Pith. sign in

REVIEW 3 major objections 6 minor 63 references

A single multi-tenant cloud record, a video-linked individual development plan, and a certified analyst marketplace can close the tooling gap facing non-elite football clubs.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-31 04:27 UTC pith:BMEP2ODW

load-bearing objection Honest design/position paper with clean requirements-to-architecture trace and a standard-but-well-specified event model; no efficacy claims and none smuggled. the 3 major comments →

arxiv 2607.28531 v1 pith:BMEP2ODW submitted 2026-07-30 cs.CY

AIx4Soccer: A Unified Platform Architecture for Football Club Management and Structured Athlete Development

Frederico Falconi Costa , Salvador Cesar Costa , Fabricio F. Costa This is my paper
classification cs.CY
keywords sports informaticsmulti-tenant SaaSathlete developmentevent-centric data modelknowledge graphtwo-sided marketplaceyouth footballdata protection
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Football clubs outside the elite tier run on fragmented tools for video, GPS, medical records, scouting, and admin, so development is often undocumented and analysis capacity is scarce. This design paper proposes AIx4Soccer One Platform: a multi-tenant cloud SaaS that keeps one integrated club and athlete record across administration, squad, training, development, and video. It embeds the PDI Framework, a recurring assess–goal–plan–evidence–review cycle mapped to established development science and tied to video clips, plus Tak Tik, a certified two-sided marketplace that matches clubs to video analysts under a 75/25 analyst/platform split. The authors also specify a future substrate in which every fact is a typed immutable event in an append-only log that induces a growing knowledge graph. A sympathetic reader cares because markets such as Brazil export the most professional players while the base of the pyramid remains under-tooled, and structured records could support both athlete development and duty of care for the many who never turn pro.

Core claim

The paper’s central claim is architectural: a multi-tenant cloud operating system that unifies club-management workflows, operationalizes a cyclical multi-dimensional individual development plan linked to video evidence, and clears analysis labor through a certified marketplace is a suitable way to address tool fragmentation, professionalization gaps, and undocumented development outside elite clubs—with a formal event-centric append-only log as the preferred long-term data model. It is a design and early-deployment position, not an efficacy study.

What carries the argument

The event-centric semantic data model: every fact is a typed immutable event in a per-tenant append-only log that induces a knowledge graph whose edges are dated, attributable, and evidence-backed; module views are folds over that log. It unifies match, training, medical, PDI, video, admin, and marketplace facts and supplies audit trails, as-of queries, and the substrate for small domain event models.

Load-bearing premise

That integration and affordable access—not coaching quality, socioeconomic context, or staff willingness to run review cycles—are the binding constraints, so clubs will adopt the system and documentation will change practice.

What would settle it

Run the planned staged studies at the test club and later sites: measure whether staff complete PDI review cadences, whether clubs hire certified Tak Tik analysts, and—with ethics review—whether video-evidenced development records change coaching practice or welfare documentation rates versus baseline ad-hoc methods.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Non-elite clubs get one connective record instead of reconciling siloed point tools.
  • PDI-style immutable development history can serve as evidence for training-club certification and formation-compensation regimes.
  • Clubs without in-house analysts can buy certified video analysis that lands directly in the athlete record and PDI goals.
  • Federation-level tenants can aggregate pathway and oversight views across subordinate clubs.
  • The populated event log becomes both a youth-data compliance trail and training data for small, tenant-bound domain models rather than frontier APIs.

