Raghav Singal
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Raghav Singal has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Garud Iyengar
Department of IEOR, Columbia University, New York, New York 10027
4 shared publications - Martin Brendan Haugh
Imperial College Business School, Imperial College, London SW7 2AZ, United Kingdom
4 shared publications - Omar Besbes
Graduate School of Business, Columbia University, New York, New York 10027
3 shared publications - Vineet Goyal
Department of Industrial Engineering and Operations Research, Columbia University, New York, New York 10027
3 shared publications - Antoine Désir
Technology and Operations Management, INSEAD, 77305 Fontainebleau, France
2 shared publications - Craig Fernandes
Department of Mechanical and Industrial Engineering, University of Toronto, Toronto, Ontario M5S 3G8, Canada
1 shared publication
Latest Publications
- 23 October 2026
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
Online Allocation with Heterogeneous Supply and Minimum Workload Guarantees
Keeping Gig Workers Working How should platforms allocate jobs when workers need a minimum amount of work to stay engaged? In their accepted paper, “Online Allocation with Heterogeneous Supply and Minimum Workload Guarantees,” Garud Iyengar and Raghav Singal study this question in online labor marketplaces, where workers differ in quality, capacity, and desired workload. The authors show that intuitive allocation rules can perform poorly. A first-come-first-served policy can favor lower-quality workers, whereas a quality-based greedy policy can cause high-capacity workers to leave when their workload requirements are not met. They develop Tracker, an online policy that combines admission control with a simple target-tracking rule: Each worker is assigned a target workload, and arriving jobs are directed toward workers who are most under their targets. Despite the stochastic nature of demand, the authors establish that Tracker is optimal. Computational experiments calibrated to a real labor marketplace show substantial improvements in allocation quality relative to standard policies, while maintaining workers’ minimum workload guarantees.
- 23 October 2026
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
Online Allocation with Heterogeneous Supply and Minimum Workload Guarantees
Keeping Gig Workers Working How should platforms allocate jobs when workers need a minimum amount of work to stay engaged? In their accepted paper, “Online Allocation with Heterogeneous Supply and Minimum Workload Guarantees,” Garud Iyengar and Raghav Singal study this question in online labor marketplaces, where workers differ in quality, capacity, and desired workload. The authors show that intuitive allocation rules can perform poorly. A first-come-first-served policy can favor lower-quality workers, whereas a quality-based greedy policy can cause high-capacity workers to leave when their workload requirements are not met. They develop Tracker, an online policy that combines admission control with a simple target-tracking rule: Each worker is assigned a target workload, and arriving jobs are directed toward workers who are most under their targets. Despite the stochastic nature of demand, the authors establish that Tracker is optimal. Computational experiments calibrated to a real labor marketplace show substantial improvements in allocation quality relative to standard policies, while maintaining workers’ minimum workload guarantees.
- 9 February 2026
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Manufacturing & Service Operations Management
Frontiers in Operations: Bounding Counterfactual Outcomes of Health Insurance Delay-and-Deny Practices
Problem definition: Health insurance delay-and-deny practices, such as requiring prior authorization, frequently restrict access to timely and necessary medical care. These policies can have severe consequences, including delayed diagnosis and treatment, leading to poor patient outcomes or even death. Counterfactual analysis offers a way to assess the impact of these practices by bounding the likelihood that a patient would have survived under alternative policies given the severe consequence under the original policy (probability of necessity). However, quantifying counterfactual probabilities is challenging in dynamic systems, such as those arising in healthcare, where disease progression is often latent and evolves over time. Methodology and results: We develop a principled framework to bound counterfactual probabilities, specifically the probability of necessity (PN), in generalized hidden Markov models (GHMMs). Our approach leverages the structure of GHMMs to construct a feasible space of structural causal models and employs polynomial optimization to compute tight bounds on PN. We integrate domain-specific knowledge to further refine these bounds. A data-driven case study on breast cancer screening and treatment demonstrates the power of our framework. We show that incorporating domain-specific knowledge can reduce the width of PN bounds significantly, thus providing actionable insights. Our computational techniques are scalable, enabling high-quality solutions for time horizons as large as 100 periods within a few hours. Managerial implications: Our framework offers a rigorous tool for evaluating the impact of healthcare policies, such as requiring prior authorization or incorrectly denying a medical screening, on patient outcomes. The ability to estimate counterfactual probabilities has applications in policy and legal analysis and healthcare management. For instance, the bounds on PN can inform court cases by quantifying the impact of an (incorrect) delay-and-deny insurance policy on a particular patient who experienced a detrimental outcome. Beyond healthcare, our methodology is applicable to dynamic latent-state systems in other domains. History: This paper has been accepted in the Manufacturing & Service Operations Management Frontiers in Operations Initiative. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2025.0030 .
