Skip to main content
Log in

Round-robin tournaments generated by the Circle Method have maximum carry-over

  • Full Length Paper
  • Series B
  • Published:
Mathematical Programming Submit manuscript

Abstract

The Circle Method is widely used in the field of sport scheduling to generate schedules for round-robin tournaments. If in such a tournament, team A played team B in its previous match and is now playing team C, team C is said to receive a carry-over effect from team B. The so-called carry-over effect value is a number that can be associated to each round-robin schedule; it represents a degree of unbalancedness of the schedule with respect to carry-over. Here, we prove that, for an even number of teams, the Circle Method generates a schedule with maximum carry-over effect value, answering an open question.

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.

Fig. 1
Fig. 2
Fig. 3

Similar content being viewed by others

References

  1. Anderson, I.: Combinatorial Designs and Tournaments. Oxford University Press, Oxford (1997)

    MATH  Google Scholar 

  2. Anderson, I.: Balancing carry-over effects in tournaments. In: Holroyd, F., Quinn, K., Rowley, C., Webb, B. (eds.) Research Notes in Mathematics: vol. 403. Combinatorial Designs and their Applications, pp. 1–16. CRC Press, Boca Raton (1997)

  3. Froncek, D.: Scheduling a tournament. In: Gallian, J.A. (ed.) Dolciani mathematical expositions: Vol 43. Mathematics and Sports, pp. 203–216, Mathematical Association of America, Washington DC (2010)

  4. De Werra, D.: Geography, games and graphs. Discret. Appl. Math. 2, 327–337 (1980)

    Article  MathSciNet  Google Scholar 

  5. De Werra, D.: Scheduling in sports. In: Hansen, P. (ed.) Annals of Discrete Mathematics: Vol.11, Studies on Graphs and Discrete Programming, pp. 381–395. North-Holland, Amsterdam (1981)

  6. Goossens, D.R., Spieksma, F.C.R.: Soccer schedules in Europe: an overview. J. Sched. 15, 641–651 (2012)

    Article  Google Scholar 

  7. Goossens, D.R., Spieksma, F.C.R.: The carry-over effect does not influence football results. J. Sports Econ. 13, 288–305 (2012)

    Article  Google Scholar 

  8. Griggs, T., Rosa, A.: A tour of european soccer schedules, or testing the popularity of \(GK_{2n}\). Bull. Inst. Comb. Appl. 18, 65–68 (1996)

    MathSciNet  MATH  Google Scholar 

  9. Januario, T., Urrutia, S., Ribeiro, C., de Werra, D.: Edge coloring: A natural model for sports scheduling. Eur. J. Oper. Res. 254(1), 1–8 (2016)

    Article  MathSciNet  Google Scholar 

  10. Kendall, G., Knust, S., Ribeiro, C., Urrutia, S.: Scheduling in sports: an annotated bibliography. Comput. Op. Res. 37, 1–19 (2010)

    Article  MathSciNet  Google Scholar 

  11. Kidd, M.P.: A tabu-search for minimising the carry-over effects value of a round-robin tournament. ORiON 26(2), 125–141 (2010)

    Article  Google Scholar 

  12. Kirkman, T.P.: On a problem in combinatorics. Camb. Dublin Math. J. 2, 191–204 (1847)

    Google Scholar 

  13. Lambrechts, E., Ficker, A.M.C., Goossens, D.R., Spieksma, F.C.R.: Round-Robin tournaments generated by the circle method have maximum carry-over. In: Louveaux, Q., Skutella, M. (eds.) Integer Programming and Combinatorial Optimization: 18th International Conference, IPCO 2016, pp. 178–189. Springer, Cham (2016)

    Google Scholar 

  14. Miyashiro, R., Matsui, T.: Minimizing the carry-over effects value in a round-robin tournament. In: Proceedings of the 6th International Conference on the Practice and Theory of Automated Timetabling, edited by E.K. Burke and H. Rudová, pp. 460–463, (2006)

  15. Russell, K.G.: Balancing carry-over effects in round robin tournaments. Biometrika 67(1), 127–131 (1980)

    Article  MathSciNet  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Annette M. C. Ficker.