Where Pith is reading between the lines

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

  • Clubs that populate a typed event graph early would own longitudinal development sequences that match-event vendors never see, which could rebalance who trains the next generation of football models.
  • Coach burden and review fatigue are the first practical failure modes to instrument, given how elite academy reforms already strain staff with documentation load.
  • Encoding guardian consent as first-class events is a reusable pattern for any youth-sport platform under strict minor-data laws, beyond this product.
  • The supply-skewed 75/25 split is classic seed pricing for two-sided markets; if cross-side effects lock in, the platform share could later adjust without the paper needing to claim a fixed economics forever.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. This systems/position paper proposes a conceptual architecture for AIx4Soccer One Platform: a multi-tenant cloud SaaS that unifies club administration, squad, training, development, and video workflows for institutional users, especially outside elite European clubs. It embeds a cyclical individual-development methodology (PDI) linked to video evidence, a certified two-sided analyst marketplace (Tak Tik) with a disclosed 75%/25% revenue split, and a formal, implementation-independent event-centric data model (typed immutable events in a per-tenant append-only log inducing a knowledge graph; Eqs. 1–8) proposed as a future substrate. Requirements R1–R8/N1–N4 are stated explicitly and traced into modules, roles, and the formalism. The authors repeatedly disclaim efficacy and report only early deployment at one Brazilian club/academy, with staged evaluation deferred.

Significance. If the design is sound as an integrative architecture—and within the paper’s stated non-claims it largely is—the contribution is a clear requirements-to-architecture mapping for a real digital-divide problem in grassroots and developing-market football, plus an honest regulatory and fairness discussion (LGPD, Digital ECA, GDPR; relative-age bias) and a well-motivated case for small domain models over frontier LLMs as the intelligence layer. Strengths include explicit requirements traceability, a clean event-sourcing formalization (immutability, prefix log, fold-derived read models, PDI/marketplace as event projections), and disciplined scope control (no fabricated outcomes). Novelty is integrative and methodological rather than algorithmic; significance for cs.CY/sports-informatics readers depends on whether the journal values design-science artifacts with deferred evaluation.

major comments (3)
  1. [§5, Table 3] §5 and Table 3: The claimed PDI/TBIL methodology contribution is only partially inspectable. The cycle shape (assess–goals–plan–evidence–review) is mapped to established IDP/Four-Corner/LTAD/deliberate-practice concepts, but “TBIL” content is explicitly withheld as proprietary. For a paper that lists the PDI Framework as a primary contribution (§1), either disclose enough pedagogical structure to make the methodology contribution falsifiable and reusable, or reframe the contribution strictly as platform support for a standard IDP cycle (with proprietary content out of scope). As written, the methodology claim exceeds what a reader can evaluate.
  2. [§3.3, §4, §7] §4 vs §3.3/§7: The implemented architecture is described as conventional modules over a shared tenant-isolated record (§4), while the formal event model is only a “proposed future substrate” and “north star,” not the current system of record. The Materials and Methods framing (§3) presents Eqs. (1)–(8) as the platform’s specification, which overstates present coupling. Please separate clearly (i) what is deployed today, (ii) what is specified for incremental migration, and (iii) which requirements are already satisfied without the event log. A short migration/compatibility subsection would make the central architectural claim checkable.
  3. [§3.4, §8, §9.1] §3.4 and §8: As a design-science paper citing Hevner et al., the evaluation plan is too thin to support even modest “suitability” claims beyond internal consistency. Stage (1)–(3) are named but lack concrete instruments (e.g., task-based usability metrics, adoption/engagement operationalizations, ethics-review timeline, success/failure criteria, and what would falsify the §9.1 binding-constraint thesis). Strengthen §3.4 with measurable evaluation questions tied back to R1–R8/N1–N4 so the deferred empirical program is specific enough to be assessed.
minor comments (6)
  1. [§3.3, Eq. (1)] Eq. (1): the codomain of t is written as t∈T while T is also used for the event-type set; use a distinct time sort (e.g., 𝕋) to avoid clashing with 𝒯.
  2. [§3.3, Eq. (8)] Eq. (8): |V_c(t)|=O(1) is too strong for a growing academy/federation tenant; “bounded relative to |L_c|” or “slowly growing” would match the intended volume argument.
  3. [§4–§7] Figures 1–6 are described but not visually available in the text dump; ensure captions stand alone and that role/permission edges in Fig. 2 match the least-privilege narrative in §4.4.
  4. [References] Reference [21] (Moran et al. 2024) is incomplete (title bracketed as placeholder). Complete bibliographic details.
  5. [Front matter] Abstract/keywords are appropriate; consider adding “design science” or “event sourcing” to keywords given §3.1 and §7.3.
  6. [§6.2] §6.2: The 75/25 split is a disclosed design parameter (Eq. 12); a brief note on whether it is fixed contractually or adjustable would help readers interpret marketplace positioning without requesting full pricing disclosure.

