete for resources, not coordinated
» Query queue with query weights
» Query weights specify how “heavy” a query is
and the query manager does not send a
query for execution until there is “room”
» Run jobs in parallel (even for same model)
» Requires synchronization
» Use a msg queue (msg-oriented middleware)
» All 40+ models score < 3 hours
534
hrs
3 hrs
Before Scaling After Scaling
11
Boosting Detection Service – Exposing the Service
» Build a Web Service & UI
» For integration
» For analysis
» For audit purposes
» For reporting
» Quantify problem
» See trends
» Effort
» Data Prep
» Modeling
» Scaling
» API / UI
» Boosting
» /api/boosting/get_top_offenders/<…>
» /api/boosting/get_offender_stats/<…>
» /api/boosting/get_counts/<…>
» Other unwanted behavior
» /api/cheating/<method>
» /api/spawn_trapping/<method>
35%
10%
10%
45%
12
Boosting Detection Service – Results
» Quantitative Results
» The pre-analytics process
» 3 to 4 people
» Detecting roughly 200 people / day
» The analytic service
» <1 person to maintain
» Detecting over 40,000 people / day
» Lower false positive rate
» Organizational Results
» Proved that analytics (vs manual / heuristic approaches)
» Could be more precise
» Run at a greater scale
» And provide a stronger ROI
» Opened the door to greater adoption of analytics
» Increase in team size
» Increase of utilization in core product systems
1313
We’re Hiring!
Arthur Von Eschen
arthur.voneschen@activision.com

Weitere ähnliche Inhalte

Andere mochten auch

Predictive Analytics World Deutschland 2015
Predictive Analytics World Deutschland 2015Predictive Analytics World Deutschland 2015
Predictive Analytics World Deutschland 2015Rising Media Ltd.
 
gridComm_corporate_summary_Street Lights
gridComm_corporate_summary_Street LightsgridComm_corporate_summary_Street Lights
gridComm_corporate_summary_Street LightsTuck Long Nge
 
Industrial Analytics and Predictive Maintenance 2017 - 2022
Industrial Analytics and Predictive Maintenance 2017 - 2022Industrial Analytics and Predictive Maintenance 2017 - 2022
Industrial Analytics and Predictive Maintenance 2017 - 2022Rising Media Ltd.
 
Harness the Power of Microsoft Azure
Harness the Power of Microsoft AzureHarness the Power of Microsoft Azure
Harness the Power of Microsoft AzureRBA
 
The Psychological Effects Of Violent Video Games Slide Show
The Psychological Effects Of Violent Video Games Slide ShowThe Psychological Effects Of Violent Video Games Slide Show
The Psychological Effects Of Violent Video Games Slide Showbzan33
 
Domain-specific Modeling and Code Generation for Cross-platform Mobile and Io...
Domain-specific Modeling and Code Generation for Cross-platform Mobile and Io...Domain-specific Modeling and Code Generation for Cross-platform Mobile and Io...
Domain-specific Modeling and Code Generation for Cross-platform Mobile and Io...Università degli Studi dell'Aquila
 
Violence In Video Games
Violence In Video GamesViolence In Video Games
Violence In Video GamesMilkStealer1
 
Converged IoT Systems: Bringing the Data Center to the Edge of Everything
Converged IoT Systems: Bringing the Data Center to the Edge of EverythingConverged IoT Systems: Bringing the Data Center to the Edge of Everything
Converged IoT Systems: Bringing the Data Center to the Edge of EverythingDana Gardner
 
Benefiting from Big Data - A New Approach for the Telecom Industry
Benefiting from Big Data - A New Approach for the Telecom Industry  Benefiting from Big Data - A New Approach for the Telecom Industry
Benefiting from Big Data - A New Approach for the Telecom Industry Persontyle
 
Predictive Analytics in Telecommunication
Predictive Analytics in TelecommunicationPredictive Analytics in Telecommunication
Predictive Analytics in TelecommunicationRising Media Ltd.
 
Process Mining based on the Internet of Events
Process Mining based on the Internet of EventsProcess Mining based on the Internet of Events
Process Mining based on the Internet of EventsRising Media Ltd.
 
Hacking health: IoT, analytics and other trends
Hacking health: IoT, analytics and other trendsHacking health: IoT, analytics and other trends
Hacking health: IoT, analytics and other trendsJim Boland
 
Redefining IoT with Innovation At Every Level - Prith Banerjee, Chief Technol...
Redefining IoT with Innovation At Every Level - Prith Banerjee, Chief Technol...Redefining IoT with Innovation At Every Level - Prith Banerjee, Chief Technol...
Redefining IoT with Innovation At Every Level - Prith Banerjee, Chief Technol...Schneider Electric
 
Machine learning workshop @DYP Pune
Machine learning workshop @DYP PuneMachine learning workshop @DYP Pune
Machine learning workshop @DYP PuneGanesh Raskar
 
Building an IoT Kafka Pipeline in Under 5 Minutes
Building an IoT Kafka Pipeline in Under 5 MinutesBuilding an IoT Kafka Pipeline in Under 5 Minutes
Building an IoT Kafka Pipeline in Under 5 MinutesSingleStore
 
Enabling Real-Time Analytics for IoT
Enabling Real-Time Analytics for IoTEnabling Real-Time Analytics for IoT
Enabling Real-Time Analytics for IoTSingleStore
 

Andere mochten auch (17)

