Can Semantic Methods Enhance Team Sports Tactics? A Methodology for Football with Broader Applications

Alessio Di Rubbo; Marco Pedroni; Mattia Neri; Paolino Zica; Remo Pareschi; Roberto Valtancoli

arxiv: 2601.00421 · v2 · submitted 2026-01-01 · 💻 cs.AI

Can Semantic Methods Enhance Team Sports Tactics? A Methodology for Football with Broader Applications

Alessio Di Rubbo , Mattia Neri , Remo Pareschi , Marco Pedroni , Roberto Valtancoli , Paolino Zica This is my paper

Pith reviewed 2026-05-16 17:40 UTC · model grok-4.3

classification 💻 cs.AI

keywords semantic vector spacesfootball tacticsteam profile aggregationvector distance metricstactical fitplayer attribute vectorscompositional semanticsstrategy recommendation

0 comments

The pith

Semantic vector methods can evaluate how well football tactics like high press or counterattack fit a team's player profile.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper applies semantic-space techniques from linguistics to team sports by representing each player as a vector of technical, physical, and psychological attributes. Team profiles are formed by aggregating these vectors with contextual weights, placing both teams and tactical templates in the same space. Distances between a team profile and a tactic vector then quantify tactical fit and potential to exploit opponents. This produces interpretable recommendations that adapt dynamically. The approach treats collective play as carrying meaning in the same way texts do, opening a path to data-driven strategy beyond traditional analysis.

Core claim

Tactical configurations are modeled as compositional semantic structures in a shared vector space, where player attribute vectors aggregate into higher-level team profiles and tactical templates such as high press, counterattack, or possession build-up are encoded analogously to linguistic concepts, so that vector-distance metrics directly compute tactical fit and opponent-exploitation potential.

What carries the argument

A shared multidimensional vector space in which player attributes form vectors, team profiles arise from contextual aggregation, and tactical templates sit as points whose Euclidean or cosine distances to team profiles measure alignment.

If this is right

Direct computation of a numerical tactical-fit score for any given squad and tactic pair.
Quantification of how effectively one tactic can exploit weaknesses in a specific opponent profile.
Generation of fine-grained diagnostic reports that highlight which player attributes drive or hinder the recommended tactic.
Dynamic updating of recommendations as player attributes or opponent data change during a season.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Real-time match data streams could feed the same vector space to simulate outcomes before a coach commits to a substitution or formation change.
The same aggregation-and-distance logic could be applied to non-sports teams such as emergency-response units or corporate project groups to optimize role assignments.
Hybrid systems might let human coaches override or refine the vector-derived suggestions while retaining the underlying attribute-level explanations.

Load-bearing premise

That player attribute vectors can be aggregated into team profiles whose distances to tactic vectors produce recommendations that are both actionable and superior to conventional tactical analysis.

What would settle it

A side-by-side test in which teams following vector-distance recommendations show no higher win rate or performance metric than teams using standard scouting and coaching methods over a controlled set of matches.

Figures

Figures reproduced from arXiv: 2601.00421 by Alessio Di Rubbo, Marco Pedroni, Mattia Neri, Paolino Zica, Remo Pareschi, Roberto Valtancoli.

**Figure 1.** Figure 1: Context tree structure for two representative macro-attributes. Leaf nodes contain raw observables from match data; intermediate nodes aggregate by functional role; root nodes are the macroattributes used in semantic distance computation. Edges represent weighted aggregation functions. Data Sources. The context tree is designed to integrate multiple data streams: • Event data (e.g., Opta, StatsBomb): pass… view at source ↗

**Figure 2.** Figure 2: System architecture of the tactical decision support prototype. Context signals are aggregated into 14 macro-attributes (team vector), matched to strategy templates via adapted semantic distance, and produce interpretable recommendations and diagnostics. 4 Prototype Implementation The prototype of the tactical Decision Support System (DSS) was implemented in Python 3.10 using standard scientific libraries … view at source ↗

