Skip to main content
Log in

Resource allocations in the best-of-k (\(k=2,3\)) contests

  • Published:
Journal of Economics Aims and scope Submit manuscript

Abstract

We study best-of-k (\(k=2,3\)) contests with two players who have heterogeneous resource budgets that decrease within the stages proportionally to the resource allocated in the previous stages such that for each resource unit that a player allocates, he loses \(\alpha\) (the fatigue parameter) units of resource from his budget. We first analyze the players’ resource allocations and show that if the resource budget is the same, although the recycled resource is larger in the best-of-three contest than in the best-of-two contest, the total resource allocation in the best-of-two contest with the optimal asymmetric tie-breaking rule might be larger than in the best-of-three contest.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

Notes

  1. There are opposite opinions such as those of Ferral and Smith (1999) who used data from professional sport leagues in the U.S. and showed that teams do not strategically allocate their resources but instead exert as much resources as possible in each of the stages.

  2. Konrad (2004) also analyzed the allocation of resources in sequential contests.

  3. It can be verified that the SOC of the maximization problems in this section are satsfied.

  4. It can be verified that the SOC of the maximization problems in this section are satsfied.

References

  • Amegashie J, Cadsby C, Song Y (2007) Competitive burnout: theory and experimental evidence. Games Econ Behav 59:213–239

    Article  Google Scholar 

  • Apesteguia J, Palacios-Huerta I (2010) Psychological pressure in competitive environments: evidence from a randomized natural experiment. Am Econ Rev 100(5):2548–2564

    Article  Google Scholar 

  • Cohen C, Sela A (2007) Contests with ties. B.E. J Theor Econ 7(1):43

    Article  Google Scholar 

  • Ferral C, Smith A (1999) A sequential game model of sports championship series: theory and estimation. Rev Econ Stat 81:704–719

    Article  Google Scholar 

  • Fudenberg D, Gilbert R, Stiglitz J, Tirole J (1983) Preemption, leapfrogging and competition in patent races. Eur Econ Rev 22:3–31

    Article  Google Scholar 

  • Harbaugh R, Klumpp T (2005) Early round upsets and championship blowouts. Econ Inq 43:316–332

    Article  Google Scholar 

  • Harris C, Vickers J (1985) Perfect equilibrium in a model of a race. Rev Econ Stud 52(2):193–209

    Article  Google Scholar 

  • Harris C, Vickers J (1987) Racing with uncertainty. Rev Econ Stud 54(1):1–21

    Article  Google Scholar 

  • Hart S (2008) Discrete Colonel Blotto and general Lotto games. Int J Game Theory 36:441–460

    Article  Google Scholar 

  • Klumpp T, Polborn M (2006) Primaries and the New Hampshire effect. J Public Econ 90:1073–1114

    Article  Google Scholar 

  • Klumpp T, Konrad K, Solomon A (2019) The dynamics of majoritarian Blotto games. Games Econ Behav 117:402–419

    Article  Google Scholar 

  • Konrad KA (2004) Bidding in hierarchies. Eur Econ Rev 48:1301–1308

    Article  Google Scholar 

  • Konrad K, Kovenock D (2009) Multi-battle contests. Games Econ Behav 66:256–274

    Article  Google Scholar 

  • Kovenock D, Roberson B (2009) Is the 50-state strategy optimal? J Theor Polit 21(2):213–236

    Article  Google Scholar 

  • Kovenock D, Roberson B (2021) Generalization of the general Lotto and Colonel Blotto games. Econ Theory 71:997–1032

    Article  Google Scholar 

  • Krumer A (2013) Best-of-two contests with psychological effects. Theory Decis 75(1):85–100

    Article  Google Scholar 

  • Krumer A (2015) The order of games in a best-of-three contest. J Sports Econ 16(2):185–200

    Article  Google Scholar 

  • Kvasov D (2007) Contests with limited resources. J Econ Theory 136:738–748

    Article  Google Scholar 

  • Lien D (1990) Corruption and allocation efficiency. J Dev Econ 33:153–164

    Article  Google Scholar 

  • Mago S, Sheremeta R, Yates A (2013) Best-of-three contest experiments: strategic versus psychological momentum. Int J Ind Organ 31(3):287–296

    Article  Google Scholar 

  • Malueg D, Yates A (2006) Equilibria in rent-seeking contests with homogeneous success functions. Econ Theory 27(3):719–727

    Article  Google Scholar 

  • Malueg D, Yates A (2010) Testing contest theory: evidence from best-of-three tennis matches. Rev Econ Stat 92:689–692