Predictive Analytics World Deutschland 2015
Predictive Analytics World Deutschland 2015Predictive Analytics World Deutschland 2015
Predictive Analytics World Deutschland 2015
 
gridComm_corporate_summary_Street Lights
gridComm_corporate_summary_Street LightsgridComm_corporate_summary_Street Lights
gridComm_corporate_summary_Street Lights
 
Industrial Analytics and Predictive Maintenance 2017 - 2022
Industrial Analytics and Predictive Maintenance 2017 - 2022Industrial Analytics and Predictive Maintenance 2017 - 2022
Industrial Analytics and Predictive Maintenance 2017 - 2022
 
Harness the Power of Microsoft Azure
Harness the Power of Microsoft AzureHarness the Power of Microsoft Azure
Harness the Power of Microsoft Azure
 
The Psychological Effects Of Violent Video Games Slide Show
The Psychological Effects Of Violent Video Games Slide ShowThe Psychological Effects Of Violent Video Games Slide Show
The Psychological Effects Of Violent Video Games Slide Show
 
Faclon Labs
Faclon Labs Faclon Labs
Faclon Labs
 
Domain-specific Modeling and Code Generation for Cross-platform Mobile and Io...
Domain-specific Modeling and Code Generation for Cross-platform Mobile and Io...Domain-specific Modeling and Code Generation for Cross-platform Mobile and Io...
Domain-specific Modeling and Code Generation for Cross-platform Mobile and Io...
 
Violence In Video Games
Violence In Video GamesViolence In Video Games
Violence In Video Games
 
Converged IoT Systems: Bringing the Data Center to the Edge of Everything
Converged IoT Systems: Bringing the Data Center to the Edge of EverythingConverged IoT Systems: Bringing the Data Center to the Edge of Everything
Converged IoT Systems: Bringing the Data Center to the Edge of Everything
 
Benefiting from Big Data - A New Approach for the Telecom Industry
Benefiting from Big Data - A New Approach for the Telecom Industry  Benefiting from Big Data - A New Approach for the Telecom Industry
Benefiting from Big Data - A New Approach for the Telecom Industry
 
Predictive Analytics in Telecommunication
Predictive Analytics in TelecommunicationPredictive Analytics in Telecommunication
Predictive Analytics in Telecommunication
 
Process Mining based on the Internet of Events
Process Mining based on the Internet of EventsProcess Mining based on the Internet of Events
Process Mining based on the Internet of Events
 
Hacking health: IoT, analytics and other trends
Hacking health: IoT, analytics and other trendsHacking health: IoT, analytics and other trends
Hacking health: IoT, analytics and other trends
 
Redefining IoT with Innovation At Every Level - Prith Banerjee, Chief Technol...
Redefining IoT with Innovation At Every Level - Prith Banerjee, Chief Technol...Redefining IoT with Innovation At Every Level - Prith Banerjee, Chief Technol...
Redefining IoT with Innovation At Every Level - Prith Banerjee, Chief Technol...
 
Machine learning workshop @DYP Pune
Machine learning workshop @DYP PuneMachine learning workshop @DYP Pune
Machine learning workshop @DYP Pune
 
Building an IoT Kafka Pipeline in Under 5 Minutes
Building an IoT Kafka Pipeline in Under 5 MinutesBuilding an IoT Kafka Pipeline in Under 5 Minutes
Building an IoT Kafka Pipeline in Under 5 Minutes
 
Enabling Real-Time Analytics for IoT
Enabling Real-Time Analytics for IoTEnabling Real-Time Analytics for IoT
Enabling Real-Time Analytics for IoT
 

Ähnlich wie Call of Duty® Cheating Detection with Analytics

DataEngConf SF16 - Three lessons learned from building a production machine l...
DataEngConf SF16 - Three lessons learned from building a production machine l...DataEngConf SF16 - Three lessons learned from building a production machine l...
DataEngConf SF16 - Three lessons learned from building a production machine l...Hakka Labs
 
Digital Attribution Modeling Using Apache Spark-(Anny Chen and William Yan, A...
Digital Attribution Modeling Using Apache Spark-(Anny Chen and William Yan, A...Digital Attribution Modeling Using Apache Spark-(Anny Chen and William Yan, A...
Digital Attribution Modeling Using Apache Spark-(Anny Chen and William Yan, A...Spark Summit
 
Presentation by Lionel Briand
Presentation by Lionel BriandPresentation by Lionel Briand
Presentation by Lionel BriandPtidej Team
 
Proper logging can catch breaches like retail PoS
Proper logging can catch breaches like retail PoSProper logging can catch breaches like retail PoS
Proper logging can catch breaches like retail PoSMichael Gough
 
Proper logging can catch breaches like retail PoS
Proper logging can catch breaches like retail PoSProper logging can catch breaches like retail PoS
Proper logging can catch breaches like retail PoSMichael Gough
 
Application metrics - Confoo 2019
Application metrics - Confoo 2019Application metrics - Confoo 2019
Application metrics - Confoo 2019Rafael Dohms
 
Large scale Click-streaming and tranaction log mining
Large scale Click-streaming and tranaction log miningLarge scale Click-streaming and tranaction log mining
Large scale Click-streaming and tranaction log miningitstuff
 
IEEE.BigData.Tutorial.2.slides
IEEE.BigData.Tutorial.2.slidesIEEE.BigData.Tutorial.2.slides
IEEE.BigData.Tutorial.2.slidesNish Parikh
 