**Figure 3.** Figure 3: Example of radar plot for the “Energetic and Balanced” scenario. The shaded blue area represents the team profile, while the orange outline indicates the ideal strategy vector. Overall, the DSS exhibited behavior consistent with expert tactical intuition while maintaining quantitative transparency through vector distances. 5.3 Stability and Explainability Analyses To evaluate stability and interpretability… view at source ↗

**Figure 4.** Figure 4: Sensitivity of adapted distance dadapt with respect to contextual weight λ across the four scenarios. Smooth trends indicate stability in the optimal strategy selection. 5.4 Attribute Contribution Analysis Aggregating results across all scenarios, [PITH_FULL_IMAGE:figures/full_fig_p022_4.png] view at source ↗

**Figure 5.** Figure 5: Relative importance of the five most influential macro-attributes across all simulations. 5.5 Critical Discussion The experiments demonstrate that a vector-based semantic model can reproduce coherent tactical reasoning without hard-coded rules. The DSS adapts dynamically to variations in physical, 22 [PITH_FULL_IMAGE:figures/full_fig_p022_5.png] view at source ↗

**Figure 6.** Figure 6: Radar plot comparing the projected halftime team profile (solid blue) with the top three recommended strategies. Build-up Play shows the closest overall alignment, while the team’s high transition speed represents surplus capability relative to this strategy’s demands. 6.5 Retrospective Analysis 6.5.1 Observed vs. Recommended Tactics The DSS recommended Build-up Play—a possession-oriented strategy emphasiz… view at source ↗

read the original abstract

This paper explores how semantic-space reasoning, traditionally used in computational linguistics, can be extended to tactical decision-making in team sports. Building on the analogy between texts and teams -- where players act as words and collective play conveys meaning -- the proposed methodology models tactical configurations as compositional semantic structures. Each player is represented as a multidimensional vector integrating technical, physical, and psychological attributes; team profiles are aggregated through contextual weighting into a higher-level semantic representation. Within this shared vector space, tactical templates such as high press, counterattack, or possession build-up are encoded analogously to linguistic concepts. Their alignment with team profiles is evaluated using vector-distance metrics, enabling the computation of tactical ``fit'' and opponent-exploitation potential. A Python-based prototype demonstrates how these methods can generate interpretable, dynamically adaptive strategy recommendations, accompanied by fine-grained diagnostic insights at the attribute level. Beyond football, the approach offers a generalizable framework for collective decision-making and performance optimization in team-based domains -- ranging from basketball and hockey to cooperative robotics and human-AI coordination systems. The paper concludes by outlining future directions toward real-world data integration, predictive simulation, and hybrid human-machine tactical intelligence.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper sketches a semantic vector model for football tactics but offers no empirical validation or data to back its claims.

read the letter

This paper tries to apply semantic vector methods from linguistics to team sports tactics, specifically football. The main pitch is that you can build vectors for players from their attributes, combine them into team profiles with some weighting, encode standard tactics like high press as vectors too, and then use distance to see how well a team fits a tactic or exploits an opponent. What's new here is the direct mapping of those linguistic techniques onto tactical templates and the idea of computing fit and exploitation potential this way. The Python prototype is mentioned as a way to get interpretable outputs at the attribute level, which could be useful for coaches or analysts. It does a reasonable job explaining the steps without too much jargon and extends the idea to other team domains like robotics. The big problem is the complete absence of any data or testing. The description talks about benefits and adaptive recommendations, but there's no match data, no comparison to existing methods, and no numbers showing that lower distances actually predict better performance. The parameters for vector size and weighting are free choices with no grounding. That makes the central claim hard to evaluate. This is for people working on AI for sports or collective decision making who want to explore new modeling ideas. A reader could get some inspiration for their own work, but it won't provide ready-to-use validated tools. I think it deserves peer review so the authors can be asked to add validation experiments. Without that, it's more of a position paper than a methods contribution.

Referee Report

2 major / 1 minor

Summary. The paper proposes extending semantic vector-space methods from computational linguistics to football tactics by representing players as multidimensional vectors of technical, physical, and psychological attributes, aggregating them via contextual weighting into team profiles, encoding tactical templates (high press, counterattack, possession build-up) as analogous vectors, and using distance metrics to compute tactical fit and opponent-exploitation potential. A Python prototype is described as producing interpretable, adaptive recommendations with attribute-level diagnostics, and the approach is positioned as generalizable to other team domains.

