Optimal Contest Design with Entry Restriction

Hanbing Liu, Ningyuan Li, Weian Li, Qi Qi, Changyuan Yu · 2025

Read Optimal Contest Design with Entry Restriction by Hanbing Liu, Ningyuan Li, Weian Li, Qi Qi, Changyuan Yu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-09014.txt.

Introduction

Hanbing Liu, Ningyuan Li, Weian Li, Qi Qi, and Changyuan Yu wrote this paper in two thousand twenty-five. It designs contests where the host sets both who may enter and how the prize purse is split.

With a fixed purse, the host picks the field size and the prize shape to pull the most effort. The authors pin down the one symmetric Bayesian Nash equilibrium of the players. They then solve the best design twice: once for the top single effort, once for the sum of all efforts. The paper is a piece of game theory for prize design.

Catalog record

CREATORHanbing Liu, Ningyuan Li, Weian Li, Qi Qi, Changyuan Yu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE160 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-09014

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Optimal Contest Design with Entry Restriction Hanbing Liu, Ningyuan Li, Weian Li, Qi Qi, Changyuan Yu Abstract This paper explores the design of contests involving n n contestants, focusing on how the designer decides on the number of contestants allowed and the prize structure with a fixed budget. We characterize the unique symmetric Bayesian Nash equilibrium of contestants and find the optimal contests design for the maximum individual effort objective and the total effort objective. A Preprint 1 Introduction In real-world competitive scenarios, shortlists are commonly used. Specifically, organizers do not admit all registered candidates into the final contest but instead select a subset of candidates to form a shortlist, eliminating others. For example, in the Olympic Games, only top-ranked athletes qualify for the final competition due to scheduling constraints. Similarly, in the International Collegiate Programming Contest, the offline finals have limited venue capacity, so only teams meeting specific requirements advance. A similar scenario is observed in corporate recruitment, where only applicants who surpass a certain threshold in their resumes are invited for interviews. This raises an important question: does a shortlist benefit the organizer, and how does it affect the competitive results? To answer this question, we use the setting of a ‘‘contest’’ as a basic competitive (or game) model and examine the impact of a shortlist on contest design. Contests are a fundamental topic in mechanism design within economics, where a contest designer determines the prize structure and winning rules to achieve specific objectives. Contestants, in turn, invest irreversible and costly efforts to compete under these rules. Since contests effectively model competitive scenarios, they have widespread real-world applications, including sports, programming competitions, resource allocation and so on. A rank-order contest is a common contest format in which prizes are allocated strictly based on contestants’ performance 1 1 1 The term “performance” is an abstract concept that varies by context. For example, in some cases, it may refer to measurable metrics such as race completion time in the Olympic 100-meter sprint or bid prices in an auction. . Specifically, the contestant with the best performance wins the first prize, the second-best wins the second prize, and so on. From a theoretical perspective, research on rank-order contests primarily focuses on characterizing contestant equilibrium in performance and optimizing contest design to achieve various objectives, such as the total effort Moldovanu and Sela (2001) , the maximum individual effort Chawla et al. (2019) , or other criteria Archak and Sundararajan (2009) ; Elkind et al. (2021) . Broadly speaking, the existence of a shortlist can incentivize admitted contestants to exert more effort, as they anticipate facing stronger competition. However, eliminating too many contestants may reduce motivation, as fewer opponents make it easier to secure a desirable prize. Thus, determining the optimal shortlist size is a crucial problem. To address this, several key questions must be considered sequentially. First, we need to understand how the shortlist size affects contestants’ performance under any given prize structure. Once this relationship is established, the next step is to jointly determine the optimal shortlist size and prize structure to achieve various objectives. This is particularly challenging for the total effort objective, as there is a tradeoff between the number of admitted contestants and the incentive effect. Intuitively, reducing the number of contestants increases individual effort, but total effort may not follow a simple monotonic pattern. Finally, it is essential to analyze how the presence of a shortlist impacts the designer’s objective by comparing the optimal contests with and without a shortlist. 1.1 Main Results Model. To address the aforementioned questions, we first construct a model of a contest with a shortlist. Consider a rank-order contest with n\geq 2 initially registered contestants. Each contestant’s ability, x_{i} , is private information drawn from a publicly known distribution F(x) . We assume that the contest designer is aware of all contestants’ abilities upon registration. This assumption is reasonable, as designers can assess abilities based on submitted registration materials, such as competition records, CVs, or asset certificates. The contest proceeds in two stages. First, the designer selects the number of shortlisted contestants, 2\leq m\leq n , and determines the prize structure \vec{V}=(V_{1},V_{2},\cdots,V_{m})\in\mathbb{R}_{1\times m} , specifying the rewards for each rank, within a budget B . The designer then ranks all n contestants by ability in non-increasing order, admitting the top m contestants to the next round while eliminating the remaining n-m contestants. In the second stage, the shortlisted contestants observe the prize structure and recognize their eligibility. They then exert costly effort to compete for the prizes, with their effort levels influenced by their abilities. Each admitted contestant aims to maximize the utility (which is equal to the obtained prize minus the cost incurred by effort) by choosing the level of effort exerted. The contest designer’s decision variables consist of two components: the number of shortlisted contestants m and the prize structure \vec{V} , to optimize two kinds of objectives, the maximum individual effort and the total effort. Based on this contest-theoretical model, our main results are divided into three parts: 1.1.1 Posterior distribution and equilibrium effort Our first result characterizes contestant equilibrium under any shortlist and prize structure. We use the symmetric Bayesian Nash equilibrium (sBNE) to define the contest’s solution, and can fully characterize the equilibrium effort function. Theorem 1 (Unique sBNE of Contestants (Sketch)) . For any ability distribution F , any size of shortlist m and any prize structure \vec{V} , the unique symmetric Bayesian Nash equilibrium exists and can be expressed in a closed-form. The sBNE is typically derived using the first-order condition of a contestant’s expected utility function. A key component of this function is the probability of ranking in each position, which, in traditional rank-order contests, is determined by the prior distribution of abilities. However, in our model, the elimination stage alters contestants’ beliefs, i.e., admitted contestants receive a signal of eligibility and update their beliefs about opponents from the prior to a posterior distribution. Therefore, the main challenge in deriving this equilibrium is accurately computing an admitted contestant’s posterior belief about opponents (see Propositions 1 and Corollary 1 for details). On the other hand, since contestants with different abilities form distinct posterior beliefs, calculating the sBNE in our model is significantly more complex due to the non-identical posterior distributions, compared to the traditional settings. 1.1.2 Optimal Contest Design With a detailed characterization of contestant equilibrium, we next address the problem of the optimal contest design under two common objectives: the maximum individual effort and the total effort. In the traditional rank-order contest design problem, determining the optimal prize structure is already complex due to its high-dimension. In our work, we not only determine the optimal size of shortlist but also design the optimal prize structure within a fixed budget, making the problem more complex. However, we provide a fundamentally different characterization of the optimal contest, as detailed below. Before presenting the detailed description of the optimal contests, we introduce a technical guideline to simplify our analysis. Specifically, we focus on a special type of contest, referred to as a “simple contest”, where the designer allocates the budget equally across the prizes. Interestingly, we identify two sufficient conditions (see as Proposition 6 ) under

Read the complete text