    Article  Google Scholar 

  • Roberson B (2006) The Colonel Blotto game. Econ Theory 29:1–24

    Article  Google Scholar 

  • Ryvkin D (2011) Fatigue in dynamic tournaments. J Econ Manag Strat 20(4):1011–1041

    Article  Google Scholar 

  • Sela A (2011) Best-of-three all-pay auctions. Econ Lett 112(1):67–70

    Article  Google Scholar 

  • Sela A (2017) Two-stage contests with effort-dependent values of winning. Rev Econ Des 21:253–272

    Google Scholar 

  • Sela A, Erez E (2013) Dynamic contests with resource constraints. Soc Choice Welf 41(4):863–882

    Article  Google Scholar 

  • Sela A, Tsachi O (2020) On the optimal allocation of prizes in best-of-three all-pay auctions. Soc Choice Welf 55:255–273

    Article  Google Scholar 

  • Snyder JM (1989) Election goals and the allocation of campaign resources. Econometrica 57:637–660

    Article  Google Scholar 

  • Szech N (2015) Tie-breaks and bid-caps in all-pay auctions. Games Econ Behav 92:138–149

    Article  Google Scholar 

  • Szymanski S (2003) The economic design of sporting contests. J Econ Lit 41(4):1137–1187

    Article  Google Scholar 

  • Tullock G (1980) Efficient rent-seeking. In: Buchanan JM, Tollison RD, Tullock G (eds) Toward a theory of rent-seeking society. Texas A&M University Press, College Station

    Google Scholar 

  • Zizzo DJ (2002) Racing with uncertainty: a patent race experiment. Int J Ind Organ 20:877–902

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Aner Sela.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

The previous title of this paper was: “Resource Allocations in Multi-Stage Contests.”.

Appendix

Appendix

1.1 Proof of Proposition 1

By (8) and (10), when \(\frac{\delta }{(1-\delta )}\ge \alpha \ge \delta ,\)

$$\begin{aligned} x_{1} & = \delta \frac{v_{1}}{\alpha }\ge (1-\delta )v_{1}=x_{2} \\ y_{1} & = \delta \frac{w_{1}}{\alpha }\ge (1-\delta )w_{1}=y_{2}. \end{aligned}$$

Otherwise, if \(\alpha >\frac{\delta }{(1-\delta )}\), we obtain that \(x_{1}<x_{2}\) and \(y_{1}<y_{2}.\) Similarly, by (9) and (10), when \(\alpha<\) \(\frac{\delta }{(1-\delta )},\)

$$\begin{aligned} x_{1} & = v_{1}>v_{1}(1-\alpha )=x_{2} \\ y_{1} & = w_{1}>w_{1}(1-\alpha )=y_{2}. \end{aligned}$$

Thus, if \(\alpha >\frac{\delta }{(1-\delta )}\), then \(x_{1}<x_{2}\) and \(y_{1}<y_{2}\), and if \(\alpha <\frac{\delta }{(1-\delta )}\), then \(x_{1}>x_{2}\) and \(y_{1}>y_{2}.\) In particular, if \(\delta \ge 0.5\), we obtain that \(x_{1}>x_{2}\) and \(y_{1}>y_{2}.\) \(\ Q.E.D.\)

1.2 Proof of Proposition 3

Assume that \(r=1\). When we divide the equations (15) by each other, we obtain that

$$\begin{aligned} \frac{\alpha _{1}x_{1}}{\alpha _{2}y_{1}}=\frac{(v_{1}-\alpha _{1}x_{1})}{ (w_{1}-\alpha _{2}y_{1})}=\frac{v_{1}}{w_{1}}, \end{aligned}$$
(39)

and when we insert (39) into (15) we obtain that

$$\begin{aligned} & \delta \frac{x_{1}}{(x_{1}+y_{1})^{2}}-(1-\delta )\frac{\alpha _{1}x_{1}}{ \alpha _{1}x_{1}+\alpha _{2}y_{1}}\frac{\alpha _{2}}{(v_{1}-\alpha _{1}x_{1}+w_{1}-\alpha _{2}y_{1})} \\ & \quad =\frac{x_{1}}{\left( x_{1}+y_{1}\right) ^{2}\left( \alpha _{1}x_{1}+\alpha _{2}y_{1}\right) \left( v_{1}+w_{1}-\alpha _{1}x_{1}-\alpha _{2}y_{1}\right) }\cdot \\ & \qquad \left( \begin{array}{c} \delta \alpha _{1}\alpha _{2}x_{1}^{2}-\alpha _{1}\alpha _{2}y_{1}^{2}-\delta \alpha _{1}^{2}x_{1}^{2}-\delta \alpha _{2}^{2}y_{1}^{2}-\alpha _{1}\alpha _{2}x_{1}^{2}+\delta \alpha _{1}\alpha _{2}y_{1}^{2}+\delta \alpha _{1}v_{1}x_{1} \\ +\delta \alpha _{1}w_{1}x_{1}+\delta \alpha _{2}v_{1}y_{1}+\delta \alpha _{2}w_{1}y_{1}-2\alpha _{1}\alpha _{2}x_{1}y_{1} \end{array} \right) \\ & \quad =0 \end{aligned}$$