Seven deadly sins of ElasticSearch Benchmarking
Seven deadly sins of ElasticSearch BenchmarkingSeven deadly sins of ElasticSearch Benchmarking
Seven deadly sins of ElasticSearch BenchmarkingFan Robbin
 
How Lyft Drives Data Discovery
How Lyft Drives Data DiscoveryHow Lyft Drives Data Discovery
How Lyft Drives Data DiscoveryNeo4j
 
Application Metrics (with Prometheus examples)
Application Metrics (with Prometheus examples)Application Metrics (with Prometheus examples)
Application Metrics (with Prometheus examples)Rafael Dohms
 
Pragmatic Machine Learning @ ML Spain
Pragmatic Machine Learning @ ML SpainPragmatic Machine Learning @ ML Spain
Pragmatic Machine Learning @ ML SpainLouis Dorard
 
Expert Network - Machine Learning Tech Days
Expert Network - Machine Learning Tech DaysExpert Network - Machine Learning Tech Days
Expert Network - Machine Learning Tech DaysStefano Tempesta
 
Application Metrics (with Prometheus examples) #PHPDD18
Application Metrics (with Prometheus examples) #PHPDD18Application Metrics (with Prometheus examples) #PHPDD18
Application Metrics (with Prometheus examples) #PHPDD18Rafael Dohms
 
AWS re:Invent 2016: Getting to Ground Truth with Amazon Mechanical Turk (MAC201)
AWS re:Invent 2016: Getting to Ground Truth with Amazon Mechanical Turk (MAC201)AWS re:Invent 2016: Getting to Ground Truth with Amazon Mechanical Turk (MAC201)
AWS re:Invent 2016: Getting to Ground Truth with Amazon Mechanical Turk (MAC201)Amazon Web Services
 
Machine learning for sensor Data Analytics
Machine learning for sensor Data AnalyticsMachine learning for sensor Data Analytics
Machine learning for sensor Data AnalyticsMATLABISRAEL
 
SmartData Webinar: Applying Neocortical Research to Streaming Analytics
SmartData Webinar: Applying Neocortical Research to Streaming AnalyticsSmartData Webinar: Applying Neocortical Research to Streaming Analytics
SmartData Webinar: Applying Neocortical Research to Streaming AnalyticsDATAVERSITY
 
Measuring Relevance in the Negative Space
Measuring Relevance in the Negative SpaceMeasuring Relevance in the Negative Space
Measuring Relevance in the Negative SpaceTrey Grainger
 
The importance of model fairness and interpretability in AI systems
The importance of model fairness and interpretability in AI systemsThe importance of model fairness and interpretability in AI systems
The importance of model fairness and interpretability in AI systemsFrancesca Lazzeri, PhD
 

Ähnlich wie Call of Duty® Cheating Detection with Analytics (20)

DataEngConf SF16 - Three lessons learned from building a production machine l...
DataEngConf SF16 - Three lessons learned from building a production machine l...DataEngConf SF16 - Three lessons learned from building a production machine l...
DataEngConf SF16 - Three lessons learned from building a production machine l...
 
Digital Attribution Modeling Using Apache Spark-(Anny Chen and William Yan, A...
Digital Attribution Modeling Using Apache Spark-(Anny Chen and William Yan, A...Digital Attribution Modeling Using Apache Spark-(Anny Chen and William Yan, A...
Digital Attribution Modeling Using Apache Spark-(Anny Chen and William Yan, A...
 
Presentation by Lionel Briand
Presentation by Lionel BriandPresentation by Lionel Briand
Presentation by Lionel Briand
 
Proper logging can catch breaches like retail PoS
Proper logging can catch breaches like retail PoSProper logging can catch breaches like retail PoS
Proper logging can catch breaches like retail PoS
 
Proper logging can catch breaches like retail PoS
Proper logging can catch breaches like retail PoSProper logging can catch breaches like retail PoS
Proper logging can catch breaches like retail PoS
 
Application metrics - Confoo 2019
Application metrics - Confoo 2019Application metrics - Confoo 2019
Application metrics - Confoo 2019
 
Large scale Click-streaming and tranaction log mining
Large scale Click-streaming and tranaction log miningLarge scale Click-streaming and tranaction log mining
Large scale Click-streaming and tranaction log mining
 
IEEE.BigData.Tutorial.2.slides
IEEE.BigData.Tutorial.2.slidesIEEE.BigData.Tutorial.2.slides
IEEE.BigData.Tutorial.2.slides
 
Seven deadly sins of ElasticSearch Benchmarking
Seven deadly sins of ElasticSearch BenchmarkingSeven deadly sins of ElasticSearch Benchmarking
Seven deadly sins of ElasticSearch Benchmarking
 
How Lyft Drives Data Discovery
How Lyft Drives Data DiscoveryHow Lyft Drives Data Discovery
How Lyft Drives Data Discovery
 
Application Metrics (with Prometheus examples)
Application Metrics (with Prometheus examples)Application Metrics (with Prometheus examples)
Application Metrics (with Prometheus examples)
 
Pragmatic Machine Learning @ ML Spain
Pragmatic Machine Learning @ ML SpainPragmatic Machine Learning @ ML Spain
Pragmatic Machine Learning @ ML Spain
 
Expert Network - Machine Learning Tech Days
Expert Network - Machine Learning Tech DaysExpert Network - Machine Learning Tech Days
Expert Network - Machine Learning Tech Days
 