- 9 February 2026
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Manufacturing & Service Operations Management
Frontiers in Operations: Bounding Counterfactual Outcomes of Health Insurance Delay-and-Deny Practices
Problem definition: Health insurance delay-and-deny practices, such as requiring prior authorization, frequently restrict access to timely and necessary medical care. These policies can have severe consequences, including delayed diagnosis and treatment, leading to poor patient outcomes or even death. Counterfactual analysis offers a way to assess the impact of these practices by bounding the likelihood that a patient would have survived under alternative policies given the severe consequence under the original policy (probability of necessity). However, quantifying counterfactual probabilities is challenging in dynamic systems, such as those arising in healthcare, where disease progression is often latent and evolves over time. Methodology and results: We develop a principled framework to bound counterfactual probabilities, specifically the probability of necessity (PN), in generalized hidden Markov models (GHMMs). Our approach leverages the structure of GHMMs to construct a feasible space of structural causal models and employs polynomial optimization to compute tight bounds on PN. We integrate domain-specific knowledge to further refine these bounds. A data-driven case study on breast cancer screening and treatment demonstrates the power of our framework. We show that incorporating domain-specific knowledge can reduce the width of PN bounds significantly, thus providing actionable insights. Our computational techniques are scalable, enabling high-quality solutions for time horizons as large as 100 periods within a few hours. Managerial implications: Our framework offers a rigorous tool for evaluating the impact of healthcare policies, such as requiring prior authorization or incorrectly denying a medical screening, on patient outcomes. The ability to estimate counterfactual probabilities has applications in policy and legal analysis and healthcare management. For instance, the bounds on PN can inform court cases by quantifying the impact of an (incorrect) delay-and-deny insurance policy on a particular patient who experienced a detrimental outcome. Beyond healthcare, our methodology is applicable to dynamic latent-state systems in other domains. History: This paper has been accepted in the Manufacturing & Service Operations Management Frontiers in Operations Initiative. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2025.0030 .
- Open Access
- 2 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 540
Peer Review Market Design: Effort-Based Matching and Admission Control
- 2 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 540
Open AccessPeer Review Market Design: Effort-Based Matching and Admission Control
- 1 September 2024
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Manufacturing & Service Operations Management
- Vol. 26 (5) , 1768-1786
Workforce Scheduling with Heterogeneous Time Preferences: Effective Wages and Workers’ Supply
Problem definition: Motivated by the debate around workers’ welfare in the gig economy, we propose a framework to evaluate current practices and possible alternatives. We study a setting in which customers seek service from workers and a platform facilitates such matches over the course of the day. The platform allocates time slots to workers using an allocation policy, and the workers are strategic agents (with respect to “when to work”) who maximize their expected utility that depends on their preferred times to work, the allocated slots, and the total availability time. The platform seeks to ensure that a sufficient number of workers is available to satisfy demand, whereas the workers aim to maximize their wage-driven utility. Methodology/results: We evaluate policies on two dimensions critical to any firm: the supply of workers across the day, and the effective wages of workers. We illustrate that several families of currently deployed policies have serious limitations. We find these limitations exist because the policies do not let workers fully express their preferences and/or cannot account for heterogeneity in such preferences. We propose a new allocation policy and establish strong performance guarantees with respect to both the workers’ supply and effective wages. The policy is simple and fully leverages the market information to reach better market outcomes. We supplement our theory with numerical experiments in the context of ride-hailing calibrated on various New York City data sets that illustrate performance across a range of markets. Managerial implications: We highlight a fundamental inefficiency of policies currently deployed that limit workers’ ability to express their preferences. By allowing workers to express their temporal preferences, and by judiciously prioritizing “full-time” workers over “part-time” workers, we can obtain a potentially significant Pareto improvement, maintaining (or even increasing) workers’ supply while increasing their effective wages. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2022.0414 .
