{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T08:51:42Z","timestamp":1791967902692,"version":"4.3.4"},"reference-count":55,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,7,13]],"date-time":"2026-07-13T00:00:00Z","timestamp":1783900800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T00:00:00Z","timestamp":1791936000000},"content-version":"vor","delay-in-days":93,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Sci Rep"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>High-frequency trading now plays out at microsecond and even nanosecond granularity, and profitability is shaped jointly by adversarial market dynamics and by acute sensitivity to execution latency. Most existing frameworks, somewhat surprisingly, still treat these two pressures in isolation, leaving robustness and timing largely uncoupled. This study sets out to unify adversarial robustness training with latency-aware policy optimization inside a single strategy-design framework, hereafter referred to as MAGAT \u2014 a multi-agent game-theoretic adversarial trading system. Methodologically, MAGAT pits a Protagonist Agent \u2014 trained by multi-agent proximal policy optimization under centralized training with decentralized execution \u2014 against an Adversary Agent that searches for worst-case perturbations using gradient-free evolution strategies; the two roles alternate in a minimax loop whose fixed point is interpreted as an approximate, not certified, equilibrium, and the approximation is itself diagnosed by tracking the joint best-response gap. A Latency-Aware Reward Shaping (LARS) term penalizes aggressive orders in proportion to the logarithm of realized delay, while an FPGA, INT8-GPU, and kernel-bypass pipeline targets sub-700-nanosecond tick-to-order latency. Performance is assessed entirely through event-driven simulation: matching-engine replay over Level-3 LOBSTER limit order book data for three U.S. equities (AAPL, MSFT, INTC, 2022), with a configurable delay-injection platform reproducing uniform, Pareto, and bursty latency profiles. Across the four stress scenarios MAGAT sustains Sharpe ratios of 1.97\u20132.31 (95% bootstrap CI\u2009\u00b1\u20090.06) and survival rates of 89\u201396% (\u00b1\u20092.1pp), against 1.18 and 69% for the strongest single-agent baseline; latency elasticity falls roughly fourfold (from 0.54 to 0.12) and the measured 99th-percentile execution latency stays near 683 nanoseconds. These results refer strictly to simulation and should not, on their own, be read as evidence of deployable live performance.<\/jats:p>","DOI":"10.1038\/s41598-026-60518-6","type":"journal-article","created":{"date-parts":[[2026,7,13]],"date-time":"2026-07-13T12:15:17Z","timestamp":1783944917000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["A multi-agent game-theoretic adversarial training framework with latency-aware reward shaping for robust high-frequency trading strategy design"],"prefix":"10.1038","volume":"16","author":[{"given":"Yuan","family":"Li","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,7,13]]},"reference":[{"key":"60518_CR1","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1146\/annurev-financial-121415-033010","volume":"8","author":"A Menkveld","year":"2016","unstructured":"Menkveld, A. The economics of high-frequency trading: Taking stock. Annual Rev. Financial Econ. 8, 1\u201324 (2016).","journal-title":"Annual Rev. Financial Econ."},{"issue":"1","key":"60518_CR2","doi-asserted-by":"publisher","first-page":"493","DOI":"10.1093\/qje\/qjab032","volume":"137","author":"M Aquilina","year":"2022","unstructured":"Aquilina, M., Budish, E. & O\u2019Neill, P. Quantifying the high-frequency trading \u2018arms race\u2019. Quart. J. Econ. 137 (1), 493\u2013564 (2022).","journal-title":"Quart. J. Econ."},{"key":"60518_CR3","unstructured":"Cartea, \u00c1., Jaimungal, S. & Penalva, J. Algorithmic and High-Frequency Trading (Cambridge University Press, 2015)."