Skip to content

A Thermometer, Not an Analyst

Working paper placing a text judge beside a live trading system as a read-only signal, comparing Jev and Claude as blind headline judges.

Paperresearch

This summary is based on how curated lists describe the project; its README was not available to review.

Overview

A working paper by Alex Vidovich that adds a text judge to a live trading system strictly as a reader: it measures headlines but never prescribes trades, hence "a thermometer, not an analyst".

It compares Jev with Claude as blind judges of news headlines and puts six gates in front of either before its output can influence action. It is aimed at practitioners thinking about how to let language models touch financial decisions safely.

What the decision model decides

TypeQuestionAnswers
ScoreHow does this headline bear on the market being traded?(inferred)—

How it works

  1. Feed headlines blind to the judge
  2. Get Jev and Claude judgments
  3. Pass judgments through six gates
  4. Only then allow any effect on the trading system

Strengths

  • Clear separation between measuring and acting
  • Head-to-head comparison of a decision model and an LLM

Caveats

  • Working paper; results not reproduced here
  • Only list summary available

At a glance

Type
Paper
Status
research

Use cases and integrations

How curated lists describe it

  • Working paper: puts a text judge next to a live trading system only as a reader that never prescribes, and compares Jev with Claude as a blind headline judge behind six gates before either can act.

    yibie/awesome-jev

Frequently asked questions

What is compared?

Jev and Claude as blind judges of headlines next to a trading system.

Does the judge place trades?

No, it only reads; six gates sit before any action.

Where is it?

https://alexvidovich.com/papers/text-judge

Related projects

  • Checks whether each cited paper supports its citing sentence: code fetches papers, Claude proves a quote, Jev scores it, a human decides.

    Serviceproduction
  • sort where the key is a description: Jev judges pairs of texts and a Bradley-Terry fit turns them into a scored ranking.

    ProjectbetaPython
  • Chinese open-source research and trading workbench whose JEV monitor watches futures contracts and judges entries and exits.

    ProjectbetaTypeScript
  • Self-hosted trading analysis platform pairing Claude briefings with Jev probabilities for signals like thesis-breaking headlines.

    Servicebeta
  • Watches X for updates relevant to your YAML profile, has Jev classify and score each post, and alerts in Slack.

    ProjectbetaTypeScript
  • Stock terminal that gathers live evidence via Valyu and asks Jev for the probability a stock trades higher in 30 days.

    ProjectprototypeTypeScript