- 1 September 2024
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Manufacturing & Service Operations Management
- Vol. 26 (5) , 1768-1786
Workforce Scheduling with Heterogeneous Time Preferences: Effective Wages and Workers’ Supply
Problem definition: Motivated by the debate around workers’ welfare in the gig economy, we propose a framework to evaluate current practices and possible alternatives. We study a setting in which customers seek service from workers and a platform facilitates such matches over the course of the day. The platform allocates time slots to workers using an allocation policy, and the workers are strategic agents (with respect to “when to work”) who maximize their expected utility that depends on their preferred times to work, the allocated slots, and the total availability time. The platform seeks to ensure that a sufficient number of workers is available to satisfy demand, whereas the workers aim to maximize their wage-driven utility. Methodology/results: We evaluate policies on two dimensions critical to any firm: the supply of workers across the day, and the effective wages of workers. We illustrate that several families of currently deployed policies have serious limitations. We find these limitations exist because the policies do not let workers fully express their preferences and/or cannot account for heterogeneity in such preferences. We propose a new allocation policy and establish strong performance guarantees with respect to both the workers’ supply and effective wages. The policy is simple and fully leverages the market information to reach better market outcomes. We supplement our theory with numerical experiments in the context of ride-hailing calibrated on various New York City data sets that illustrate performance across a range of markets. Managerial implications: We highlight a fundamental inefficiency of policies currently deployed that limit workers’ ability to express their preferences. By allowing workers to express their temporal preferences, and by judiciously prioritizing “full-time” workers over “part-time” workers, we can obtain a potentially significant Pareto improvement, maintaining (or even increasing) workers’ supply while increasing their effective wages. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2022.0414 .
- 8 March 2024
- Published by SAGE Publications in Production and Operations Management
- Vol. 33 (3) , 775-794
Model-Free Approximate Bayesian Learning for Large-Scale Conversion Funnel Optimization
The flexibility of choosing the ad action as a function of the consumer state is critical for modern-day marketing campaigns. We study the problem of identifying the optimal sequential personalized interventions that maximize the adoption probability for a new product. We model consumer behavior by a conversion funnel that captures the state of each consumer (e.g., interaction history with the firm) and allows the consumer behavior to vary as a function of both her state and firm’s sequential interventions. We show our model captures consumer behavior with very high accuracy (out-of-sample area under the curve of over 0.95) in a real-world email marketing dataset. However, it results in a very large-scale learning problem, where the firm must learn the state-specific effects of various interventions from consumer interactions. We propose a novel attribution-based decision-making algorithm for this problem that we call model-free approximate Bayesian learning. Our algorithm inherits the interpretability and scalability of Thompson sampling for bandits and maintains an approximate belief over the value of each state-specific intervention. The belief is updated as the algorithm interacts with the consumers. Despite being an approximation to the Bayes update, we prove the asymptotic optimality of our algorithm and analyze its convergence rate. We show that our algorithm significantly outperforms traditional approaches on extensive simulations calibrated to a real-world email marketing dataset.
- 8 March 2024
- Published by SAGE Publications in Production and Operations Management
- Vol. 33 (3) , 775-794
Model-Free Approximate Bayesian Learning for Large-Scale Conversion Funnel Optimization
The flexibility of choosing the ad action as a function of the consumer state is critical for modern-day marketing campaigns. We study the problem of identifying the optimal sequential personalized interventions that maximize the adoption probability for a new product. We model consumer behavior by a conversion funnel that captures the state of each consumer (e.g., interaction history with the firm) and allows the consumer behavior to vary as a function of both her state and firm’s sequential interventions. We show our model captures consumer behavior with very high accuracy (out-of-sample area under the curve of over 0.95) in a real-world email marketing dataset. However, it results in a very large-scale learning problem, where the firm must learn the state-specific effects of various interventions from consumer interactions. We propose a novel attribution-based decision-making algorithm for this problem that we call model-free approximate Bayesian learning. Our algorithm inherits the interpretability and scalability of Thompson sampling for bandits and maintains an approximate belief over the value of each state-specific intervention. The belief is updated as the algorithm interacts with the consumers. Despite being an approximation to the Bayes update, we prove the asymptotic optimality of our algorithm and analyze its convergence rate. We show that our algorithm significantly outperforms traditional approaches on extensive simulations calibrated to a real-world email marketing dataset.
