Performance of Machine Learning Algorithms in Predicting Game Outcome from Drafts in Dota 2

Conference paper
Part of the Communications in Computer and Information Science book series (CCIS, volume 661)


In this paper we suggest the first systematic review and compare performance of most frequently used machine learning algorithms for prediction of the match winner from the teams’ drafts in Dota 2 computer game. Although previous research attempted this task with simple models, weve made several improvements in our approach aiming to take into account interactions among heroes in the draft. For that purpose we’ve tested the following machine learning algorithms: Naive Bayes classifier, Logistic Regression and Gradient Boosted Decision Trees. We also introduced Factorization Machines for that task and got our best results from them. Besides that, we found that model’s prediction accuracy depends on skill level of the players. We’ve prepared publicly available dataset which takes into account shortcomings of data used in previous research and can be used further for algorithms development, testing and benchmarking.


Online games Predictive models Dota 2 Factorization machines MOBA 



This paper was prepared within the framework of a subsidy granted to HSE by the Government of Russian Federation for implementation of the Global Competitiveness Program.


  1. 1.
    Taylor, T.: Raising the Stakes: E-sports and the Professionalization of Computer Gaming. MIT Press, Cambridge (2012)Google Scholar
  2. 2.
    Ontanón, S., Synnaeve, G., Uriarte, A., Richoux, F., Churchill, D., Preuss, M.: A survey of real-time strategy game AI research and competition in starcraft. IEEE Trans. Comput. Intell. AI Games 5(4), 293–311 (2013)CrossRefGoogle Scholar
  3. 3.
    Synnaeve, G., Bessiere, P.: A Bayesian model for opening prediction in RTS games with application to StarCraft. In: IEEE Conference on Computational Intelligence and Games (CIG), pp. 281–288. IEEE (2011)Google Scholar
  4. 4.
    Bosc, G., Tan, P., Boulicaut, J.-F., Raissi, C., Kaytoue, M.: A pattern mining approach to study strategy balance in RTS games. IEEE Trans. Comput. Intell. AI Games PP(99), 1–1 (2015)CrossRefGoogle Scholar
  5. 5.
    Cavadenti, O., Codocedo, V., Boulicaut, J.-F., Kaytoue, M.: When cyberathletes conceal their game: clustering confusion matrices to identify avatar aliases. In: 2015 IEEE International Conference on Data Science and Advanced Analytics (DSAA), pp. 1–10. IEEE, October 2015Google Scholar
  6. 6.
    Conley, K., Perry, D.: How does he saw me? A recommendation engine for picking heroes in Dota 2. Technical report (2013)Google Scholar
  7. 7.
    Nuangjumnonga, T., Mitomo, H.: Leadership development through online gaming. In: 19th ITS Biennial Conference: Moving Forward with Future Technologies: Opening a Platform for All, Bangkok, pp. 1–24 (2012)Google Scholar
  8. 8.
    Pobiedina, N., Neidhardt, J.: On successful team formation. Technical report (2013)Google Scholar
  9. 9.
    Rioult, F., Métivier, J.-P., Helleu, B., Scelles, N., Durand, C.: Mining tracks of competitive video games. AASRI Procedia 8, 82–87 (2014). SECSCrossRefGoogle Scholar
  10. 10.
    Drachen, A., Yancey, M., Maguire, J., Chu, D., Wang, I. Y., Mahlmann, T., Schubert, M., Klabajan, D.: Skill-based differences in spatio-temporal team behaviour in defence of the Ancients 2 (Dota 2). In: Games Media Entertainment (GEM), Dota 2, vol. 2, pp. 1–8. IEEE (2014)Google Scholar
  11. 11.
    Eggert, C., Herrlich, M., Smeddinck, J., Malaka, R.: classification of player roles in the team-based multi-player game Dota 2. In: Chorianopoulos, K., Divitini, M., Hauge, J.B., Jaccheri, L., Malaka, R. (eds.) ICEC 2015. LNCS, vol. 9353, pp. 112–125. Springer, Heidelberg (2015). doi: 10.1007/978-3-319-24589-8_9 CrossRefGoogle Scholar
  12. 12.
    Agarwala, A., Pearce, M.: Learning Dota 2 team compositions. Technical report, Stanford University (2014)Google Scholar
  13. 13.
    Song, K., Zhang, T., Ma, C.: Predicting the winning side of DotA2. Technical report, Stanford University (2015)Google Scholar
  14. 14.
    Kalyanaraman, K.: To win or not to win? A prediction model to determine the outcome of a DotA2 match. Technical report, University of California San Diego (2014)Google Scholar
  15. 15.
    Kinkade, N., Jolla, L., Lim, K.: Dota 2 win prediction. Technical report, University of California, San Diego (2015)Google Scholar
  16. 16.
    Johansson, F., Wikström, J., Johansson, F.: Result prediction by mining replays in Dota 2. Ph.D. thesis, Blekinge Institute of Technology (2015)Google Scholar
  17. 17.
    Rendle, S.: Factorization machines. In: Proceedings of the IEEE International Conference on Data Mining, ICDM 2010, Washington, DC, USA, pp. 995–1000. IEEE Computer Society (2010)Google Scholar
  18. 18.
    Freudenthaler, C., Schmidt-Thieme, L., Rendle, S.: Bayesian factorization machines. In: Workshop on Sparse Representation and Low-rank Approximation, Neural Information Processing Systems (NIPS-WS) (2011)Google Scholar
  19. 19.
    Bayer, I.: fastFM: a library for factorization machines. CoRR, abs/1505.00641 (2015)Google Scholar
  20. 20.
    Chen, T., Guestrin, C.: XGBoost: a Scalable Tree Boosting System. arXiv e-prints, March 2016Google Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  1. 1.International Laboratory for Applied Network ResearchNational Research University Higher School of EconomicsMoscowRussia
  2. 2.Yandex Data FactoryMoscowRussia
  3. 3.Moscow Institute of Physics and TechnologyMoscowRussia
  4. 4.National Research University Higher School of EconomicsMoscowRussia
  5. 5.Institute for Information Transmission ProblemsMoscowRussia

Personalised recommendations