},{"issue":"3","key":"60518_CR4","doi-asserted-by":"publisher","first-page":"325","DOI":"10.1080\/14697688.2012.757636","volume":"13","author":"JD Farmer","year":"2013","unstructured":"Farmer, J. D. & Skouras, S. An ecological perspective on the future of computer trading. Quant. Finance. 13 (3), 325\u2013346 (2013).","journal-title":"Quant. Finance"},{"issue":"4","key":"60518_CR5","doi-asserted-by":"publisher","first-page":"646","DOI":"10.1016\/j.finmar.2013.05.003","volume":"16","author":"J Hasbrouck","year":"2013","unstructured":"Hasbrouck, J. & Saar, G. Low-latency trading. J. Financial Markets. 16 (4), 646\u2013679 (2013).","journal-title":"J. Financial Markets"},{"issue":"4","key":"60518_CR6","first-page":"39","volume":"20","author":"DH Bailey","year":"2017","unstructured":"Bailey, D. H., Borwein, J. M., L\u00f3pez de, M., Prado & Zhu, Q. J. The probability of backtest overfitting. J. Comput. Finance. 20 (4), 39\u201369 (2017).","journal-title":"J. Comput. Finance"},{"key":"60518_CR7","doi-asserted-by":"crossref","unstructured":"Spooner, T., Fearnley, J., Savani, R. & Koukorinis, A. Market making via reinforcement learning, in Proc. 17th International Conference on Autonomous Agents and MultiAgent Systems (AAMAS), pp. 434\u2013442, (2018).","DOI":"10.65109\/YYQW8667"},{"key":"60518_CR8","unstructured":"Ganesh, S. et al. Reinforcement learning for market making in a multi-agent dealer market, arXiv preprint arXiv:1911.05892, (2019)."},{"key":"60518_CR9","doi-asserted-by":"crossref","unstructured":"Ba\u015far, T. & Olsder, G. J. Dynamic Noncooperative Game Theory 2nd edn (SIAM, 1999).","DOI":"10.1137\/1.9781611971132"},{"key":"60518_CR10","unstructured":"Vadori, N. et al. Towards multi-agent reinforcement learning driven over-the-counter market simulations, arXiv preprint arXiv:2210.07184, 42\u201356 (2022)."},{"issue":"267","key":"60518_CR11","first-page":"1","volume":"23","author":"T Spooner","year":"2022","unstructured":"Spooner, T. & Savani, R. Robust market making via adversarial reinforcement learning. J. Mach. Learn. Res. 23 (267), 1\u201354 (2022).","journal-title":"J. Mach. Learn. Res."},{"issue":"6","key":"60518_CR12","doi-asserted-by":"publisher","first-page":"1051","DOI":"10.1080\/14697688.2022.2039403","volume":"22","author":"C Moallemi","year":"2022","unstructured":"Moallemi, C. & Wang, M. A reinforcement learning approach to optimal execution. Quant. Finance. 22 (6), 1051\u20131069 (2022).","journal-title":"Quant. Finance"},{"key":"60518_CR13","doi-asserted-by":"crossref","unstructured":"Frey, S. Y. et al. JAX-LOB: A GPU-accelerated limit order book simulator to unlock large scale reinforcement learning for trading, in Proc. 4th ACM International Conference on AI in Finance, pp. 583\u2013591, (2023).","DOI":"10.1145\/3604237.3626880"},{"key":"60518_CR14","doi-asserted-by":"crossref","unstructured":"Kalogiannis, F., Anagnostides, I. & Sandholm, T. Learning equilibria in adversarial team Markov games: A nonconvex-hidden-concave min-max optimization problem, in Advances in Neural Information Processing Systems (NeurIPS), (2024).","DOI":"10.52202\/079017-2948"},{"key":"60518_CR15","doi-asserted-by":"crossref","unstructured":"Belcak, P., Calliess, J. P. & Zohren, S. Fast agent-based simulation framework with applications to reinforcement learning and the study of trading latency effects, in Multi-Agent-Based Simulation XXII, Springer, 42\u201356, (2022).","DOI":"10.1007\/978-3-030-94548-0_4"},{"key":"60518_CR16","doi-asserted-by":"crossref","unstructured":"Shapley, L. Stochastic games, Proceedings of the National Academy of Sciences, vol. 39, no. 10, pp. 1095\u20131100, (1953).","DOI":"10.1073\/pnas.39.10.1953"},{"issue":"11","key":"60518_CR17","doi-asserted-by":"publisher","first-page":"1709","DOI":"10.1080\/14697688.2013.803148","volume":"13","author":"MD Gould","year":"2013","unstructured":"Gould, M. D. et al. Limit order books. Quant. Finance. 13 (11), 1709\u20131748 (2013).","journal-title":"Quant. Finance"},{"key":"60518_CR18","unstructured":"Amato, C. An introduction to centralized training for decentralized execution in cooperative multi-agent reinforcement learning, arXiv preprint arXiv:2409.03052, (2024)."