New AI Models Value Offensive Handball Actions.

Julius Broermann, Oliver M\"uller, Michael D\"oring, Jochen Baumeister· August 14, 2026 View original

Key takeaways

  • New models, H-xT and H-VAEP, adapt advanced football analytics to handball.
  • These models provide more robust and nuanced player evaluations than traditional metrics.
  • H-VAEP specifically excels at highlighting contributions to offensive build-up play.
  • The open-source release facilitates adoption by professional handball clubs.

Who benefits

SportsData AnalyticsEntertainment

Summary

This paper introduces Handball-xT (H-xT) and Handball-VAEP (H-VAEP), the first adaptations of advanced football analytics frameworks for professional handball, using five seasons of Bundesliga data. These models provide stable and discriminative player ratings by valuing on-the-ball actions and build-up play.

Traditional methods for evaluating handball players often rely on simple statistics, which fail to capture the full impact of multi-player offensive sequences. Drawing inspiration from successful football analytics frameworks like Expected Threat (xT) and Valuing Actions by Estimating Probabilities (VAEP), researchers have now developed specialized versions for handball. These new models, named Handball-xT (H-xT) and Handball-VAEP (H-VAEP), leverage extensive event data from the Handball Bundesliga. H-xT uses a handball-specific court zoning layout, proving more robust than standard grids in simulations. H-VAEP was optimized by refining its feature space and context length to prevent bias from team identity. Evaluations show H-VAEP produces highly consistent, distinct, and intuitive player ratings that effectively highlight contributions to offensive build-up play. The complete code repository is being released to assist professional clubs in deploying these analytical tools.

Why it matters

Sports analytics professionals can leverage these new models to gain deeper insights into player performance beyond traditional metrics, identifying undervalued contributions to offensive play.

How to implement this in your domain

  1. 1Download and integrate the open-source code repository into existing sports analytics platforms.
  2. 2Collect detailed tracking-derived event data for handball matches, similar to the Bundesliga dataset used.
  3. 3Apply H-xT to analyze court positioning and threat generation during offensive plays.
  4. 4Utilize H-VAEP to generate comprehensive player ratings that account for build-up contributions.
  5. 5Train coaching staff and scouts on interpreting these advanced metrics for player development and recruitment.

Original post by Julius Broermann, Oliver M\"uller, Michael D\"oring, Jochen Baumeister

"arXiv:2608.12926v1 Announce Type: new Abstract: Traditional player evaluation in professional handball relies on basic box-score metrics or heuristic indices, which fail to credit the multi-player build-up chain. While football (soccer) analytics has adopted Expected Threat (xT)…"

View on X

Originally posted by Julius Broermann, Oliver M\"uller, Michael D\"oring, Jochen Baumeister on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses

More in AI Research

AI Engineering & DevToolsAI ResearchAI Investing

FlowLOB Generates Realistic, Controllable Limit Order Books Efficiently

This paper introduces FlowLOB, a conditional flow-matching generator for Limit Order Book (LOB) trajectories that offers realistic market dynamics, efficient sampling, and controllable scenario generation, outperforming existing agent-based and deep generative simulators. FlowLOB achieves high fidelity with significantly fewer computational steps than diffusion models and transfers effectively to unseen instruments.

Zhuohan Wang, Andreea Bacalum, Ollie Olby, Carmine Ventre, Namid StillmanAug 14, 2026
AI Engineering & DevToolsAI Research

Auditing Reveals Bias in Neural Combinatorial Optimization Benchmarks

This paper audits test-time budget allocation in Neural Combinatorial Optimization (NCO) solvers, revealing that reported gains from non-uniform sampling often stem from "sampling luck" rather than true allocation benefits on in-distribution data. It proposes a correction procedure and demonstrates real gains under distribution shift, emphasizing the need for rigorous evaluation.

Jinhyung BaeAug 14, 2026
AI Engineering & DevToolsAI Research

Diffusion Models Solve Mixed-Integer Optimization Problems Faster

This paper introduces Constrained Graph Diffusion (CGD), a novel learning-based approach that uses a graph-based generative diffusion model to approximately solve mixed-integer optimization problems (MIPs). CGD integrates a training-free feasibility projection operator into the diffusion process, significantly improving solution quality and feasibility while achieving substantial speedups over traditional numerical solvers.

Vincenzo Di Vito, Mehdi Taghizadeh, Deepjyoti Deka, Kaarthik Sundar, Ferdinando FiorettoAug 14, 2026