Since \(y_{1}=\frac{w_{1}\alpha _{1}x_{1}}{v_{1}\alpha _{2}}\), we obtain

$$\begin{aligned} & \delta \alpha _{1}\alpha _{2}x_{1}^{2}-\alpha _{1}\alpha _{2}\left( \frac{ w_{1}\alpha _{1}x_{1}}{v_{1}\alpha _{2}}\right) ^{2}-\delta \alpha _{1}^{2}x_{1}^{2}-\delta \alpha _{2}^{2}\left( \frac{w_{1}\alpha _{1}x_{1}}{ v_{1}\alpha _{2}}\right) ^{2}-\alpha _{1}\alpha _{2}x_{1}^{2}\\ & \qquad +\delta \alpha _{1}\alpha _{2}\left( \frac{w_{1}\alpha _{1}x_{1}}{v_{1}\alpha _{2}}\right) ^{2} \\ & \qquad +\delta \alpha _{1}v_{1}x_{1}+\delta \alpha _{1}w_{1}x_{1}+\delta \alpha _{2}v\left( \frac{w_{1}\alpha _{1}x_{1}}{v_{1}\alpha _{2}}\right) +\delta \alpha _{2}w_{1}\left( \frac{w_{1}\alpha _{1}x_{1}}{v_{1}\alpha _{2}}\right) \\ & \qquad -2\alpha _{1}\alpha _{2}x_{1}\left( \frac{w_{1}\alpha _{1}x_{1}}{v_{1}\alpha _{2}}\right) \\ & \quad =0. \end{aligned}$$

The solution of the last equation gives us player 1’s resource allocation,

$$\begin{aligned} x_{1}=\frac{\delta \alpha _{2}v_{1}\left( v_{1}+w_{1}\right) ^{2}}{\alpha _{1}\alpha _{2}(\delta v_{1}^{2}+\delta w_{1}^{2})+(1-\delta )((v_{1}\alpha _{2})^{2}+(w_{1}\alpha _{1})^{2})+2\alpha _{1}\alpha _{2}v_{1}w_{1}}, \end{aligned}$$

and then player 2’s resource allocation is

$$\begin{aligned} y_{1}=\frac{\delta \alpha _{1}w_{1}\left( v_{1}+w_{1}\right) ^{2}}{\alpha _{1}\alpha _{2}(\delta v_{1}^{2}+\delta w_{1}^{2})+(1-\delta )((v_{1}\alpha _{2})^{2}+(w_{1}\alpha _{1})^{2})+2\alpha _{1}\alpha _{2}v_{1}w_{1}}. \end{aligned}$$

Note that when \(\alpha =\alpha _{1}=\alpha _{2}\), we obtain the symmetric solution given by (8) as follows:

$$\begin{aligned} x_{1}=\frac{\delta v_{1}}{\alpha _{1}},{ }y_{1}=\frac{\delta w_{1}}{ \alpha _{2}}. \end{aligned}$$

Now, assume that \(a_{2}\ge a_{1}\). Then we have

$$\begin{aligned} x_{1}\ge \frac{\delta \alpha _{2}v_{1}\left( v_{1}+w_{1}\right) ^{2}}{ (\alpha _{2})^{2}(\delta v_{1}^{2}+\delta w_{1}^{2})+(1-\delta )(\alpha _{2})^{2}((v_{1})^{2}+(w_{1})^{2})+2(\alpha _{2})^{2}v_{1}w_{1}}=\frac{ \delta v_{1}}{\alpha _{2}}, \end{aligned}$$

and

$$\begin{aligned} x_{1}\le \frac{\delta \alpha _{2}v_{1}\left( v_{1}+w_{1}\right) ^{2}}{ (\alpha _{1})^{2}(\delta v_{1}^{2}+\delta w_{1}^{2})+(1-\delta )(\alpha _{1})^{2}((v_{1})^{2}+(w_{1})^{2})+2(\alpha _{1})^{2}v_{1}w_{1}}=\frac{ \alpha _{2}\delta v_{1}}{(\alpha _{1})^{2}}. \end{aligned}$$