Application Metrics (with Prometheus examples) #PHPDD18
Application Metrics (with Prometheus examples) #PHPDD18Application Metrics (with Prometheus examples) #PHPDD18
Application Metrics (with Prometheus examples) #PHPDD18
 
AWS re:Invent 2016: Getting to Ground Truth with Amazon Mechanical Turk (MAC201)
AWS re:Invent 2016: Getting to Ground Truth with Amazon Mechanical Turk (MAC201)AWS re:Invent 2016: Getting to Ground Truth with Amazon Mechanical Turk (MAC201)
AWS re:Invent 2016: Getting to Ground Truth with Amazon Mechanical Turk (MAC201)
 
SRE in Apiary
SRE in ApiarySRE in Apiary
SRE in Apiary
 
Machine learning for sensor Data Analytics
Machine learning for sensor Data AnalyticsMachine learning for sensor Data Analytics
Machine learning for sensor Data Analytics
 
SmartData Webinar: Applying Neocortical Research to Streaming Analytics
SmartData Webinar: Applying Neocortical Research to Streaming AnalyticsSmartData Webinar: Applying Neocortical Research to Streaming Analytics
SmartData Webinar: Applying Neocortical Research to Streaming Analytics
 
Measuring Relevance in the Negative Space
Measuring Relevance in the Negative SpaceMeasuring Relevance in the Negative Space
Measuring Relevance in the Negative Space
 
The importance of model fairness and interpretability in AI systems
The importance of model fairness and interpretability in AI systemsThe importance of model fairness and interpretability in AI systems
The importance of model fairness and interpretability in AI systems
 

Mehr von Rising Media Ltd.

Data Science at Roche: From Exploration to Productionization - Frank Block
Data Science at Roche: From Exploration to Productionization - Frank BlockData Science at Roche: From Exploration to Productionization - Frank Block
Data Science at Roche: From Exploration to Productionization - Frank BlockRising Media Ltd.
 
Cost-Effective Personalisation Platform for 30M Users of Ringier Axel Springe...
Cost-Effective Personalisation Platform for 30M Users of Ringier Axel Springe...Cost-Effective Personalisation Platform for 30M Users of Ringier Axel Springe...
Cost-Effective Personalisation Platform for 30M Users of Ringier Axel Springe...Rising Media Ltd.
 
Uplift Modelling as a Tool for Making Causal Inferences at Shopify - Mojan Hamed
Uplift Modelling as a Tool for Making Causal Inferences at Shopify - Mojan HamedUplift Modelling as a Tool for Making Causal Inferences at Shopify - Mojan Hamed
Uplift Modelling as a Tool for Making Causal Inferences at Shopify - Mojan HamedRising Media Ltd.
 
Behind the Buzzword: Understanding Customer Data Platforms in the Light of Pr...
Behind the Buzzword: Understanding Customer Data Platforms in the Light of Pr...Behind the Buzzword: Understanding Customer Data Platforms in the Light of Pr...
Behind the Buzzword: Understanding Customer Data Platforms in the Light of Pr...Rising Media Ltd.
 
Data Science Development Lifecycle - Everyone Talks About it, Nobody Really K...
Data Science Development Lifecycle - Everyone Talks About it, Nobody Really K...Data Science Development Lifecycle - Everyone Talks About it, Nobody Really K...
Data Science Development Lifecycle - Everyone Talks About it, Nobody Really K...Rising Media Ltd.
 
Creating Community at WeWork through Graph Embeddings with node2vec - Karry Lu
Creating Community at WeWork through Graph Embeddings with node2vec - Karry LuCreating Community at WeWork through Graph Embeddings with node2vec - Karry Lu
Creating Community at WeWork through Graph Embeddings with node2vec - Karry LuRising Media Ltd.
 
More than 10 Blue Links: Advanced-Level SERP Optimisation
More than 10 Blue Links: Advanced-Level SERP OptimisationMore than 10 Blue Links: Advanced-Level SERP Optimisation
More than 10 Blue Links: Advanced-Level SERP OptimisationRising Media Ltd.
 
How to Get Great Results Across Every Marketing Channel
How to Get Great Results Across Every Marketing ChannelHow to Get Great Results Across Every Marketing Channel
How to Get Great Results Across Every Marketing ChannelRising Media Ltd.
 
Don’t Freak Out! Tips for Mobile and Voice Search
Don’t Freak Out! Tips for Mobile and Voice SearchDon’t Freak Out! Tips for Mobile and Voice Search
Don’t Freak Out! Tips for Mobile and Voice SearchRising Media Ltd.
 
The Scout24 Data Landscape Manifesto: Building an Opinionated Data Platform
The Scout24 Data Landscape Manifesto: Building an Opinionated Data PlatformThe Scout24 Data Landscape Manifesto: Building an Opinionated Data Platform
The Scout24 Data Landscape Manifesto: Building an Opinionated Data PlatformRising Media Ltd.
 
Prescriptive ohne Predictive: Regression ist noch nicht tot! ROMI bei Unitymedia
Prescriptive ohne Predictive: Regression ist noch nicht tot! ROMI bei UnitymediaPrescriptive ohne Predictive: Regression ist noch nicht tot! ROMI bei Unitymedia
Prescriptive ohne Predictive: Regression ist noch nicht tot! ROMI bei UnitymediaRising Media Ltd.
 
