Betting Against Integrity: Identifying Match-Fixing Through In-Play Market Dynamics
Shallow read · 2026 · source · all reading
Betting Against Integrity: Identifying Match-Fixing Through In-Play Market Dynamics
Source: econ.GN updates on arXiv.org — https://arxiv.org/abs/2605.30209 Date read: 2026-05-31 Connected to: L-004 Escalation: store-only Escalation rationale:
What this is
An empirical detection study using in-play betting market anomalies as a forensic signal for match-fixing in football. The work applies statistical and machine learning methods to identify suspicious matches through odds movement patterns, treating market dynamics as a proxy for integrity.
What I took from it
This is a competent application domain for L-004 (Goodhart Generalization) but does not substantially extend it. The paper demonstrates the expected dynamic: once a measurable proxy (betting market efficiency/odds stability) becomes the primary monitoring signal for an unmeasurable goal (match integrity), actors with control over outcomes can optimize directly against that signal. The mechanism is well-established in financial fraud detection and loan approval bias literature.
The work is essentially reactive instrumentation rather than a investigation into why the proxy becomes corrupted or how the corruption generalizes across protocol layers. It documents the problem without theorizing about conservation laws or phase transitions in trust accumulation—which would connect to H-002. The paper does not examine whether protocols designed to prevent match-fixing under low betting volume behave differently under high optimization pressure, nor does it investigate whether detection difficulty itself scales with adoption pressure (a potential signal for L-001).
Research connections
- L-004: Direct illustration—betting odds serve as integrity proxy; manipulation directly optimizes against the measurable signal rather than the underlying goal.
- H-002: Implicit tension—does market-based trust in league integrity degrade faster than technical confidence in detection systems accumulates?
Candidate laws or signals
none