Similarly,

$$\begin{aligned} y_{1}\ge \frac{\delta \alpha _{1}w_{1}\left( v_{1}+w_{1}\right) ^{2}}{ (\alpha _{2})^{2}(\delta v_{1}^{2}+\delta w_{1}^{2})+(1-\delta )(\alpha _{2})^{2}((v_{1})^{2}+(w_{1})^{2})+2(\alpha _{2})^{2}v_{1}w_{1}}=\frac{ \alpha _{1}\delta w_{1}}{(\alpha _{2})^{2}}, \end{aligned}$$

and

$$\begin{aligned} y_{1}\le \frac{\delta \alpha _{1}w_{1}\left( v_{1}+w_{1}\right) ^{2}}{ (\alpha _{1})^{2}(\delta v_{1}^{2}+\delta w_{1}^{2})+(1-\delta )(\alpha _{1})^{2}((v_{1})^{2}+(w_{1})^{2})+2(\alpha _{1})^{2}v_{1}w_{1}}=\frac{ \delta w_{1}}{\alpha _{1}}. \end{aligned}$$

Thus, if \(\frac{\alpha _{2}\delta }{(\alpha _{1})^{2}}\le 1\), the resource allocation of both players in the first stage are given by (16) and if \(\frac{\alpha _{1}\delta }{ (\alpha _{2})^{2}}>1\), both players allocate their resource budgets in the first stage. Q.E.D.

1.3 Proof of Proposition 4

According to the players’ resource allocations in the first stage given by (31), (32), and in the second stage given by (33), (34), (35), and (36), we have

  1. 1.

    If \(\alpha \le \frac{1}{3}\), and \(\beta \le \frac{1}{2}\)

    $$\begin{aligned} x_{1} & = v_{1}>x_{2}=v_{1}(1-\alpha )>x_{3}=v_{1}(1-\alpha )(1-\beta ) \\ y_{1} & = w_{1}>y_{2}=w_{1}(1-\alpha )>y_{3}=w_{1}(1-\alpha )(1-\beta ). \end{aligned}$$
  2. 2.

    If \(\alpha \le \frac{1}{3}\), and \(\beta >\frac{1}{2}\)

    $$\begin{aligned} x_{1} & = v_{1}>x_{2}=\frac{v_{1}(1-\alpha )}{2\beta }>x_{3}=\frac{ v_{1}(1-\alpha )}{2} \\ y_{1} & = w_{1}>y_{2}=\frac{w_{1}(1-\alpha )}{2\beta }>y_{3}=\frac{ w_{1}(1-\alpha )}{2}. \end{aligned}$$
  3. 3.

    If \(\alpha >\frac{1}{3}\), and \(\beta \le \frac{1}{2}\)

    $$\begin{aligned} x_{1} & = \frac{v_{1}}{3\alpha }>x_{2}=\frac{2}{3}v_{1}>x_{3}=\frac{2}{3} v_{1}(1-\beta ) \\ y_{1} & = \frac{w_{1}}{3\alpha }>y_{2}=\frac{2}{3}w_{1}>x_{3}=\frac{2}{3} w_{1}(1-\beta ). \end{aligned}$$
  4. 4.

    If \(\alpha >\frac{1}{3}\), and \(\beta >\frac{1}{2}\)

    $$\begin{aligned} x_{1} & = \frac{v_{1}}{3\alpha }>\frac{2}{3}v_{1}(1-\beta )\text { and }x_{2}= \frac{v_{1}}{3\beta }>x_{3}=\frac{v_{1}}{3} \\ y_{1} & = \frac{w_{1}}{3\alpha }>y_{3}=\frac{w_{1}}{3}\text { and }y_{2}=\frac{ w_{1}}{3\beta }>y_{3}=\frac{w_{1}}{3}. \end{aligned}$$

In all of the above four cases, the resource allocations in the third stage are smaller than in the previous ones. However, in cases 1–3, independent of the values of the fatigue parameters \(\alpha\) and \(\beta\), the resource allocations of the first stage are larger than that of the second one. However, in case 4, the resource allocation of the first stage is larger than that of the second one if the fatigue parameter of the first stage \(\alpha\) is smaller than the fatigue parameter of the second stage \(\beta\); namely, \(\alpha >\frac{1}{3}\), and \(\beta >\frac{1}{2}\) and \(\alpha <\beta .\) \(\ Q.E.D.\)