Significance. If the vector-based distances reliably predict performance gains, the work would supply a novel, interpretable framework for tactical optimization that emphasizes fine-grained diagnostics and cross-domain applicability. The linguistic analogy and emphasis on dynamic adaptation are conceptually promising strengths, but the absence of any empirical grounding leaves the significance prospective rather than demonstrated.

major comments (2)

[Methodology description] No equations, parameter definitions, or formal specifications are supplied for constructing player vectors, performing contextual aggregation into team profiles, encoding tactic templates, or selecting the distance metric used to compute fit. This is load-bearing for the central claim that the resulting distances yield actionable recommendations.
[Prototype section] The Python prototype is presented as demonstrating interpretable outputs, yet no dataset, baseline comparison, statistical test, or correlation with match outcomes (win rate, possession metrics, expert ratings) is reported. Without such validation the claim that smaller distances produce superior strategies cannot be assessed.

minor comments (1)

The mapping from linguistic composition to collective play is stated at a high level; a more precise correspondence (e.g., which aggregation operation mirrors which linguistic operator) would improve clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback, which identifies key areas where the manuscript can be strengthened. We address each major comment point by point below, committing to revisions that add formal rigor and clarify the prototype's scope without overstating its current validation.

read point-by-point responses

Referee: [Methodology description] No equations, parameter definitions, or formal specifications are supplied for constructing player vectors, performing contextual aggregation into team profiles, encoding tactic templates, or selecting the distance metric used to compute fit. This is load-bearing for the central claim that the resulting distances yield actionable recommendations.

Authors: We agree that formal specifications are essential for reproducibility and to substantiate the central claims. In the revised manuscript we will add explicit definitions: player vectors as v_p = [t_p, ph_p, psy_p] with each component a normalized [0,1] attribute score; contextual aggregation into team profile T = sum_{p in team} w_p * v_p where weights w_p are derived from positional context and formation; tactic templates encoded as fixed vectors T_template in the same space; and fit computed via cosine distance d(T, T_template) = 1 - (T · T_template) / (|T| |T_template|). These additions will directly support the interpretability and actionability arguments. revision: yes
Referee: [Prototype section] The Python prototype is presented as demonstrating interpretable outputs, yet no dataset, baseline comparison, statistical test, or correlation with match outcomes (win rate, possession metrics, expert ratings) is reported. Without such validation the claim that smaller distances produce superior strategies cannot be assessed.

Authors: The prototype is intended strictly as an illustrative implementation on synthetic data to show workflow and diagnostic output, not as empirical validation. We will revise the section to state this limitation explicitly, include concrete example runs with attribute-level diagnostics, and add a forward-looking subsection outlining planned validation against real match datasets (e.g., Opta event data) using baselines such as expert-rated tactics and outcome correlations. This keeps the current paper focused on the methodological contribution while addressing the validation gap. revision: partial

Circularity Check

1 steps flagged

Tactical 'fit' reduces to vector distance by construction in the proposed semantic space

specific steps

fitted input called prediction [Abstract]
"Their alignment with team profiles is evaluated using vector-distance metrics, enabling the computation of tactical ``fit'' and opponent-exploitation potential."

The tactical fit is defined as the result of applying the vector-distance metric to the constructed team profiles and tactic vectors. Since the paper presents this distance computation as the method for obtaining fit (without separate validation against real performance data), the claimed output is statistically forced by the input choices of vector aggregation, weighting, and metric selection.

full rationale

The paper proposes a methodology that constructs player attribute vectors, aggregates them into team profiles via contextual weighting, encodes tactics as analogous vectors, and computes 'fit' and exploitation potential directly via vector-distance metrics. This central output is equivalent to the chosen aggregation and distance functions by definition, with no independent derivation, external benchmark, or correlation to match outcomes shown. The linguistic analogy is assumed without validation that would ground the parameters outside the model. This constitutes fitted-input-called-prediction circularity at the core claim level, warranting a moderate score of 6 rather than higher (no self-citation chain or uniqueness theorem is invoked).