Circularity Check

0 steps flagged

No circular derivation: design specs and definitions, not fitted or self-justifying predictions

full rationale

This is a systems/position paper that states requirements (R1–R8, N1–N4), proposes a conceptual SaaS architecture, a PDI cycle mapped to external development literature, a marketplace design choice (75%/25%), and a formal event-sourcing-style data model (Eqs. 1–8). Nothing is presented as an empirical prediction or first-principles derivation that reduces to its own fitted inputs. Eq. (11) defines a cycle-progress ratio from the development sub-log; Eq. (12) states a disclosed revenue-split design parameter—both are definitions, not forecasts. Literature mappings (Table 3) and citations to VAEP, SPADL, Rochet–Tirole, Fowler event sourcing, etc. illustrate compatibility with external work; there is no load-bearing self-citation chain, uniqueness theorem imported from the authors, or ansatz smuggled from prior author papers. The paper repeatedly disclaims efficacy and reports no outcome data. Residual founder/patent interest is not circularity under the stated rules. Derivation chain is self-contained specification, not circular.

Axiom & Free-Parameter Ledger

2 free parameters · 6 axioms · 4 invented entities

The load-bearing content is a requirements-driven systems design resting on established platform-economics, multi-tenancy, event-sourcing, and athlete-development literatures, plus product-specific constructs (PDI/TBIL, Tak Tik split, AIx4Soccer tenancy) introduced without empirical validation. Free parameters are design choices and literature-cited event-volume scales, not fits to outcome data. Invented entities are software/methodology artifacts, not physical posits; none have independent external evidence of efficacy yet.

free parameters (2)
  • Tak Tik revenue split (π_analyst=0.75p, π_platform=0.25p) = 75%/25%
    Hand-chosen design parameter disclosed for positioning (§6.2, Eq. 12); not derived from data in the paper; central to the marketplace side of the contribution.
  • Expected recorded events per match E[N_match] = ≈1.6–2.0×10^3
    Order-of-magnitude volume used to motivate unbounded log growth (Eq. 8, §7.1); taken from commercial/open event datasets rather than measured on the platform.
axioms (6)
  • domain assumption Append-only immutable event logs with state-as-fold (event sourcing) provide auditability, temporal queries, and retroactive read models suitable for compliance and development records.
    Invoked throughout §3.3 and §7.3 via Eqs. (3)–(7) and citations to Fowler/Kleppmann/industrial event-sourcing studies; assumed beneficial despite known schema-evolution costs.
  • domain assumption Two-sided platform participation depends on price structure across sides and on quality signals under incomplete information (Rochet–Tirole; Akerlof/Spence).
    Grounds Tak Tik certification and skewed supply-side split in §6; standard economics, not re-derived.
  • domain assumption Multi-dimensional cyclical IDPs (assess–goals–plan–evidence–review) mapped to FA Four-Corner/LTAD/deliberate practice are the right operational unit for athlete development documentation.
    §5 and Table 3 treat this shape as given by sport-pedagogy literature; efficacy in grassroots Brazilian settings is not shown.
  • ad hoc to paper Entities relate only through events; application views are pure projections of the per-tenant log under logical tenant isolation.
    Design commitment of the proposed substrate (Eqs. 5–7, N1); implementation-independent but not forced by nature—chosen architecture.
  • domain assumption Small domain-specific models inside the compliance perimeter are preferable to frontier LLM APIs for core system-of-record duties on minors’ data.
    §9.5 argument from hallucination risk, LGPD/Digital ECA, and SLM empirical literature; normative design stance, not measured here.
  • standard math Standard mathematical structures: totally ordered time, prefix-ordered logs, labeled multigraphs, left folds, subsumption partial order on event types.
    Used to define Eqs. (1)–(8) without novel mathematics.
invented entities (4)
  • AIx4Soccer One Platform (multi-tenant club OS) no independent evidence
    purpose: Unify administration, squad, training, PDI, video, and communication on one tenant-isolated record for clubs/federations.
    Central product artifact; only early single-club deployment reported; no public efficacy or scale evidence.
  • PDI Framework / PDI/TBIL methodology no independent evidence
    purpose: Operationalize cyclical, video-evidenced, athlete-owned individual development plans inside the platform.
    Named methodology with proprietary TBIL pedagogical content explicitly undisclosed (§5.3); shape traced to prior IDP literature but the branded method is new to this paper.
  • Tak Tik certified video-analysis marketplace no independent evidence
    purpose: Match clubs lacking analysts to certified supply under a fixed 75/25 split and platform certification signal.
    Companion two-sided market design; no liquidity, match-rate, or quality outcome data provided.
  • Proposed event-centric semantic substrate (typed event tuple + per-tenant append-only log → knowledge graph) no independent evidence
    purpose: Serve as future system-of-record unifying match, training, medical, DEV, video, admin, and marketplace facts.
    Formalized in §3.3/§7 as north-star architecture, explicitly not current implementation; builds on known event sourcing and sports event vocabularies.

