JevMade hello@JevMade.com
← Back to evaluation videos

JevMade field notes / Video guide

I Tested Jev on 300 AI Trading Decisions — The Benchmark Exposed This

The creator backtests TypeSafe's Jev model across 300 market decisions, assessing typed outputs for trade plans, win rates, and self-contradictions.

Original by The Fintech BuilderEvaluationIntermediate14 min 51 sec Published Source reviewed

Before you press play

What you’ll find in the video

  1. In the creator’s backtest, some short decisions contradicted the accompanying execution fields, which said not to trade.
  2. The creator reports that higher-confidence decisions did not produce higher returns in this sample.
  3. The post-hoc subgroup without contradictory fields performed better than the contradictory subgroup; this is an observed split, not a validated trading rule.
Worth knowing

Gemini-assisted video/transcript review. This historical-data backtest covers five tickers and small trade subgroups. It does not establish an out-of-sample trading edge, account for every live execution risk, or make high-confidence decisions safe for real capital.

I Tested Jev on 300 AI Trading Decisions — The Benchmark Exposed This