- Open Access
- 1 January 2023
- Published by Elsevier in SSRN Electronic Journal
Bounding Counterfactuals in Hidden Markov Models and Beyond
Answering a counterfactual query (CQ) in a hidden Markov model (HMM) requires embedding the HMM in a so-called structural causal model (SCM). In general, however, we don't have sufficient information to fully specify the SCM and so we cannot provide an exact answer to the CQ. Nonetheless, the known structure of the given HMM implies we should only consider SCMs that have no hidden confounding. We use this observation to optimize over the space of feasible SCMs to compute upper and lower bounds on the CQ and to the best of our knowledge, we are the first to compute such bounds in a HMM. We consider two applications of our framework. The first application concerns the well-known dishonest casino HMM. In this case, the CQ is the casino’s expected winnings that are attributable to cheating (EWAC) and we show that bounds on the EWAC can be computed as the solutions to simple linear programs. Our second application concerns a patient who died from breast cancer after being denied regular screenings by her insurance company. We propose an extended HMM to model the problem dynamics and we bound the CQ - the so-called probability of necessity - by formulating and solving polynomial programs. We also show how informative bounds can still be computed even when the original polynomial programs are too difficult to solve. Finally, we disprove a recent conjecture suggesting that the Gumbel-max distribution is the unique mechanism for modeling SCMs with the so-called counterfactual stability (CS) property.
- 1 January 2023
- Published by Elsevier in SSRN Electronic Journal
Open AccessBounding Counterfactuals in Hidden Markov Models and Beyond
Answering a counterfactual query (CQ) in a hidden Markov model (HMM) requires embedding the HMM in a so-called structural causal model (SCM). In general, however, we don't have sufficient information to fully specify the SCM and so we cannot provide an exact answer to the CQ. Nonetheless, the known structure of the given HMM implies we should only consider SCMs that have no hidden confounding. We use this observation to optimize over the space of feasible SCMs to compute upper and lower bounds on the CQ and to the best of our knowledge, we are the first to compute such bounds in a HMM. We consider two applications of our framework. The first application concerns the well-known dishonest casino HMM. In this case, the CQ is the casino’s expected winnings that are attributable to cheating (EWAC) and we show that bounds on the EWAC can be computed as the solutions to simple linear programs. Our second application concerns a patient who died from breast cancer after being denied regular screenings by her insurance company. We propose an extended HMM to model the problem dynamics and we bound the CQ - the so-called probability of necessity - by formulating and solving polynomial programs. We also show how informative bounds can still be computed even when the original polynomial programs are too difficult to solve. Finally, we disprove a recent conjecture suggesting that the Gumbel-max distribution is the unique mechanism for modeling SCMs with the so-called counterfactual stability (CS) property.
- 1 October 2022
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Management Science
- Vol. 68 (10) , 7457-7479
Shapley Meets Uniform: An Axiomatic Framework for Attribution in Online Advertising
One of the central challenges in online advertising is attribution, namely, assessing the contribution of individual advertiser actions such as emails, display ads, and search ads to eventual conversion. Several heuristics are used for attribution in practice; however, most do not have any formal justification. The main contribution in this work is to propose an axiomatic framework for attribution in online advertising. We show that the most common heuristics can be cast under the framework and illustrate how these may fail. We propose a novel attribution metric, which we refer to as counterfactual adjusted Shapley value (CASV), which inherits the desirable properties of the traditional Shapley value while overcoming its shortcomings in the online advertising context. We also propose a Markovian model for the user journey through the conversion funnel, in which ad actions may have disparate impacts at different stages. We use the Markovian model to compare our metric with commonly used metrics. Furthermore, under the Markovian model, we establish that the CASV metric coincides with an adjusted “unique-uniform” attribution scheme. This scheme is efficiently implementable and can be interpreted as a correction to the commonly used uniform attribution scheme. We supplement our theoretical developments with numerical experiments using a real-world large-scale data set. This paper was accepted by David Simchi-Levi, revenue management and market analytics.