pith-pipeline@v1.2.0-daily-grok45 · 22113 in / 4099 out tokens · 83742 ms · 2026-07-31T04:27:04.414745+00:00 · methodology

0 comments
read the original abstract

Football clubs, academies, and federations operate a growing but fragmented portfolio of digital tools: separate systems for video analysis, GPS/performance tracking, medical records, scouting, and administration. This fragmentation is most acute outside the elite European clubs that can afford integration, producing a digital divide that disadvantages grassroots clubs in developing markets such as Brazil, paradoxically the world's largest exporter of professional players. This paper presents, at a conceptual level, the architecture of "AIx4Soccer One Platform," a multi-tenant cloud SaaS operating system that unifies club-management workflows and embeds a structured athlete-development methodology, the PDI Framework (Plano de Desenvolvimento Individual / Individual Development Plan). We describe two companion components: "Tak Tik," a certified two-sided marketplace connecting clubs with video analysts under a 75%/25% (analyst/platform) revenue split, and the PDI/TBIL methodology, which links development plans to video evidence and periodic review. As Materials and Methods, we state explicit requirements and give a formal, implementation-independent specification of the platform's proposed future substrate: an event-centric semantic data model in which every fact is a typed, immutable event in an append-only log that induces a growing knowledge graph. We situate the design against the literature on athlete-development frameworks, sports-analytics workflows, two-sided-market economics, and multi-tenant SaaS patterns, and discuss youth data-protection obligations (Brazil's LGPD and 2025 Digital ECA; the EU GDPR), algorithmic-fairness risks in talent evaluation, and why small, domain-specific models, rather than frontier LLMs, are the appropriate intelligence layer. This is a design and early-deployment paper, not an empirical evaluation, making no efficacy claims.

Figures

Figures reproduced from arXiv: 2607.28531 by Fabricio F. Costa, Frederico Falconi Costa, Salvador Cesar Costa.

Figure 1
Figure 1. Figure 1: Layered architecture of AIx4Soccer One Platform (conceptual). All module reads/writes pass through the tenant-isolated data layer. 9 [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Stakeholder roles and data flows (conceptual). Solid arrows denote writes; dashed arrows denote scoped reads. Tak Tik connects through the Integration Layer. 4.5 Comparison with the fragmented status quo [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The PDI cycle (conceptual). The athlete sits at the hub as owner of the plan; video evidence links goal-setting and review; guardians participate in reviews for minors. 5.3 Mapping PDI to established concepts [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Tak Tik two-sided market (conceptual): matching and certification at the center, cross-side network effects, and the 75%/25% analyst/platform payment split. 14 [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: summarizes the complete model as a wheel. At the hub sits the athlete and their integrated record; the inner ring is the PDI development cycle that continuously turns around the athlete; the middle ring comprises the six platform modules that operationalize club workflows; and the outer ring holds the stakeholder roles who read from and write to the record under least-privilege permissions. Two orbitals co… view at source ↗
Figure 6
Figure 6. Figure 6: Proposed event-centric semantic data model: the EVENT (hexagon) as the universal unit of record with who/what/when/where/outcome/evidence semantics; an event-type taxonomy (inner ring); en￾tities linked to one another only through events (outer ring); an append-only log that grows into a knowledge graph as the platform is populated. Formal counterpart: Eqs. (1)–(8). 17 [PITH_FULL_IMAGE:figures/full_fig_p0… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