Reinforcement Learning - Learning from Experience like a Human
Reinforcement Learning - Learning from Experience like a HumanReinforcement Learning - Learning from Experience like a Human
Reinforcement Learning - Learning from Experience like a HumanRising Media Ltd.
 
Mindful Analytics - Wie Achtsamkeit uns noch besser macht
Mindful Analytics - Wie Achtsamkeit uns noch besser machtMindful Analytics - Wie Achtsamkeit uns noch besser macht
Mindful Analytics - Wie Achtsamkeit uns noch besser machtRising Media Ltd.
 
Data Science Development with Impact
Data Science Development with ImpactData Science Development with Impact
Data Science Development with ImpactRising Media Ltd.
 
Predictive Analytics World for Business Deutschland 2018
Predictive Analytics World for Business Deutschland 2018Predictive Analytics World for Business Deutschland 2018
Predictive Analytics World for Business Deutschland 2018Rising Media Ltd.
 
Predictive Analytics World for Business Germany 2018
Predictive Analytics World for Business Germany 2018Predictive Analytics World for Business Germany 2018
Predictive Analytics World for Business Germany 2018Rising Media Ltd.
 
The Centrality of a Detailed Understanding of your Audience
The Centrality of a Detailed Understanding of your AudienceThe Centrality of a Detailed Understanding of your Audience
The Centrality of a Detailed Understanding of your AudienceRising Media Ltd.
 
Der steinige Weg zum automatisierten Data Science Produkt – Empfehlungen und ...
Der steinige Weg zum automatisierten Data Science Produkt – Empfehlungen und ...Der steinige Weg zum automatisierten Data Science Produkt – Empfehlungen und ...
Der steinige Weg zum automatisierten Data Science Produkt – Empfehlungen und ...Rising Media Ltd.
 
SpiegelMining – Data Science auf Spiegel Online
SpiegelMining – Data Science auf Spiegel Online SpiegelMining – Data Science auf Spiegel Online
SpiegelMining – Data Science auf Spiegel Online Rising Media Ltd.
 

Mehr von Rising Media Ltd. (20)

Data Science at Roche: From Exploration to Productionization - Frank Block
Data Science at Roche: From Exploration to Productionization - Frank BlockData Science at Roche: From Exploration to Productionization - Frank Block
Data Science at Roche: From Exploration to Productionization - Frank Block
 
Cost-Effective Personalisation Platform for 30M Users of Ringier Axel Springe...
Cost-Effective Personalisation Platform for 30M Users of Ringier Axel Springe...Cost-Effective Personalisation Platform for 30M Users of Ringier Axel Springe...
Cost-Effective Personalisation Platform for 30M Users of Ringier Axel Springe...
 
Uplift Modelling as a Tool for Making Causal Inferences at Shopify - Mojan Hamed
Uplift Modelling as a Tool for Making Causal Inferences at Shopify - Mojan HamedUplift Modelling as a Tool for Making Causal Inferences at Shopify - Mojan Hamed
Uplift Modelling as a Tool for Making Causal Inferences at Shopify - Mojan Hamed
 
Behind the Buzzword: Understanding Customer Data Platforms in the Light of Pr...
Behind the Buzzword: Understanding Customer Data Platforms in the Light of Pr...Behind the Buzzword: Understanding Customer Data Platforms in the Light of Pr...
Behind the Buzzword: Understanding Customer Data Platforms in the Light of Pr...
 
Data Science Development Lifecycle - Everyone Talks About it, Nobody Really K...
Data Science Development Lifecycle - Everyone Talks About it, Nobody Really K...Data Science Development Lifecycle - Everyone Talks About it, Nobody Really K...
Data Science Development Lifecycle - Everyone Talks About it, Nobody Really K...
 
Creating Community at WeWork through Graph Embeddings with node2vec - Karry Lu
Creating Community at WeWork through Graph Embeddings with node2vec - Karry LuCreating Community at WeWork through Graph Embeddings with node2vec - Karry Lu
Creating Community at WeWork through Graph Embeddings with node2vec - Karry Lu
 
More than 10 Blue Links: Advanced-Level SERP Optimisation
More than 10 Blue Links: Advanced-Level SERP OptimisationMore than 10 Blue Links: Advanced-Level SERP Optimisation
More than 10 Blue Links: Advanced-Level SERP Optimisation
 
How to Get Great Results Across Every Marketing Channel
How to Get Great Results Across Every Marketing ChannelHow to Get Great Results Across Every Marketing Channel
How to Get Great Results Across Every Marketing Channel
 
Don’t Freak Out! Tips for Mobile and Voice Search
Don’t Freak Out! Tips for Mobile and Voice SearchDon’t Freak Out! Tips for Mobile and Voice Search
Don’t Freak Out! Tips for Mobile and Voice Search
 
The Scout24 Data Landscape Manifesto: Building an Opinionated Data Platform
The Scout24 Data Landscape Manifesto: Building an Opinionated Data PlatformThe Scout24 Data Landscape Manifesto: Building an Opinionated Data Platform
The Scout24 Data Landscape Manifesto: Building an Opinionated Data Platform
 
Prescriptive ohne Predictive: Regression ist noch nicht tot! ROMI bei Unitymedia
Prescriptive ohne Predictive: Regression ist noch nicht tot! ROMI bei UnitymediaPrescriptive ohne Predictive: Regression ist noch nicht tot! ROMI bei Unitymedia
Prescriptive ohne Predictive: Regression ist noch nicht tot! ROMI bei Unitymedia
 