- 1 October 2022
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Management Science
- Vol. 68 (10) , 7457-7479
Shapley Meets Uniform: An Axiomatic Framework for Attribution in Online Advertising
One of the central challenges in online advertising is attribution, namely, assessing the contribution of individual advertiser actions such as emails, display ads, and search ads to eventual conversion. Several heuristics are used for attribution in practice; however, most do not have any formal justification. The main contribution in this work is to propose an axiomatic framework for attribution in online advertising. We show that the most common heuristics can be cast under the framework and illustrate how these may fail. We propose a novel attribution metric, which we refer to as counterfactual adjusted Shapley value (CASV), which inherits the desirable properties of the traditional Shapley value while overcoming its shortcomings in the online advertising context. We also propose a Markovian model for the user journey through the conversion funnel, in which ad actions may have disparate impacts at different stages. We use the Markovian model to compare our metric with commonly used metrics. Furthermore, under the Markovian model, we establish that the CASV metric coincides with an adjusted “unique-uniform” attribution scheme. This scheme is efficiently implementable and can be interpreted as a correction to the commonly used uniform attribution scheme. We supplement our theoretical developments with numerical experiments using a real-world large-scale data set. This paper was accepted by David Simchi-Levi, revenue management and market analytics.
- 1 May 2021
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 69 (3) , 950-973
A Finite Time Analysis of Temporal Difference Learning with Linear Function Approximation
Temporal difference learning (TD) is a simple iterative algorithm widely used for policy evaluation in Markov reward processes. Bhandari et al. prove finite time convergence rates for TD learning with linear function approximation. The analysis follows using a key insight that establishes rigorous connections between TD updates and those of online gradient descent. In a model where observations are corrupted by i.i.d. noise, convergence results for TD follow by essentially mirroring the analysis for online gradient descent. Using an information-theoretic technique, the authors also provide results for the case when TD is applied to a single Markovian data stream where the algorithm’s updates can be severely biased. Their analysis seamlessly extends to the study of TD learning with eligibility traces and Q-learning for high-dimensional optimal stopping problems.
- 1 May 2021
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Operations Research
- Vol. 69 (3) , 950-973
A Finite Time Analysis of Temporal Difference Learning with Linear Function Approximation
Temporal difference learning (TD) is a simple iterative algorithm widely used for policy evaluation in Markov reward processes. Bhandari et al. prove finite time convergence rates for TD learning with linear function approximation. The analysis follows using a key insight that establishes rigorous connections between TD updates and those of online gradient descent. In a model where observations are corrupted by i.i.d. noise, convergence results for TD follow by essentially mirroring the analysis for online gradient descent. Using an information-theoretic technique, the authors also provide results for the case when TD is applied to a single Markovian data stream where the algorithm’s updates can be severely biased. Their analysis seamlessly extends to the study of TD learning with eligibility traces and Q-learning for high-dimensional optimal stopping problems.
- 1 January 2021
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Management Science
- Vol. 67 (1) , 72-92
How to Play Fantasy Sports Strategically (and Win)
Daily fantasy sports (DFS) is a multibillion-dollar industry with millions of annual users and widespread appeal among sports fans across a broad range of popular sports. Building on recent work, we provide a coherent framework for constructing DFS portfolios where we explicitly model the behavior of other DFS players. We formulate an optimization problem that accurately describes the DFS problem for a risk-neutral decision maker in both double-up and top-heavy payoff settings. Our formulation maximizes the expected reward subject to feasibility constraints, and we relate this formulation to mean-variance optimization and the outperformance of stochastic benchmarks. Using this connection, we show how the problem can be reduced to the problem of solving a series of binary quadratic programs. We also propose an algorithm for solving the problem where the decision maker can submit multiple entries to the DFS contest. This algorithm is motivated by submodularity properties of the objective function and by some new results on parimutuel betting. One of the contributions of our work is the introduction of a Dirichlet-multinomial data-generating process for modeling opponents’ team selections, and we estimate the parameters of this model via Dirichlet regressions. A further benefit to modeling opponents’ team selections is that it enables us to estimate the value, in a DFS setting, of both insider trading and collusion. We demonstrate the value of our framework by applying it to DFS contests during the 2017 National Football League season. This paper was accepted by Baris Ata, stochastic models and simulation.