63 extracted references · 12 canonical work pages

  1. [1]

    doi:10.1162/154247603322493212

    Rochet, J.-C., &Tirole, J.(2003).PlatformCompetitioninTwo-SidedMarkets.Journal of the European Economic Association, 1(4), 990–1029. doi:10.1162/154247603322493212

    doi:10.1162/154247603322493212 2003

  2. [2]

    Rochet, J.-C., & Tirole, J. (2006). Two-Sided Markets: A Progress Report.RAND Journal of Eco- nomics, 37(3), 645–667

    2006

  3. [3]

    A., Krampe, R

    Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The Role of Deliberate Practice in the Acquisition of Expert Performance.Psychological Review, 100(3), 363–406

    1993

  4. [4]

    Academic Emergency Medicine, 15(11), 988–994

    Ericsson, K.A.(2008).DeliberatePracticeandAcquisitionofExpertPerformance: AGeneralOverview. Academic Emergency Medicine, 15(11), 988–994. doi:10.1111/j.1553-2712.2008.00227.x

    arXiv 2008

  5. [5]

    N., & Maitra, M

    Macnamara, B. N., & Maitra, M. (2019). The role of deliberate practice in expert performance: revisiting Ericsson, Krampe & Tesch-Römer (1993).Royal Society Open Science, 6(8), 190327. doi:10.1098/rsos.190327

    work page doi:10.1098/rsos.190327 2019

  6. [6]

    (2013).Long-Term Athlete Development

    Balyi, I., Way, R., & Higgs, C. (2013).Long-Term Athlete Development. Human Kinetics

    2013

  7. [7]

    Ford, P., et al. (2011). The Long-Term Athlete Development model: Physiological evidence and appli- cation.Journal of Sports Sciences, 29(4), 389–402. doi:10.1080/02640414.2010.536849

    arXiv 2011

  8. [8]

    Elite Player Performance Plan (EPPP).https://www.premierleague.com/en/ footballandcommunity/youth-development/eppp

    Premier League. Elite Player Performance Plan (EPPP).https://www.premierleague.com/en/ footballandcommunity/youth-development/eppp

  9. [9]

    The FA’s 4 Corner Model.https://www.thefa.com/bootroom/resources/ coaching/the-fas-4-corner-model

    The Football Association. The FA’s 4 Corner Model.https://www.thefa.com/bootroom/resources/ coaching/the-fas-4-corner-model

  10. [10]

    doi:10.1080/02640414.2024.2335583

    Professional soccer players’ perceptions of the English FA 4-corner model in supporting first-team transitions (2024).Journal of Sports Sciences. doi:10.1080/02640414.2024.2335583

    arXiv 2024

  11. [11]

    Michel Sablon: Communicating a national playing philosophy to professional clubs.https://www.fifatrainingcentre.com/

    FIFA Training Centre. Michel Sablon: Communicating a national playing philosophy to professional clubs.https://www.fifatrainingcentre.com/

  12. [12]

    The Belgian Blueprint: How a small nation became a European football superpower

    CNN (2021). The Belgian Blueprint: How a small nation became a European football superpower. https://www.cnn.com/interactive/2021/06/sport/belgium-euro-2020-spt-intl/

    2021

  13. [13]

    A brief history of German youth academies.https://www.dfl.de/en/ topics/youth-academies/

    DFL Deutsche Fußball Liga. A brief history of German youth academies.https://www.dfl.de/en/ topics/youth-academies/

  14. [14]

    R., et al

    Thornton, H. R., et al. (2019). Developing Athlete Monitoring Systems in Team Sports: Data Anal- ysis and Visualization.International Journal of Sports Physiology and Performance, 14(6), 698–705. doi:10.1123/ijspp.2018-0169

    work page doi:10.1123/ijspp.2018-0169 2019

  15. [15]

