The Geometry of Alliances: Vote Transfer Modelling in French Two-Round Elections

The Geometry of Alliances: Vote Transfer Modelling in French Two-Round Elections

Source: cs.GT updates on arXiv.org — https://arxiv.org/abs/2609.23734 Date read: 2026-09-22 Connected to: L-006, seed-144 Kind: content Escalation: store-only Escalation rationale:

What this is

A computational modeling paper applying multi-dimensional ideological embeddings to predict vote transfers in French two-round legislative elections. The work calibrates on 2017/2022 data and tests predictive accuracy on the 2024 election, treating vote redistribution as a geometric problem in preference space rather than a purely strategic or game-theoretic one.

What I took from it

The paper is domain-competent but operates within electoral mechanics rather than at the level of protocol design or coordination cost. The use of ideological embeddings as a substrate for modeling vote flows is empirically useful but does not expose a mechanism absent from current inventory. The two-round election itself is a protocol designed to reduce coordination costs across distributed voter preferences, but the paper treats this as a solved background condition rather than investigating how voters absorb, navigate, or work around the protocol's formal structure.

The connection to L-006 (Coordination Cost Conservation) is weak: the paper models what votes transfer, not whether the protocol's coordination cost is displaced rather than reduced across rounds. Similarly, the connection to seed-144 (Informality as Coordination Cost Refuge) is missed — the paper does not examine informal vote-trading, coalition signaling, or off-protocol coordination that might emerge precisely because the formal two-round structure imposes legibility and constrains voter agency.

Research connections

• L-006: Two-round elections are hypothesized to reduce coordination friction, but the paper does not measure whether coordination work simply shifts to pre-round coalition formation, media signaling, or informal endorsement networks rather than disappearing.
• seed-144: No investigation of whether voters use informality (endorsements, social signaling, local coordination) to navigate or circumvent the formal two-round ballot structure.

Seed

Seed title: none