1.4 Proof of Proposition 5

By (12), player 1’s expected payoff in the best-of-two contest is

$$\begin{aligned} u_{best2}^{1}=\frac{v_{1}^{r}}{v_{1}^{r}+w_{1}^{r}}, \end{aligned}$$

and by (37), player 1’s expected payoff in the best-of-three contest is

$$\begin{aligned} u_{best3}^{1}=\frac{v_{1}^{2r}}{\left( v_{1}^{r}+w_{1}^{r}\right) ^{3}} \left( v_{1}^{r}+3w_{1}^{r}\right) . \end{aligned}$$

Thus, we have

$$\begin{aligned} u_{best2}^{1}-u_{best3}^{1}=\frac{v_{1}^{r}}{v_{1}^{r}+w_{1}^{r}} -\frac{v_{1}^{2r}}{\left( v_{1}^{r}+w_{1}^{r}\right) ^{3}}\left( v_{1}^{r}+3w_{1}^{r}\right) =-v_{1}^{r}\frac{w_{1}^{r}}{\left( v_{1}^{r}+w_{1}^{r}\right) ^{3}}\left( v_{1}^{r}-w_{1}^{r}\right) . \end{aligned}$$

We can see that \(u_{best3}^{1}\ge u_{best2}^{1}\) iff \(v_{1}\ge w_{1}\). The result is similar for player 2. Q.E.D.

1.5 Proof of Proposition 6

We assume that the fatigue parameter is the same for both stages of the best-of-three contest and that it is equal to the fatigue parameter of the best-of-two contest. We denote this parameter by \(\alpha\). According to the analysis of the resource allocations in the previous sections, the players’ total effort in the best-of-two contest is maximized when \(\delta \ge \alpha\) and is equal to

$$\begin{aligned} R_{best2}=(v_{1}+w_{1})(2-\alpha ). \end{aligned}$$

Then, we have the following three cases:

  1. 1.

    When \(\alpha \ge \frac{1}{2}\), the resource constraints in all the stages of the best-of-three contest are not binding. Then, the players’ total resource allocations in the best-of-three contest is

    $$\begin{aligned} R_{best3}=\frac{v_{1}+w_{1}}{3\alpha }+\frac{v_{1}+w_{1}}{3\alpha }+\frac{ v_{1}+w_{1}}{3}. \end{aligned}$$

    The difference between the total resource in both contests is

    $$\begin{aligned} R_{best3}-R_{best2}=\frac{1}{3\alpha }\left( v_{1}+w_{1}\right) \left( 3\alpha ^{2}-5\alpha +2\right) \end{aligned}$$

    Thus, \(R_{best3}-R_{best2}>0\) if \(0\le \alpha <\frac{2}{3}\); otherwise, if \(\alpha \ge \frac{2}{3}\) we obtain that \(R_{best3}-R_{best2}\le 0.\)

  2. 2.

    When \(\frac{1}{3}\le \alpha <\frac{1}{2}\), the resource constraint in the first stage of the best-of-three contest is binding but in the second stage is not. Then, the players’ total resource allocation is

    $$\begin{aligned} R_{best3}=\frac{v_{1}+w_{1}}{3\alpha }+\frac{2(v_{1}+w_{1})}{3}+\frac{ 2(v_{1}+w_{1})(1-\alpha )}{3}. \end{aligned}$$

    The difference between the total resource in both contests is

    $$\begin{aligned} R_{best3}-R_{best2}=\frac{1}{3\alpha }\left( \alpha -1\right) ^{2}\left( v_{1}+w_{1}\right) >0. \end{aligned}$$
  3. 3.

    When \(\alpha <\frac{1}{3}\), the resource constraint in the best-of-three contest is binding and the players’ total resource allocation is

    $$\begin{aligned} R_{best3}=(v_{1}+w_{1})+(v_{1}+w_{1})(1-\alpha )+(v_{1}+w_{1})(1-\alpha )^{2}. \end{aligned}$$

    Then

    $$\begin{aligned} R_{best3}-R_{best2}=\left( \alpha -1\right) ^{2}\left( v_{1}+w_{1}\right) >0. \end{aligned}$$

    Q.E.D.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Sela, A. Resource allocations in the best-of-k (\(k=2,3\)) contests. J Econ 139, 235–260 (2023). https://doi.org/10.1007/s00712-023-00827-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00712-023-00827-w

Keywords

JEL Classification

Navigation