Reinforcement Learning - Learning from Experience like a Human
Reinforcement Learning - Learning from Experience like a HumanReinforcement Learning - Learning from Experience like a Human
Reinforcement Learning - Learning from Experience like a Human
 
Mindful Analytics - Wie Achtsamkeit uns noch besser macht
Mindful Analytics - Wie Achtsamkeit uns noch besser machtMindful Analytics - Wie Achtsamkeit uns noch besser macht
Mindful Analytics - Wie Achtsamkeit uns noch besser macht
 
Data Science Development with Impact
Data Science Development with ImpactData Science Development with Impact
Data Science Development with Impact
 
Predictive Analytics World for Business Deutschland 2018
Predictive Analytics World for Business Deutschland 2018Predictive Analytics World for Business Deutschland 2018
Predictive Analytics World for Business Deutschland 2018
 
Predictive Analytics World for Business Germany 2018
Predictive Analytics World for Business Germany 2018Predictive Analytics World for Business Germany 2018
Predictive Analytics World for Business Germany 2018
 
The Centrality of a Detailed Understanding of your Audience
The Centrality of a Detailed Understanding of your AudienceThe Centrality of a Detailed Understanding of your Audience
The Centrality of a Detailed Understanding of your Audience
 
Der steinige Weg zum automatisierten Data Science Produkt – Empfehlungen und ...
Der steinige Weg zum automatisierten Data Science Produkt – Empfehlungen und ...Der steinige Weg zum automatisierten Data Science Produkt – Empfehlungen und ...
Der steinige Weg zum automatisierten Data Science Produkt – Empfehlungen und ...
 
Data Alchemy
Data AlchemyData Alchemy
Data Alchemy
 
SpiegelMining – Data Science auf Spiegel Online
SpiegelMining – Data Science auf Spiegel Online SpiegelMining – Data Science auf Spiegel Online
SpiegelMining – Data Science auf Spiegel Online
 

Kürzlich hochgeladen

1:1定制(UQ毕业证)昆士兰大学毕业证成绩单修改留信学历认证原版一模一样
1:1定制(UQ毕业证)昆士兰大学毕业证成绩单修改留信学历认证原版一模一样1:1定制(UQ毕业证)昆士兰大学毕业证成绩单修改留信学历认证原版一模一样
1:1定制(UQ毕业证)昆士兰大学毕业证成绩单修改留信学历认证原版一模一样vhwb25kk
 
20240419 - Measurecamp Amsterdam - SAM.pdf
20240419 - Measurecamp Amsterdam - SAM.pdf20240419 - Measurecamp Amsterdam - SAM.pdf
20240419 - Measurecamp Amsterdam - SAM.pdfHuman37
 
DBA Basics: Getting Started with Performance Tuning.pdf
DBA Basics: Getting Started with Performance Tuning.pdfDBA Basics: Getting Started with Performance Tuning.pdf
DBA Basics: Getting Started with Performance Tuning.pdfJohn Sterrett
 
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default  Presentation : Data Analysis Project PPTPredictive Analysis for Loan Default  Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPTBoston Institute of Analytics
 
How we prevented account sharing with MFA
How we prevented account sharing with MFAHow we prevented account sharing with MFA
How we prevented account sharing with MFAAndrei Kaleshka
 
9711147426✨Call In girls Gurgaon Sector 31. SCO 25 escort service
9711147426✨Call In girls Gurgaon Sector 31. SCO 25 escort service9711147426✨Call In girls Gurgaon Sector 31. SCO 25 escort service
9711147426✨Call In girls Gurgaon Sector 31. SCO 25 escort servicejennyeacort
 
9654467111 Call Girls In Munirka Hotel And Home Service
9654467111 Call Girls In Munirka Hotel And Home Service9654467111 Call Girls In Munirka Hotel And Home Service
9654467111 Call Girls In Munirka Hotel And Home ServiceSapana Sha
 
INTERNSHIP ON PURBASHA COMPOSITE TEX LTD
INTERNSHIP ON PURBASHA COMPOSITE TEX LTDINTERNSHIP ON PURBASHA COMPOSITE TEX LTD
INTERNSHIP ON PURBASHA COMPOSITE TEX LTDRafezzaman
 
Advanced Machine Learning for Business Professionals
Advanced Machine Learning for Business ProfessionalsAdvanced Machine Learning for Business Professionals
Advanced Machine Learning for Business ProfessionalsVICTOR MAESTRE RAMIREZ
 
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM TRACKING WITH GOOGLE ANALYTICS.pptx
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM  TRACKING WITH GOOGLE ANALYTICS.pptxEMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM  TRACKING WITH GOOGLE ANALYTICS.pptx
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM TRACKING WITH GOOGLE ANALYTICS.pptxthyngster
 
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一F sss
 
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...Boston Institute of Analytics
 
PKS-TGC-1084-630 - Stage 1 Proposal.pptx
PKS-TGC-1084-630 - Stage 1 Proposal.pptxPKS-TGC-1084-630 - Stage 1 Proposal.pptx
PKS-TGC-1084-630 - Stage 1 Proposal.pptxPramod Kumar Srivastava
 
Defining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryDefining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryJeremy Anderson
 