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The framework rests on the untested text-team analogy and the assumption that attribute vectors can be defined and aggregated meaningfully; no free parameters are numerically specified but dimensionality and weighting choices are implicit.

free parameters (2)

vector dimensionality
The number of dimensions used to embed player attributes is not stated and must be selected or fitted.
contextual weighting coefficients
Weights applied when aggregating player vectors into team profiles are not given and would require definition or tuning.

axioms (2)

ad hoc to paper Players act as words and collective play conveys meaning in a manner analogous to linguistic composition.
Explicitly stated as the foundational building block of the methodology.
domain assumption Technical, physical, and psychological attributes can be integrated into a single multidimensional vector per player.
Required for the vector representation step.

pith-pipeline@v0.9.0 · 5524 in / 1471 out tokens · 84758 ms · 2026-05-16T17:40:38.006551+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

30 extracted references · 30 canonical work pages

[1]

Visual analysis of pressure in football.Data Mining and Knowledge Discovery, 31(6):1793–1839, 2017

Gennady Andrienko, Natalia Andrienko, Guido Budziak, Jason Dykes, Georg Fuchs, Tatiana von Landesberger, and Hendrik Weber. Visual analysis of pressure in football.Data Mining and Knowledge Discovery, 31(6):1793–1839, 2017. doi: 10.1007/s10618-017-0513-2

work page doi:10.1007/s10618-017-0513-2 2017
[2]

Pascal Bauer and Gabriel Anzer. Data-driven detection of counterpressing in professional football: A supervised machine learning task based on synchronized positional and event data with expert-based feature extraction.Data Mining and Knowledge Discovery, 35(5): 2009–2049, 2021. doi: 10.1007/s10618-021-00763-7

work page doi:10.1007/s10618-021-00763-7 2009
[3]

BERT : Pre-training of Deep Bidirectional Transformers for Language Understanding

Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. BERT: Pre-training of deep bidirectional transformers for language understanding. InProceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 4171–4186, 2019. doi: 10.18653/v1/N19-1423

work page doi:10.18653/v1/n19-1423 2019
[4]

Is a compact organization important for defensive success in elite soccer? Analysis based on player tracking data.International Journal of Sports Science & Coaching, 2024

Leander Forcher, Leon Forcher, Stefan Altmann, Darko Jekauc, and Matthias Kempe. Is a compact organization important for defensive success in elite soccer? Analysis based on player tracking data.International Journal of Sports Science & Coaching, 2024. doi: 10.1177/17479541231172695. Online first. 31

work page doi:10.1177/17479541231172695 2024
[5]

Ghisellini, R

R. Ghisellini, R. Pareschi, M. Pedroni, and G. B. Raggi. Recommending actionable strategies: A semantic approach to integrating analytical frameworks with decision heuristics. Information, 16(3):192, 2025. doi: 10.3390/info16030192

work page doi:10.3390/info16030192 2025
[6]

From extraction to synthesis: Entangled heuristics for agent-augmented strategic reasoning, 2025

Renato Ghisellini, Remo Pareschi, Marco Pedroni, and Giovanni Battista Raggi. From extraction to synthesis: Entangled heuristics for agent-augmented strategic reasoning, 2025. URLhttps://arxiv.org/abs/2507.13768

work page arXiv 2025
[7]

CourtVision: New visual and spatial analytics for the NBA

Kirk Goldsberry. CourtVision: New visual and spatial analytics for the NBA. InMIT Sloan Sports Analytics Conference, Boston, MA,

work page
[8]

URL https://www.sloansportsconference.com/research-papers/ courtvision-new-visual-and-spatial-analytics-for-the-nba

work page
[9]

Tactical knowledge in team sports from a constructivist and cognitivist perspective.Quest, 47:490–505, 11 1995

Jean-Francis Grehaigne and Paul Godbout. Tactical knowledge in team sports from a constructivist and cognitivist perspective.Quest, 47:490–505, 11 1995. doi: 10.1080/ 00336297.1995.10484171

work page arXiv 1995
[10]