Additional information

This work is supported by the Interuniversity Attraction Poles Programme initiated by the Belgian Science Policy Office.

Appendix: Verification of the Bound

Appendix: Verification of the Bound

We prove the claim that

$$\begin{aligned} \frac{{\left( \ell - 3\right) } n^{2} - {\left( \ell ^{2} - 2 \, \ell - 13\right) } n - 12 \, \ell - 12}{{\left( \ell + 1\right) } n} > 0 \end{aligned}$$
(11)

when \(\ell \) lies between 3 and \(n-3\). We use the assumption that \(n \ge 10\) as well.

Since n and \(\ell \) are positive, it is sufficient to check that the numerator is positive:

$$\begin{aligned} (\ell - 3) n^{2} - {\left( \ell ^{2} - 2 \, \ell - 13\right) } n - 12 \, \ell - 12 > 0. \end{aligned}$$

The left hand side is a quadratic polynomial in both n and \(\ell \). Given a value for n we want to know for which values of \(\ell \) this expression is positive. So we rewrite the previous expression ordered by powers of \(\ell \):

$$\begin{aligned} -n \ell ^2+(n^2+2n-12)\ell +(-3n^2+13n -12). \end{aligned}$$

This quadratic polynomial has a strictly negative leading coefficient, so it is positive between its zero points which are given by

$$\begin{aligned} \ell _1 = \frac{n^{2} + 2 \, n - \sqrt{n^{4} - 8 \, n^{3} + 32 \, n^{2} - 96 \, n + 144} - 12}{2 \, n},\\ \ell _2 = \frac{n^{2} + 2 \, n + \sqrt{n^{4} - 8 \, n^{3} + 32 \, n^{2} - 96 \, n + 144} - 12}{2 \, n}. \end{aligned}$$

Now the claim can be formulated as follows: the interval over \(\ell \) where expression (11) is positive, contains the interval \([3, n-3]\), or equivalently, we have: \(\ell _1 < 3\) and \(\ell _2 > n-3\). To prove the first inequality for all values of n, it is sufficient to show that (i) \(\ell _1\) is strictly increasing, and (ii) has as limit 3 when n goes to infinity. To calculate this limit, we multiply numerator and denominator by \(n^2 + 2 n - 12 + \sqrt{n^4 - 8 n^3 + 32 n^2 - 96 n + 144}\), and get

$$\begin{aligned} \ell _1= & {} \frac{12 n^3 - 52 n^2+ 48 n}{2 n(n^2 + 2 n + \sqrt{n^4 - 8 n^3 + 32 n^2 - 96 n + 144} - 12)}, \\= & {} \frac{6 n^2 - 26 n+ 24 }{n^2 + 2 n + \sqrt{n^4 - 8 n^3 + 32 n^2 - 96 n + 144} - 12}.\\ \end{aligned}$$

Clearly, \(\lim \nolimits _{n\rightarrow \infty } \ell _1 = 3\).

To prove that \(\ell _1\) is strictly increasing, we consider the derivative of \(\ell _1\) with respect to n:

$$\begin{aligned} \ell '_1 = -\frac{n^{4} - 4 \, n^{3} - {\left( n^{2} + 12\right) }\sqrt{n^{4} - 8 \, n^{3} + 32 \, n^{2} - 96 \, n + 144} + 48 \, n - 144}{2 n^{2}\sqrt{n^{4} - 8 \, n^{3} + 32 \, n^{2} - 96 \, n + 144}}. \end{aligned}$$

Notice that \(n^{4} - 8 \, n^{3} + 32 \, n^{2} - 96 \, n + 144\) is always positive (indeed, this polynomial has four imaginary roots). So it is sufficient to show that the numerator is positive:

$$\begin{aligned} -n^{4} + 4 n^{3} + (n^{2} + 12)\sqrt{n^{4} - 8 \, n^{3} + 32 \, n^{2} - 96 n + 144} - 48 n + 144 > 0. \end{aligned}$$
(12)