    Romann, M., et al. (2021). Relative Age-Related Biases in Objective and Subjective Assessments of Performance in Talented Youth Soccer Players.Frontiers in Sports and Active Living, 3, 664231. doi:10.3389/fspor.2021.664231 22

    arXiv 2021

  16. [16]

    Kalra, S., & Prabhakar, T. V. (2018). Implementation Patterns for Multi-Tenancy (SaaS design pat- terns)

    2018

  17. [17]

    E., et al

    Whiting, M. E., et al. (2020). Reputation Agent: Prompting Fair Reviews in Gig Markets.Proceedings of The Web Conference. arXiv:2005.06022

    Pith/arXiv arXiv 2020

  18. [18]

    FIFA Talent Development Scheme.https://inside.fifa.com/talent-development

    FIFA. FIFA Talent Development Scheme.https://inside.fifa.com/talent-development

  19. [19]

    International transfers reach historic high in 2024 (Global Transfer Report 2024).https:// inside.fifa.com/

    FIFA. International transfers reach historic high in 2024 (Global Transfer Report 2024).https:// inside.fifa.com/

    2024

  20. [20]

    Monthly Report no100 (May 2025): player-exporting nations 2020–2025

    CIES Football Observatory. Monthly Report no100 (May 2025): player-exporting nations 2020–2025. https://football-observatory.com/

    2025

  21. [21]

    Moran, J., et al. (2024). [Study on academy-to-professional progression rates.]International Journal of Sports Science & Coaching. University of Essex

    2024

  22. [22]

    29, as amended by Lei no 12.395/2011

    Lei no 9.615, de 24 de março de 1998 (Lei Pelé), Art. 29, as amended by Lei no 12.395/2011. Presidência da República / Planalto

    1998

  23. [23]

    Certificado de Clube Formador.https://portaldegovernanca

    Confederação Brasileira de Futebol. Certificado de Clube Formador.https://portaldegovernanca. cbf.com.br/certificado-clube-formador

  24. [24]

    Regulations on the Status and Transfer of Players (RSTP), Articles 20–21 and Annexes 4–5 (Training Compensation; Solidarity Mechanism).https://inside.fifa.com/

    FIFA. Regulations on the Status and Transfer of Players (RSTP), Articles 20–21 and Annexes 4–5 (Training Compensation; Solidarity Mechanism).https://inside.fifa.com/

  25. [25]

    14 (children and adolescents)

    Lei Geral de Proteção de Dados (LGPD), Lei no 13.709/2018, Art. 14 (children and adolescents). Brazil

    2018

  26. [26]

    Digital Statute of the Child and Adolescent (Digital ECA), Lei no 15.211/2025. Brazil

    2025

  27. [27]

    Regulation (EU) 2016/679 (General Data Protection Regulation, GDPR)

    2016

  28. [28]

    Impacto do Futebol Brasileiro (2018 data)

    Ernst & Young / CBF. Impacto do Futebol Brasileiro (2018 data). Confederação Brasileira de Futebol

    2018

  29. [29]

    Decroos, T., Bransen, L., Van Haaren, J., & Davis, J. (2019). Actions Speak Louder than Goals: Valuing Player Actions in Soccer.Proceedings of the 25th ACM SIGKDD Conference (KDD ’19), 1851–1861. doi:10.1145/3292500.3330758

    arXiv 2019

  30. [30]

    Pappalardo, L., Cintia, P., Rossi, A., Massucco, E., Ferragina, P., Pedreschi, D., & Giannotti, F. (2019). A public data set of spatio-temporal match events in soccer competitions.Scientific Data, 6,

    2019

  31. [31]

    J., Smith, J

    Anzer, G., Arnsmeyer, K., Bauer, P., Bekkers, J., Brefeld, U., Davis, J., Evans, N., Kempe, M., Robertson, S. J., Smith, J. W., & Van Haaren, J. (2025). Common Data Format (CDF): A Standardized Format for Match-Data in Football (Soccer). arXiv:2505.15820