Gudmundsson and M

J. Gudmundsson and M. Horton. Spatio-temporal analysis of team sports – a survey.ACM Computing Surveys, 50(2), 2017

work page 2017
[11]

Flexibility, stability, and adaptability of team playing style as key determinants of within-season performance in football

Qixiang He, Ying Hwa Kee, and John Komar. Flexibility, stability, and adaptability of team playing style as key determinants of within-season performance in football. InProceedings of the 9th International Performance Analysis Workshop and Conference & 5th IACSS Conference (PACSS 2021), volume 1426 ofAdvances in Intelligent Systems and Computing, pages 69...

work page doi:10.1007/978-3-030-99333-7_10 2021
[12]

Kullback and R

S. Kullback and R. A. Leibler. On information and sufficiency.Annals of Mathematical Statistics, 1951

work page 1951
[13]

1991 , publisher =

Jianhua Lin. Divergence measures based on the shannon entropy.IEEE Transactions on Information Theory, 37(1):145–151, January 1991. doi: 10.1109/18.61115

work page doi:10.1109/18.61115 1991
[14]

Defending in 4-4-2 or 5-3-2 formation? Small differences in footballers’ collective tactical behaviours.Journal of Sports Sciences, 40(7):793–805, 2022

Brandon Low, Robert Rein, Steffen Schwab, and Daniel Memmert. Defending in 4-4-2 or 5-3-2 formation? Small differences in footballers’ collective tactical behaviours.Journal of Sports Sciences, 40(7):793–805, 2022. doi: 10.1080/02640414.2021.1993655

work page doi:10.1080/02640414.2021.1993655 2022
[15]

Performance analysis in football: A critical review and implications for future research.Journal of Sports Sciences, 31(6):639–676, 2013

Rob Mackenzie and Chris Cushion. Performance analysis in football: A critical review and implications for future research.Journal of Sports Sciences, 31(6):639–676, 2013. doi: 10.1080/02640414.2012.746720. PMID: 23249092

work page doi:10.1080/02640414.2012.746720 2013
[16]

Salmon, Adam D

Scott McLean, Paul M. Salmon, Adam D. Gorman, Gemma J. M. Read, and Colin Solomon. What’s in a game? A systems approach to enhancing performance analysis in football. PLOS ONE, 12(2):e0172565, 2017. doi: 10.1371/journal.pone.0172565

work page doi:10.1371/journal.pone.0172565 2017
[17]

Mikolov, K

T. Mikolov, K. Chen, G. Corrado, and J. Dean. Efficient estimation of word representations in vector space. InICLR, 2013

work page 2013
[18]

Dolev Mutzari, Tonmoay Deb, Cristian Molinaro, Andrea Pugliese, V. S. Subrahmanian, and Sarit Kraus. Defending a city from multi-drone attacks: A sequential stackelberg security games approach, 2025. URLhttps://arxiv.org/abs/2508.11380

work page arXiv 2025
[19]

A public data set of spatio-temporal match events in soccer competitions.Scientific Data, 6:236, 2019

Luca Pappalardo, Paolo Cintia, Alessio Rossi, Emanuele Massucco, Paolo Ferragina, Dino Pedreschi, and Fosca Giannotti. A public data set of spatio-temporal match events in soccer competitions.Scientific Data, 6:236, 2019. doi: 10.1038/s41597-019-0247-7. 32

work page doi:10.1038/s41597-019-0247-7 2019
[20]

Measuring the effectiveness of playing strategies at soccer.Journal of the Royal Statistical Society: Series D (The Statistician), 46(4):541–550,

Richard Pollard and Charles Reep. Measuring the effectiveness of playing strategies at soccer.Journal of the Royal Statistical Society: Series D (The Statistician), 46(4):541–550,

work page
[21]

doi: 10.1111/1467-9884.00108

work page doi:10.1111/1467-9884.00108
[22]

Sentence- BERT : Sentence Embeddings using S iamese BERT -Networks

Nils Reimers and Iryna Gurevych. Sentence-BERT: Sentence embeddings using siamese BERT-networks. InProceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 3982–3992, Hong Kong, China, 2019. Association for Computational Linguistic...