- 1 January 2021
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Management Science
- Vol. 67 (1) , 72-92
How to Play Fantasy Sports Strategically (and Win)
Daily fantasy sports (DFS) is a multibillion-dollar industry with millions of annual users and widespread appeal among sports fans across a broad range of popular sports. Building on recent work, we provide a coherent framework for constructing DFS portfolios where we explicitly model the behavior of other DFS players. We formulate an optimization problem that accurately describes the DFS problem for a risk-neutral decision maker in both double-up and top-heavy payoff settings. Our formulation maximizes the expected reward subject to feasibility constraints, and we relate this formulation to mean-variance optimization and the outperformance of stochastic benchmarks. Using this connection, we show how the problem can be reduced to the problem of solving a series of binary quadratic programs. We also propose an algorithm for solving the problem where the decision maker can submit multiple entries to the DFS contest. This algorithm is motivated by submodularity properties of the objective function and by some new results on parimutuel betting. One of the contributions of our work is the introduction of a Dirichlet-multinomial data-generating process for modeling opponents’ team selections, and we estimate the parameters of this model via Dirichlet regressions. A further benefit to modeling opponents’ team selections is that it enables us to estimate the value, in a DFS setting, of both insider trading and collusion. We demonstrate the value of our framework by applying it to DFS contests during the 2017 National Football League season. This paper was accepted by Baris Ata, stochastic models and simulation.
- Open Access
- 23 May 2019
- Published by Elsevier in SSRN Electronic Journal
How to Play Fantasy Sports Strategically (and Win)
Daily Fantasy Sports (DFS) is a multi-billion dollar industry with millions of annual users and widespread appeal among sports fans across a broad range of popular sports. Building on the recent work of Hunter, Vielma and Zaman (2016), we provide a coherent framework for constructing DFS portfolios where we explicitly model the behavior of other DFS players. We formulate an optimization problem that accurately describes the DFS problem for a risk-neutral decision-maker in both double-up and top-heavy payoff settings. Our formulation maximizes the expected reward subject to feasibility constraints and we relate this formulation to mean-variance optimization and the out-performance of stochastic benchmarks. Using this connection, we show how the problem can be reduced to the problem of solving a series of binary quadratic programs. We also propose an algorithm for solving the problem where the decision-maker can submit multiple entries to the DFS contest. This algorithm is motivated in part by some new results on parimutuel betting which can be viewed as a special case of a DFS contest. One of the contributions of our work is the introduction of a Dirichlet-multinomial data generating process for modeling opponents' team selections and we estimate the parameters of this model via Dirichlet regressions. A further benefit to modeling opponents' team selections is that it enables us to estimate the value in a DFS setting of both insider trading and and collusion. We demonstrate the value of our framework by applying it to DFS contests during the 2017 NFL season.
- 23 May 2019
- Published by Elsevier in SSRN Electronic Journal
Open AccessHow to Play Fantasy Sports Strategically (and Win)
Daily Fantasy Sports (DFS) is a multi-billion dollar industry with millions of annual users and widespread appeal among sports fans across a broad range of popular sports. Building on the recent work of Hunter, Vielma and Zaman (2016), we provide a coherent framework for constructing DFS portfolios where we explicitly model the behavior of other DFS players. We formulate an optimization problem that accurately describes the DFS problem for a risk-neutral decision-maker in both double-up and top-heavy payoff settings. Our formulation maximizes the expected reward subject to feasibility constraints and we relate this formulation to mean-variance optimization and the out-performance of stochastic benchmarks. Using this connection, we show how the problem can be reduced to the problem of solving a series of binary quadratic programs. We also propose an algorithm for solving the problem where the decision-maker can submit multiple entries to the DFS contest. This algorithm is motivated in part by some new results on parimutuel betting which can be viewed as a special case of a DFS contest. One of the contributions of our work is the introduction of a Dirichlet-multinomial data generating process for modeling opponents' team selections and we estimate the parameters of this model via Dirichlet regressions. A further benefit to modeling opponents' team selections is that it enables us to estimate the value in a DFS setting of both insider trading and and collusion. We demonstrate the value of our framework by applying it to DFS contests during the 2017 NFL season.