    Pith/arXiv arXiv 2025

  32. [32]

    kloppy: standardizing soccer tracking and event data (open-source library).https://kloppy

    PySport. kloppy: standardizing soccer tracking and event data (open-source library).https://kloppy. pysport.org

  33. [33]

    4085.https://ceur-ws.org/Vol-4085/paper56.pdf

    An Ontology for the Common Data Format on Football Match Data (2025).CEUR Workshop Proceed- ings, Vol. 4085.https://ceur-ws.org/Vol-4085/paper56.pdf

    2025

  34. [34]

    Gudmundsson, J., & Horton, M. (2017). Spatio-Temporal Analysis of Team Sports.ACM Computing Surveys, 50(2), 1–34. doi:10.1145/3054132

    doi:10.1145/3054132 2017

  35. [35]

    Mendes-Neves, T., Meireles, L., & Mendes-Moreira, J. (2024). Towards a foundation large events model for soccer.Machine Learning, 113, 8687–8709. doi:10.1007/s10994-024-06606-y

    work page doi:10.1007/s10994-024-06606-y 2024

  36. [36]

    Baron, E., Hocevar, D., & Salehe, Z. (2024). A Foundation Model for Soccer. arXiv:2407.14558. 23

    Pith/arXiv arXiv 2024

  37. [37]

    Fowler, M. (2005). Event Sourcing. martinfowler.com.https://martinfowler.com/eaaDev/ EventSourcing.html

    2005

  38. [38]

    Overeem, M., Spoor, M., Jansen, S., & Brinkkemper, S. (2021). An empirical characterization of event sourced systems and their schema evolution — Lessons from industry.Journal of Systems and Software, 178, 110970. doi:10.1016/j.jss.2021.110970

    arXiv 2021

  39. [39]

    (2017).Designing Data-Intensive Applications

    Kleppmann, M. (2017).Designing Data-Intensive Applications. O’Reilly Media

    2017

  40. [40]

    D., et al

    Wilkinson, M. D., et al. (2016). The FAIR Guiding Principles for scientific data management and stewardship.Scientific Data, 3, 160018. doi:10.1038/sdata.2016.18

    doi:10.1038/sdata.2016.18 2016

  41. [41]

    M., & Reilly, T

    Williams, A. M., & Reilly, T. (2000). Talent identification and development in soccer.Journal of Sports Sciences, 18(9), 657–667. doi:10.1080/02640410050120041

    work page doi:10.1080/02640410050120041 2000

  42. [42]

    T., Pereira, A., & Araújo, D

    Sarmento, H., Anguera, M. T., Pereira, A., & Araújo, D. (2018). Talent Identification and Development in Male Football: A Systematic Review.Sports Medicine, 48, 907–931. doi:10.1007/s40279-017-0851-7

    work page doi:10.1007/s40279-017-0851-7 2018

  43. [43]

    M., & Philippaerts, R

    Vaeyens, R., Lenoir, M., Williams, A. M., & Philippaerts, R. M. (2008). Talent Identification and Development Programmes in Sport.Sports Medicine, 38(9), 703–714. doi:10.2165/00007256-200838090- 00001

    work page doi:10.2165/00007256-200838090- 2008

  44. [44]

    Cobley, S., Baker, J., Wattie, N., & McKenna, J. (2009). Annual Age-Grouping and Athlete Develop- ment: A Meta-Analytical Review of Relative Age Effects in Sport.Sports Medicine, 39(3), 235–256. doi:10.2165/00007256-200939030-00005

    work page doi:10.2165/00007256-200939030-00005 2009

  45. [45]

    P., Lloyd, R

    Cumming, S. P., Lloyd, R. S., Oliver, J. L., Eisenmann, J. C., & Malina, R. M. (2017). Bio-banding in Sport.Strength and Conditioning Journal, 39(2), 34–47. doi:10.1519/SSC.0000000000000281

    work page doi:10.1519/ssc.0000000000000281 2017

  46. [46]