work page doi:10.18653/v1/d19-1410 2019
[23]

Rein and D

R. Rein and D. Memmert. Big data and tactical analysis in elite soccer: Future challenges and opportunities for sports science.Journal of Sports Sciences, 34(7):639–650, 2016

work page 2016
[24]

Dajo Sanders, Mathieu Heijboer, Matthijs K. C. Hesselink, Tony Myers, and Ibrahim Akubat. Analysing a cycling grand tour: Can we monitor fatigue with intensity or load ratios? Journal of Sports Sciences, 36(12):1385–1391, 2018. doi: 10.1080/02640414.2017.1388669

work page doi:10.1080/02640414.2017.1388669 2018
[25]

Teresa Anguera, Jorge Campaniço, Nuno Matos, and José C

Hugo Sarmento, Rui Marcelino, M. Teresa Anguera, Jorge Campaniço, Nuno Matos, and José C. Leitão. Match analysis in football: a systematic review.Journal of Sports Sciences, 32(20):1831–1843, 2014. doi: 10.1080/02640414.2014.898852

work page doi:10.1080/02640414.2014.898852 2014
[26]

StatsBomb Open Data.https://github.com/statsbomb/open-data, 2024

StatsBomb. StatsBomb Open Data.https://github.com/statsbomb/open-data, 2024. Accessed: 2024

work page 2024
[27]

Turban, R

E. Turban, R. Sharda, and D. Delen.Decision Support and Business Intelligence Systems. Pearson, 2011

work page 2011
[28]

P. D. Turney and P. Pantel. From frequency to meaning: Vector space models of semantics. Journal of Artificial Intelligence Research, 2010

work page 2010
[29]

Learning How to V ote with Principles: Axiomatic Insights Into the Collective Decisions of Neural Networks.J

Peter D. Turney and Patrick Pantel. From frequency to meaning: Vector space models of semantics.Journal of Artificial Intelligence Research, 37:141–188, 2010. doi: 10.1613/jair. 2934

work page doi:10.1613/jair 2010
[30]

Weinberg and Daniel Gould.Foundations of Sport and Exercise Psychology

Robert S. Weinberg and Daniel Gould.Foundations of Sport and Exercise Psychology. Human Kinetics, Champaign, IL, 8th edition, 2023. ISBN 978-1-7182-0759-2. 33

work page 2023

[1] [1]

Visual analysis of pressure in football.Data Mining and Knowledge Discovery, 31(6):1793–1839, 2017

Gennady Andrienko, Natalia Andrienko, Guido Budziak, Jason Dykes, Georg Fuchs, Tatiana von Landesberger, and Hendrik Weber. Visual analysis of pressure in football.Data Mining and Knowledge Discovery, 31(6):1793–1839, 2017. doi: 10.1007/s10618-017-0513-2

work page doi:10.1007/s10618-017-0513-2 2017

[2] [2]

Pascal Bauer and Gabriel Anzer. Data-driven detection of counterpressing in professional football: A supervised machine learning task based on synchronized positional and event data with expert-based feature extraction.Data Mining and Knowledge Discovery, 35(5): 2009–2049, 2021. doi: 10.1007/s10618-021-00763-7

work page doi:10.1007/s10618-021-00763-7 2009

[3] [3]

BERT : Pre-training of Deep Bidirectional Transformers for Language Understanding

Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. BERT: Pre-training of deep bidirectional transformers for language understanding. InProceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT), pages 4171–4186, 2019. doi: 10.18653/v1/N19-1423

work page doi:10.18653/v1/n19-1423 2019

[4] [4]

Is a compact organization important for defensive success in elite soccer? Analysis based on player tracking data.International Journal of Sports Science & Coaching, 2024

Leander Forcher, Leon Forcher, Stefan Altmann, Darko Jekauc, and Matthias Kempe. Is a compact organization important for defensive success in elite soccer? Analysis based on player tracking data.International Journal of Sports Science & Coaching, 2024. doi: 10.1177/17479541231172695. Online first. 31

work page doi:10.1177/17479541231172695 2024

[5] [5]