From idea to production in a day – Leveraging Azure ML and Streamlit to build...
From idea to production in a day – Leveraging Azure ML and Streamlit to build...From idea to production in a day – Leveraging Azure ML and Streamlit to build...
From idea to production in a day – Leveraging Azure ML and Streamlit to build...Florian Roscheck
 
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改yuu sss
 
Call Girls In Dwarka 9654467111 Escorts Service
Call Girls In Dwarka 9654467111 Escorts ServiceCall Girls In Dwarka 9654467111 Escorts Service
Call Girls In Dwarka 9654467111 Escorts ServiceSapana Sha
 
GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]📊 Markus Baersch
 

Kürzlich hochgeladen (20)

1:1定制(UQ毕业证)昆士兰大学毕业证成绩单修改留信学历认证原版一模一样
1:1定制(UQ毕业证)昆士兰大学毕业证成绩单修改留信学历认证原版一模一样1:1定制(UQ毕业证)昆士兰大学毕业证成绩单修改留信学历认证原版一模一样
1:1定制(UQ毕业证)昆士兰大学毕业证成绩单修改留信学历认证原版一模一样
 
20240419 - Measurecamp Amsterdam - SAM.pdf
20240419 - Measurecamp Amsterdam - SAM.pdf20240419 - Measurecamp Amsterdam - SAM.pdf
20240419 - Measurecamp Amsterdam - SAM.pdf
 
DBA Basics: Getting Started with Performance Tuning.pdf
DBA Basics: Getting Started with Performance Tuning.pdfDBA Basics: Getting Started with Performance Tuning.pdf
DBA Basics: Getting Started with Performance Tuning.pdf
 
Call Girls in Saket 99530🔝 56974 Escort Service
Call Girls in Saket 99530🔝 56974 Escort ServiceCall Girls in Saket 99530🔝 56974 Escort Service
Call Girls in Saket 99530🔝 56974 Escort Service
 
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default  Presentation : Data Analysis Project PPTPredictive Analysis for Loan Default  Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
 
How we prevented account sharing with MFA
How we prevented account sharing with MFAHow we prevented account sharing with MFA
How we prevented account sharing with MFA
 
9711147426✨Call In girls Gurgaon Sector 31. SCO 25 escort service
9711147426✨Call In girls Gurgaon Sector 31. SCO 25 escort service9711147426✨Call In girls Gurgaon Sector 31. SCO 25 escort service
9711147426✨Call In girls Gurgaon Sector 31. SCO 25 escort service
 
9654467111 Call Girls In Munirka Hotel And Home Service
9654467111 Call Girls In Munirka Hotel And Home Service9654467111 Call Girls In Munirka Hotel And Home Service
9654467111 Call Girls In Munirka Hotel And Home Service
 
INTERNSHIP ON PURBASHA COMPOSITE TEX LTD
INTERNSHIP ON PURBASHA COMPOSITE TEX LTDINTERNSHIP ON PURBASHA COMPOSITE TEX LTD
INTERNSHIP ON PURBASHA COMPOSITE TEX LTD
 
Advanced Machine Learning for Business Professionals
Advanced Machine Learning for Business ProfessionalsAdvanced Machine Learning for Business Professionals
Advanced Machine Learning for Business Professionals
 
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM TRACKING WITH GOOGLE ANALYTICS.pptx
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM  TRACKING WITH GOOGLE ANALYTICS.pptxEMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM  TRACKING WITH GOOGLE ANALYTICS.pptx
EMERCE - 2024 - AMSTERDAM - CROSS-PLATFORM TRACKING WITH GOOGLE ANALYTICS.pptx
 
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一
办理学位证中佛罗里达大学毕业证,UCF成绩单原版一比一
 
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...
NLP Data Science Project Presentation:Predicting Heart Disease with NLP Data ...
 
E-Commerce Order PredictionShraddha Kamble.pptx
E-Commerce Order PredictionShraddha Kamble.pptxE-Commerce Order PredictionShraddha Kamble.pptx
E-Commerce Order PredictionShraddha Kamble.pptx
 
PKS-TGC-1084-630 - Stage 1 Proposal.pptx
PKS-TGC-1084-630 - Stage 1 Proposal.pptxPKS-TGC-1084-630 - Stage 1 Proposal.pptx
PKS-TGC-1084-630 - Stage 1 Proposal.pptx
 
Defining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryDefining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data Story
 
From idea to production in a day – Leveraging Azure ML and Streamlit to build...
From idea to production in a day – Leveraging Azure ML and Streamlit to build...From idea to production in a day – Leveraging Azure ML and Streamlit to build...
From idea to production in a day – Leveraging Azure ML and Streamlit to build...
 
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改
专业一比一美国俄亥俄大学毕业证成绩单pdf电子版制作修改
 
Call Girls In Dwarka 9654467111 Escorts Service
Call Girls In Dwarka 9654467111 Escorts ServiceCall Girls In Dwarka 9654467111 Escorts Service
Call Girls In Dwarka 9654467111 Escorts Service
 
GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]GA4 Without Cookies [Measure Camp AMS]
GA4 Without Cookies [Measure Camp AMS]
 