},{"key":"60518_CR19","unstructured":"Rashid, T. et al. QMIX: Monotonic value function factorisation for deep multi-agent reinforcement learning, in Proc. 35th International Conference on Machine Learning (ICML), pp. 4295\u20134304, (2018)."},{"key":"60518_CR20","doi-asserted-by":"crossref","unstructured":"Yu, C. et al. The surprising effectiveness of PPO in cooperative multi-agent games, in Advances in Neural Information Processing Systems (NeurIPS) Datasets and Benchmarks Track, (2022).","DOI":"10.52202\/068431-1787"},{"key":"60518_CR21","doi-asserted-by":"crossref","unstructured":"Foerster, J., Farquhar, G., Afouras, T., Nardelli, N. & Whiteson, S. Counterfactual multi-agent policy gradients, in Proc. 32nd AAAI Conference on Artificial Intelligence, pp. 2974\u20132982, (2018).","DOI":"10.1609\/aaai.v32i1.11794"},{"key":"60518_CR22","unstructured":"Osborne, M. J. & Rubinstein, A. A Course in Game Theory (MIT Press, 1994)."},{"key":"60518_CR23","doi-asserted-by":"crossref","unstructured":"Nash, J. F. Equilibrium points in n-person games, Proceedings of the National Academy of Sciences, vol. 36, no. 1, pp. 48\u201349, (1950).","DOI":"10.1073\/pnas.36.1.48"},{"issue":"1","key":"60518_CR24","doi-asserted-by":"publisher","first-page":"195","DOI":"10.1137\/070699652","volume":"39","author":"C Daskalakis","year":"2009","unstructured":"Daskalakis, C., Goldberg, P. W. & Papadimitriou, C. H. The complexity of computing a Nash equilibrium. SIAM J. Comput. 39 (1), 195\u2013259 (2009).","journal-title":"SIAM J. Comput."},{"key":"60518_CR25","unstructured":"Chen, Y., Tao, Y. & Yang, Y. On the convergence of fictitious play: A decomposition approach, in Proc. 31st International Joint Conference on Artificial Intelligence (IJCAI), pp. 182\u2013188, (2022)."},{"key":"60518_CR26","doi-asserted-by":"crossref","unstructured":"Adam, L., Hor\u00e1k, K., Kroupa, T. & Bo\u0161ansk\u00fd, B. Double oracle algorithm for computing equilibria in continuous games, in Proc. 35th AAAI Conference on Artificial Intelligence, pp. 5070\u20135077, (2021).","DOI":"10.1609\/aaai.v35i6.16641"},{"issue":"1","key":"60518_CR27","doi-asserted-by":"publisher","first-page":"219","DOI":"10.1162\/089976600300015961","volume":"12","author":"K Doya","year":"2000","unstructured":"Doya, K. Reinforcement learning in continuous time and space. Neural Comput. 12 (1), 219\u2013245 (2000).","journal-title":"Neural Comput."},{"key":"60518_CR28","unstructured":"Singh, S. P., Jaakkola, T. & Jordan, M. I. Reinforcement learning with soft state aggregation, in Advances in Neural Information Processing Systems, vol. 7, pp. 361\u2013368, (1995)."},{"key":"60518_CR29","unstructured":"Kleinrock, L. Queueing Systems, Volume I: Theory (Wiley-Interscience, 1975)."},{"issue":"6","key":"60518_CR30","doi-asserted-by":"publisher","first-page":"835","DOI":"10.1109\/90.650143","volume":"5","author":"ME Crovella","year":"1997","unstructured":"Crovella, M. E. & Bestavros, A. Self-similarity in World Wide Web traffic: Evidence and possible causes. IEEE\/ACM Trans. Networking. 5 (6), 835\u2013846 (1997).","journal-title":"IEEE\/ACM Trans. Networking"},{"key":"60518_CR31","doi-asserted-by":"crossref","unstructured":"Barndorff-Nielsen, O. E. & Shephard, N. Realised power variation and stochastic volatility models, Bernoulli, vol. 9, no. 2, pp. 243\u2013265, (2003).","DOI":"10.3150\/bj\/1068128977"},{"key":"60518_CR32","doi-asserted-by":"crossref","unstructured":"Doostmohammadian, M., Ramesh, N. K. & Aghasi, A. Delay-tolerant augmented-consensus-based distributed directed optimization, Systems & Control Letters, vol. 205, art. no. 106260, (2025).","DOI":"10.1016\/j.sysconle.2025.106260"},{"key":"60518_CR33","doi-asserted-by":"crossref","unstructured":"Makridis, E. et al. Distributed optimization with gradient tracking over heterogeneous delay-prone directed networks, in Proc. 2024 European Control Conference (ECC), pp. 2312\u20132319, (2024).","DOI":"10.23919\/ECC64448.2024.10590963"},{"key":"60518_CR34","doi-asserted-by":"crossref","unstructured":"Huang, R., Polak, T. & LOBSTER. : Limit order book reconstruction system, Available at SSRN: (2011). https:\/\/ssrn.com\/abstract=1977207.","DOI":"10.2139\/ssrn.1977207"},{"issue":"1","key":"60518_CR35","doi-asserted-by":"publisher","first-page":"47","DOI":"10.1093\/jjfinec\/nbt003","volume":"12","author":"R Cont","year":"2014","unstructured":"Cont, R., Kukanov, A. & Stoikov, S. The price impact of order book events. J. Financial Econometrics. 