Ghisellini, R

R. Ghisellini, R. Pareschi, M. Pedroni, and G. B. Raggi. Recommending actionable strategies: A semantic approach to integrating analytical frameworks with decision heuristics. Information, 16(3):192, 2025. doi: 10.3390/info16030192

work page doi:10.3390/info16030192 2025

[6] [6]

From extraction to synthesis: Entangled heuristics for agent-augmented strategic reasoning, 2025

Renato Ghisellini, Remo Pareschi, Marco Pedroni, and Giovanni Battista Raggi. From extraction to synthesis: Entangled heuristics for agent-augmented strategic reasoning, 2025. URLhttps://arxiv.org/abs/2507.13768

work page arXiv 2025

[7] [7]

CourtVision: New visual and spatial analytics for the NBA

Kirk Goldsberry. CourtVision: New visual and spatial analytics for the NBA. InMIT Sloan Sports Analytics Conference, Boston, MA,

work page

[8] [8]

URL https://www.sloansportsconference.com/research-papers/ courtvision-new-visual-and-spatial-analytics-for-the-nba

work page

[9] [9]

Tactical knowledge in team sports from a constructivist and cognitivist perspective.Quest, 47:490–505, 11 1995

Jean-Francis Grehaigne and Paul Godbout. Tactical knowledge in team sports from a constructivist and cognitivist perspective.Quest, 47:490–505, 11 1995. doi: 10.1080/ 00336297.1995.10484171

work page arXiv 1995

[10] [10]

Gudmundsson and M

J. Gudmundsson and M. Horton. Spatio-temporal analysis of team sports – a survey.ACM Computing Surveys, 50(2), 2017

work page 2017

[11] [11]

Flexibility, stability, and adaptability of team playing style as key determinants of within-season performance in football

Qixiang He, Ying Hwa Kee, and John Komar. Flexibility, stability, and adaptability of team playing style as key determinants of within-season performance in football. InProceedings of the 9th International Performance Analysis Workshop and Conference & 5th IACSS Conference (PACSS 2021), volume 1426 ofAdvances in Intelligent Systems and Computing, pages 69...

work page doi:10.1007/978-3-030-99333-7_10 2021

[12] [12]

Kullback and R

S. Kullback and R. A. Leibler. On information and sufficiency.Annals of Mathematical Statistics, 1951

work page 1951

[13] [13]

1991 , publisher =

Jianhua Lin. Divergence measures based on the shannon entropy.IEEE Transactions on Information Theory, 37(1):145–151, January 1991. doi: 10.1109/18.61115

work page doi:10.1109/18.61115 1991

[14] [14]

Defending in 4-4-2 or 5-3-2 formation? Small differences in footballers’ collective tactical behaviours.Journal of Sports Sciences, 40(7):793–805, 2022

Brandon Low, Robert Rein, Steffen Schwab, and Daniel Memmert. Defending in 4-4-2 or 5-3-2 formation? Small differences in footballers’ collective tactical behaviours.Journal of Sports Sciences, 40(7):793–805, 2022. doi: 10.1080/02640414.2021.1993655

work page doi:10.1080/02640414.2021.1993655 2022

[15] [15]

Performance analysis in football: A critical review and implications for future research.Journal of Sports Sciences, 31(6):639–676, 2013

Rob Mackenzie and Chris Cushion. Performance analysis in football: A critical review and implications for future research.Journal of Sports Sciences, 31(6):639–676, 2013. doi: 10.1080/02640414.2012.746720. PMID: 23249092

work page doi:10.1080/02640414.2012.746720 2013

[16] [16]

Salmon, Adam D

Scott McLean, Paul M. Salmon, Adam D. Gorman, Gemma J. M. Read, and Colin Solomon. What’s in a game? A systems approach to enhancing performance analysis in football. PLOS ONE, 12(2):e0172565, 2017. doi: 10.1371/journal.pone.0172565

work page doi:10.1371/journal.pone.0172565 2017

[17] [17]