Call of Duty® Cheating Detection with Analytics

  • 1. 1 Cheating Detection in Call of Duty® Arthur Von Eschen, PhD Sr. Director, Game Analytics Activision Predictive Analytics World Berlin, 2014
  • 2. 2 High Level… then Real Stuff » Have you heard of Call of Duty? » Do you play Call of Duty? » High Level » Provide context » Real Stuff » Actual analytics work
  • 3. 3 Console Gaming New to Analytics Game Analytics – Our Place @ Activision Independent Studio Model (autonomy)
  • 4. 4 Game Analytics – Our Place @ Activision Central Studios Started Aug 2012
  • 5. 5 Game Analytics » No customers » Thought reporting was analytics » Completely dismissed that ‘math’ or a ‘computer’ could do something better than a human » No desire for analytics » Needed an internal case study » Didn’t try to sell accuracy, better decision making, etc » I didn’t talk about analytics » Sold as ‘automation of manual work’ » They focus on more valuable work » Built analytic models as a comparison to heuristic rules » Cheating (or Boosting) detection » Was this first project / case study
  • 6. 6 Call of Duty® » A first person shooter » Over 30 million people play » For Black Ops 2 » 4.6 billion hours played » 6.5 trillion shots fired » 227 billion grenades thrown » 386 billion kills made » Multiplayer is the most popular mode » Video » Bad player behavior can ruin the experience
  • 7. 7 Boosting (Cheating) Detection Boosting » Most frequent of the unwanted behaviors » Most diverse patterns The problem » Integrity of leaderboards and competitions » Players 20% more likely to quit playing for day » Players 40% more likely to drop out of match early We built an analytic service » Similar to fraud detection » Multiple models » Decision centric » Runs at scale Boosting » /api/boosting/get_top_offenders/<…> » /api/boosting/get_offender_stats/<…> » /api/boosting/get_counts/<…> Other unwanted behavior » /api/cheating/<method> » /api/spawn_trapping/<method>
  • 8. 8 Boosting Detection Service - Analytics » Modeling - Classification » Objective is to reduce false-positives » Trained against known behaviors » Tried logistic regression, decision trees, neural nets, rnd forest » Settled on gradient boosting machine (GBM), AKA boosted trees » Ensemble modeling technique » Over 30 models in production gbm.perf(m.deaths.gbm.final,plot.it = TRUE,oobag.curve = FALSE,overlay = TRUE,method="cv") system.time( test.pred_prob <- predict(m.deaths.gbm.final,data.deaths.test,type="response",n.trees=200,n.cores=4) ) test.pred <- ifelse(test.pred_prob > 0.5,1,0) confusionMatrix(data=test.pred,reference=data.deaths.test$boosted,positive = "1") m.deaths.gbm.final <- gbm(boosted ~ death_space_killer_anomaly_x + death_space_victim_anomaly_y + death_space_victim_anomaly_x + death_space_killer_anomaly_y + nemesis_deaths_ratio + nemesis_deaths_total + hd_ratio + death_freq_anomaly + nemesis_distance_avg + bitch_kills_ratio, data=rbind(data.deaths.train, data.deaths.test), shrinkage = 0.06, n.trees=200, cv.folds=10, interaction.depth=29,distribution="bernoulli") gbm.perf(m.deaths.gbm.final,plot.it = TRUE,oobag.curve = FALSE,overlay = TRUE,method="cv") test.pred_prob <- predict(m.deaths.gbm,rbind(data.deaths.train, data.deaths.test), type="response",n.cores=4) test.pred <- ifelse(test.pred_prob > 0.5,1,0) confusionMatrix(data=test.pred,reference=rbind(data.deaths.train, data.deaths.test)$boosted,positive = "1")
  • 9. 9 Boosting Detection Service - Analytics » Modeling - Identify new methods » Over 10 models in production » Modeling work wasn’t the hardest part » Each model processes over 100 billion records each day » Applying data transformation » Applying sophisticated algorithms » Single model took 12 – 14 hours to score our 30 million players » On a large analytics cluster » We needed to score over 40 models every day
  • 10. 10 Boosting Detection Service - Scaling » Scaling was the most difficult step » Database optimization » Break apart queries, reuse datasets, caching, flat-file import, chunk data » Break process into multiple jobs and split across server nodes » Out of memory on DB, CPUs can’t handle » Jobs compete for resources, not coordinated » Query queue with query weights » Query weights specify how “heavy” a query is and the query manager does not send a query for execution until there is “room” » Run jobs in parallel (even for same model) » Requires synchronization » Use a msg queue (msg-oriented middleware) » All 40+ models score < 3 hours 534 hrs 3 hrs Before Scaling After Scaling
  • 11. 11 Boosting Detection Service – Exposing the Service » Build a Web Service & UI » For integration » For analysis » For audit purposes » For reporting » Quantify problem » See trends » Effort » Data Prep » Modeling » Scaling » API / UI » Boosting » /api/boosting/get_top_offenders/<…> » /api/boosting/get_offender_stats/<…> » /api/boosting/get_counts/<…> » Other unwanted behavior » /api/cheating/<method> » /api/spawn_trapping/<method> 35% 10% 10% 45%
  • 12. 12 Boosting Detection Service – Results » Quantitative Results » The pre-analytics process » 3 to 4 people » Detecting roughly 200 people / day » The analytic service » <1 person to maintain » Detecting over 40,000 people / day » Lower false positive rate » Organizational Results » Proved that analytics (vs manual / heuristic approaches) » Could be more precise » Run at a greater scale » And provide a stronger ROI » Opened the door to greater adoption of analytics » Increase in team size » Increase of utilization in core product systems
  • 13. 1313 We’re Hiring! Arthur Von Eschen arthur.voneschen@activision.com