12 (1), 47\u201388 (2014).","journal-title":"J. Financial Econometrics"},{"issue":"8","key":"60518_CR36","doi-asserted-by":"publisher","first-page":"1315","DOI":"10.1080\/14697688.2015.1032546","volume":"15","author":"AN Kercheval","year":"2015","unstructured":"Kercheval, A. N. & Zhang, Y. Modelling high-frequency limit order book dynamics with support vector machines. Quant. Finance. 15 (8), 1315\u20131329 (2015).","journal-title":"Quant. Finance"},{"key":"60518_CR37","unstructured":"L\u00f3pez de, M. & Prado Advances in Financial Machine Learning (Wiley, 2018)."},{"key":"60518_CR38","unstructured":"Goodfellow, I. et al. Generative adversarial nets, in Advances in Neural Information Processing Systems, vol. 27, pp. 2672\u20132680, (2014)."},{"key":"60518_CR39","unstructured":"Schaul, T., Quan, J., Antonoglou, I. & Silver, D. Prioritized experience replay, in Proc. 4th International Conference on Learning Representations (ICLR), (2016)."},{"key":"60518_CR40","unstructured":"Salimans, T., Ho, J., Chen, X., Srinivas, S. & Abbeel, P. Evolution strategies as a scalable alternative to reinforcement learning, arXiv preprint arXiv:1703.03864, (2017)."},{"key":"60518_CR41","doi-asserted-by":"crossref","unstructured":"Byrd, D., Hybinette, M. & Balch, T. H. ABIDES: Towards high-fidelity multi-agent market simulation, in Proc. ACM SIGSIM Conference on Principles of Advanced Discrete Simulation (SIGSIM-PADS), pp. 11\u201322, (2020).","DOI":"10.1145\/3384441.3395986"},{"key":"60518_CR42","unstructured":"Ng, A. Y., Harada, D. & Russell, S. Policy invariance under reward transformations: Theory and application to reward shaping, in Proc. 16th International Conference on Machine Learning (ICML), pp. 278\u2013287, (1999)."},{"key":"60518_CR43","unstructured":"Schulman, J., Wolski, F., Dhariwal, P., Radford, A. & Klimov, O. Proximal policy optimization algorithms, arXiv preprint arXiv:1707.06347, (2017)."},{"key":"60518_CR44","unstructured":"Corporation, I. Data Plane Development Kit (DPDK): Programmer\u2019s guide, 2023. [Online]. Available: https:\/\/doc.dpdk.org\/guides\/"},{"key":"60518_CR45","doi-asserted-by":"publisher","unstructured":"Leber, C., Geib, B. & Litz, H. High frequency trading acceleration using FPGAs, in Proc. 2011 21st Int. Conf. on Field Programmable Logic and Applications (FPL), pp. 317\u2013322, (2011). https:\/\/doi.org\/10.1109\/FPL.2011.64","DOI":"10.1109\/FPL.2011.64"},{"issue":"2","key":"60518_CR46","doi-asserted-by":"publisher","first-page":"118","DOI":"10.3905\/jpm.2011.37.2.118","volume":"37","author":"D Easley","year":"2011","unstructured":"Easley, D., de L\u00f3pez, M. & O\u2019Hara, M. The microstructure of the \u2018Flash Crash\u2019: Flow toxicity, liquidity crashes, and the probability of informed trading. J. Portf. Manag. 37 (2), 118\u2013128 (2011).","journal-title":"J. Portf. Manag."},{"key":"60518_CR47","unstructured":"Haarnoja, T., Zhou, A., Abbeel, P. & Levine, S. Soft actor-critic: Off-policy maximum entropy deep reinforcement learning with a stochastic actor, in Proc. 35th International Conference on Machine Learning (ICML), pp. 1861\u20131870, (2018)."},{"issue":"3","key":"60518_CR48","doi-asserted-by":"publisher","first-page":"217","DOI":"10.1080\/14697680701381228","volume":"8","author":"M Avellaneda","year":"2008","unstructured":"Avellaneda, M. & Stoikov, S. High-frequency trading in a limit order book. Quant. Finance. 