Mikolov, K

T. Mikolov, K. Chen, G. Corrado, and J. Dean. Efficient estimation of word representations in vector space. InICLR, 2013

work page 2013

[18] [18]

Dolev Mutzari, Tonmoay Deb, Cristian Molinaro, Andrea Pugliese, V. S. Subrahmanian, and Sarit Kraus. Defending a city from multi-drone attacks: A sequential stackelberg security games approach, 2025. URLhttps://arxiv.org/abs/2508.11380

work page arXiv 2025

[19] [19]

A public data set of spatio-temporal match events in soccer competitions.Scientific Data, 6:236, 2019

Luca Pappalardo, Paolo Cintia, Alessio Rossi, Emanuele Massucco, Paolo Ferragina, Dino Pedreschi, and Fosca Giannotti. A public data set of spatio-temporal match events in soccer competitions.Scientific Data, 6:236, 2019. doi: 10.1038/s41597-019-0247-7. 32

work page doi:10.1038/s41597-019-0247-7 2019

[20] [20]

Measuring the effectiveness of playing strategies at soccer.Journal of the Royal Statistical Society: Series D (The Statistician), 46(4):541–550,

Richard Pollard and Charles Reep. Measuring the effectiveness of playing strategies at soccer.Journal of the Royal Statistical Society: Series D (The Statistician), 46(4):541–550,

work page

[21] [21]

doi: 10.1111/1467-9884.00108

work page doi:10.1111/1467-9884.00108

[22] [22]

Sentence- BERT : Sentence Embeddings using S iamese BERT -Networks

Nils Reimers and Iryna Gurevych. Sentence-BERT: Sentence embeddings using siamese BERT-networks. InProceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 3982–3992, Hong Kong, China, 2019. Association for Computational Linguistic...

work page doi:10.18653/v1/d19-1410 2019

[23] [23]

Rein and D

R. Rein and D. Memmert. Big data and tactical analysis in elite soccer: Future challenges and opportunities for sports science.Journal of Sports Sciences, 34(7):639–650, 2016

work page 2016

[24] [24]

Dajo Sanders, Mathieu Heijboer, Matthijs K. C. Hesselink, Tony Myers, and Ibrahim Akubat. Analysing a cycling grand tour: Can we monitor fatigue with intensity or load ratios? Journal of Sports Sciences, 36(12):1385–1391, 2018. doi: 10.1080/02640414.2017.1388669

work page doi:10.1080/02640414.2017.1388669 2018

[25] [25]

Teresa Anguera, Jorge Campaniço, Nuno Matos, and José C

Hugo Sarmento, Rui Marcelino, M. Teresa Anguera, Jorge Campaniço, Nuno Matos, and José C. Leitão. Match analysis in football: a systematic review.Journal of Sports Sciences, 32(20):1831–1843, 2014. doi: 10.1080/02640414.2014.898852

work page doi:10.1080/02640414.2014.898852 2014

[26] [26]

StatsBomb Open Data.https://github.com/statsbomb/open-data, 2024

StatsBomb. StatsBomb Open Data.https://github.com/statsbomb/open-data, 2024. Accessed: 2024

work page 2024

[27] [27]

Turban, R

E. Turban, R. Sharda, and D. Delen.Decision Support and Business Intelligence Systems. Pearson, 2011

work page 2011

[28] [28]

P. D. Turney and P. Pantel. From frequency to meaning: Vector space models of semantics. Journal of Artificial Intelligence Research, 2010

work page 2010

[29] [29]

Learning How to V ote with Principles: Axiomatic Insights Into the Collective Decisions of Neural Networks.J

Peter D. Turney and Patrick Pantel. From frequency to meaning: Vector space models of semantics.Journal of Artificial Intelligence Research, 37:141–188, 2010. doi: 10.1613/jair. 2934

work page doi:10.1613/jair 2010

[30] [30]

Weinberg and Daniel Gould.Foundations of Sport and Exercise Psychology

Robert S. Weinberg and Daniel Gould.Foundations of Sport and Exercise Psychology. Human Kinetics, Champaign, IL, 8th edition, 2023. ISBN 978-1-7182-0759-2. 33

work page 2023