We obtain the following series of equivalent inequalities.

$$\begin{aligned} (n^{2} + 12)\sqrt{n^4 - 8n^3 + 32n^2 - 96n + 144}> & {} n^{4} - 4 n^{3} + 48 n - 144 \Longleftrightarrow \\ (n^{2} + 12)\sqrt{n^4 - 8n^3 + 32n^2 - 96n + 144}> & {} (n^{2} + 12)( n^2 - 4n -12) \Longleftrightarrow \\ \sqrt{n^4 - 8n^3 + 32n^2 - 96n + 144}> & {} n^2 - 4n -12 \Longleftrightarrow \\ n^4 - 8n^3 + 32n^2 - 96n + 144> & {} (n^2 - 4n -12)^2 \Longleftrightarrow \\ 8n(5n-24)+(n^2 - 4n -12)^2> & {} (n^2 - 4n -12)^2 \Longleftrightarrow \\ 8n(5n-24)> & {} 0 \end{aligned}$$

This is true since we assume that \(n \ge 10\); it follows that \(\ell _1<3\) if \(n \ge 10\).

For the other inequality \(\ell _2>n-3\), it is sufficient to prove that (i) this inequality holds for \(n = 8\), and (ii) that the derivative of \(\ell _2\) with respect to n is greater than, or equal to, 1. This condition on the derivative implies that if n increases with 1 that \(\ell _2\) increases with at least 1.

One can deduce that for \(n=8\) the value of \(\ell _2 \approx 8.71058907144937 > n - 3\).

Next, the derivative of \(\ell _2\) equals

$$\begin{aligned} \frac{n^{4} - 4 \, n^{3} + {\left( n^{2} + 12\right) }\sqrt{n^{4} - 8 \, n^{3} + 32 \, n^{2} - 96 \, n + 144} + 48 \, n - 144}{2 n^{2} \sqrt{n^{4} - 8 \, n^{3} + 32 \, n^{2} - 96 \, n + 144}}, \end{aligned}$$

and this has to be greater than, or equal to, 1. Thus, the following inequality has to hold

$$\begin{aligned} {n^{4} - 4 \, n^{3} - {\left( n^{2} - 12\right) }\sqrt{n^{4} - 8 \, n^{3} + 32 \, n^{2} - 96 \, n + 144} + 48 \, n - 144} \ge 0. \end{aligned}$$

The calculation to show this is analogous to the previous one. Indeed, we have:

$$\begin{aligned} (n^{2} - 12)\sqrt{n^4 - 8n^3 + 32n^2 - 96n + 144}\le & {} n^{4} - 4n^3 + 48n - 144 \Longleftrightarrow \\ (n^{2} - 12)\sqrt{n^4 - 8n^3 + 32n^2 - 96n + 144}\le & {} (n^{2} - 12)(n^2 - 4n + 12) \Longleftrightarrow \\ \sqrt{n^4 - 8n^3 + 32n^2 - 96n + 144}\le & {} n^2 - 4n + 12 \Longleftrightarrow \\ n^4 - 8n^3 + 32n^2 - 96n + 144\le & {} (n^2 - 4n + 12)^2 \Longleftrightarrow \\ n^4 - 8n^3 + 32n^2 - 96n + 144\le & {} n^4 - 8n^3 + 40n^2 -96n + 144 \Longleftrightarrow \\ 0\le & {} 8n^2 \end{aligned}$$

So the claim that \(\ell _2>n-3\), holds for \(n \ge 10\).

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lambrechts, E., Ficker, A.M.C., Goossens, D.R. et al. Round-robin tournaments generated by the Circle Method have maximum carry-over. Math. Program. 172, 277–302 (2018). https://doi.org/10.1007/s10107-017-1115-x

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10107-017-1115-x

Keywords

Navigation