Backtest Evidence: Do PreReason Briefings Improve AI Trading Decisions?

7 controlled backtest runs measuring whether structured market briefings improve LLM trading agent performance. Treatment arms receiving PreReason briefings consistently outperform controls across different models, time windows, and market conditions.

Key Results

All 7 Runs

RunNamePeriodTicksArmsTreatmentControlDeltaModel
1Where It Started WorkingSep 1 - Nov 30, 2025913-3.16%-8.44%+5.28ppOpus 4.6
2Model-Agnostic ConfirmationSep 1 - Nov 22, 2025833-2.03%-6.49%+4.46ppSonnet 4.5
3Briefing RestructuringSep 1, 2025 - Feb 28, 20261813-10.35%-15.69%+5.34ppOpus 4.6
4Architecture BreakthroughSep 1, 2025 - Mar 4, 20261853+0.21%-14.69%+14.90ppOpus 4.6
5Confirmation RunSep 1, 2025 - Mar 6, 20261873+8.46%-6.31%+14.77ppOpus 4.6
6First 4-Arm RCTSep 1, 2025 - Mar 6, 20261874+3.18%-9.51%+12.69ppOpus 4.6
7Strongest ResultSep 1, 2025 - Mar 21, 20262024+7.83%-8.14%+15.97ppOpus 4.6

Explore the Evidence

Briefings Tested

Browse all 17 briefings | Try the briefings free