8 (3), 217\u2013224 (2008).","journal-title":"Quant. Finance"},{"issue":"1","key":"60518_CR49","doi-asserted-by":"publisher","first-page":"49","DOI":"10.3905\/jpm.1994.409501","volume":"21","author":"WF Sharpe","year":"1994","unstructured":"Sharpe, W. F. The Sharpe ratio. J. Portf. Manag. 21 (1), 49\u201358 (1994).","journal-title":"J. Portf. Manag."},{"issue":"3","key":"60518_CR50","doi-asserted-by":"publisher","first-page":"549","DOI":"10.1287\/opre.1090.0780","volume":"58","author":"R Cont","year":"2010","unstructured":"Cont, R., Stoikov, S. & Talreja, R. A stochastic model for order book dynamics. Oper. Res. 58 (3), 549\u2013563 (2010).","journal-title":"Oper. Res."},{"key":"60518_CR51","doi-asserted-by":"crossref","unstructured":"Bouchaud, J. P., Bonart, J., Donier, J. & Gould, M. Trades, Quotes and Prices: Financial Markets Under the Microscope (Cambridge University Press, 2018).","DOI":"10.1017\/9781316659335"},{"key":"60518_CR52","doi-asserted-by":"crossref","unstructured":"Wang, Z., Ventre, C. & Polukarov, M. Robust market making: To quote, or not to quote, in Proc. 4th ACM International Conference on AI in Finance, pp. 664\u2013672, (2023).","DOI":"10.1145\/3604237.3626858"},{"key":"60518_CR53","unstructured":"Lowe, R. et al. Multi-agent actor-critic for mixed cooperative-competitive environments, in Advances in Neural Information Processing Systems, vol. 30, pp. 6379\u20136390, (2017)."},{"issue":"1","key":"60518_CR54","doi-asserted-by":"publisher","first-page":"53","DOI":"10.1109\/MSP.2017.2765202","volume":"35","author":"A Creswell","year":"2018","unstructured":"Creswell, A. et al. Generative adversarial networks: An overview. IEEE. Signal. Process. Mag. 35 (1), 53\u201365 (2018).","journal-title":"IEEE. Signal. Process. Mag."},{"key":"60518_CR55","doi-asserted-by":"publisher","unstructured":"Vyetrenko, S. et al. Get real: realism metrics for robust limit order book market simulations, in Proc. 1st ACM Int. Conf. on AI in Finance (ICAIF \u201920), art. no. 2, pp. 1\u20138, (2020). https:\/\/doi.org\/10.1145\/3383455.3422561","DOI":"10.1145\/3383455.3422561"}],"container-title":["Scientific Reports"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-60518-6","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-60518-6.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-60518-6.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,14]],"date-time":"2026-10-14T08:23:55Z","timestamp":1791966235000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-60518-6"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,13]]},"references-count":55,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["60518"],"URL":"https:\/\/doi.org\/10.1038\/s41598-026-60518-6","relation":{},"ISSN":["2045-2322"],"issn-type":[{"value":"2045-2322","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,13]]},"assertion":[{"value":"9 March 2026","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"29 June 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"13 July 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"This study was approved by the Research Ethics Committee of Beijing Technology and Business University (Reference Number: IRB-2024-03176). All procedures were conducted in accordance with the Declaration of Helsinki and relevant national regulations. As this research involves computational modeling and analysis of publicly available historical market data without human participants, individual consent to participate was not applicable.","order":1,"name":"Ethics","label":"Ethics approval and consent to participate","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"All authors have reviewed the manuscript and consent to its publication. No identifiable information regarding participants has been included.","order":2,"name":"Ethics","label":"Consent for publication","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":3,"name":"Ethics","label":"Competing interests","group":{"name":"EthicsHeading","label":"Declarations"}}],"article-number":"32124"}}