    F., et al

    Bergeron, M. F., et al. (2015). International Olympic Committee consensus statement on youth athletic development.British Journal of Sports Medicine, 49, 843–851. doi:10.1136/bjsports-2015-094962

    work page doi:10.1136/bjsports-2015-094962 2015

  47. [47]

    C., et al

    Bourdon, P. C., et al. (2017). Monitoring Athlete Training Loads: Consensus Statement.International Journal of Sports Physiology and Performance, 12(s2), S2-161–S2-170. doi:10.1123/IJSPP.2017-0208

    work page doi:10.1123/ijspp.2017-0208 2017

  48. [48]

    Gabbett, T. J. (2016). The training—injury prevention paradox: should athletes be training smarter and harder?British Journal of Sports Medicine, 50, 273–280. doi:10.1136/bjsports-2015-095788

    work page doi:10.1136/bjsports-2015-095788 2016

  49. [49]

    Mackenzie, R., & Cushion, C. (2013). Performance analysis in football: A critical review and implica- tions for future research.Journal of Sports Sciences, 31(6), 639–676. doi:10.1080/02640414.2012.746720

    arXiv 2013

  50. [50]

    Hughes, M., & Franks, I. M. (Eds.) (2004).Notational Analysis of Sport(2nd ed.). Routledge

    2004

  51. [51]

    Relvas, H., Littlewood, M., Nesti, M., Gilbourne, D., & Richardson, D. (2010). Organizational Struc- tures and Working Practices in Elite European Professional Football Clubs.European Sport Manage- ment Quarterly, 10(2), 165–187. doi:10.1080/16184740903559891

    work page doi:10.1080/16184740903559891 2010

  52. [52]

    Spence, M. (1973). Job Market Signaling.The Quarterly Journal of Economics, 87(3), 355–374. doi:10.2307/1882010

    doi:10.2307/1882010 1973

  53. [53]

    Akerlof, G. A. (1970). The Market for “Lemons”: Quality Uncertainty and the Market Mechanism.The Quarterly Journal of Economics, 84(3), 488–500. doi:10.2307/1879431

    doi:10.2307/1879431 1970

  54. [54]

    G., Van Alstyne, M

    Parker, G. G., Van Alstyne, M. W., & Choudary, S. P. (2016).Platform Revolution. W. W. Norton & Company

    2016

  55. [55]

    Bommasani, R., et al. (2021). On the Opportunities and Risks of Foundation Models. arXiv:2108.07258

    Pith/arXiv arXiv 2021

  56. [56]

    Ji, Z., et al. (2023). Survey of Hallucination in Natural Language Generation.ACM Computing Surveys, 55(12), 1–38. doi:10.1145/3571730 24

    doi:10.1145/3571730 2023

  57. [57]

    Gunasekar, S., et al. (2023). Textbooks Are All You Need. arXiv:2306.11644

    Pith/arXiv arXiv 2023

  58. [58]

    Abdin, M., et al. (2024). Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone. arXiv:2404.14219

    Pith/arXiv arXiv 2024

  59. [59]

    C., & Molchanov, P

    Belcak, P., Heinrich, G., Diao, S., Fu, Y., Dong, X., Muralidharan, S., Lin, Y. C., & Molchanov, P. (2025). Small Language Models are the Future of Agentic AI. arXiv:2506.02153

    Pith/arXiv arXiv 2025

  60. [60]

    Zhao, J., et al. (2024). LoRA Land: 310 Fine-tuned LLMs that Rival GPT-4, A Technical Report. arXiv:2405.00732

    Pith/arXiv arXiv 2024

  61. [61]

    Bucher, M. J. J., & Martini, M. (2024). Fine-Tuned ‘Small’ LLMs (Still) Significantly Outperform Zero-Shot Generative AI Models in Text Classification. arXiv:2406.08660

    Pith/arXiv arXiv 2024

  62. [62]

    R., March, S

    Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design Science in Information Systems Research.MIS Quarterly, 28(1), 75–105. doi:10.2307/25148625 25

    doi:10.2307/25148625 2004

  63. [236]

    doi:10.1038/s41597-019-0247-7

    doi:10.1038/s41597-019-0247-7