Can conflict risk time defense stocks well enough to beat the S&P 500?
The signal is the Caldara–Iacoviello daily Geopolitical Risk index, turned into a 0–10 conflict score: the 7-day mean, ranked against the trailing 5 years. Each week of scores can be traded the Tuesday after the week ends, so the backtests have no look-ahead.
V1 · Geoconflict-Trade Testerfrozen
The first interface, kept as the reference. The V2 engine reproduces it exactly.
- Defense rotation backtest on the daily conflict score, 2011 onward
- Scenario simulator: two conflict scenarios side by side
- Scenario model: editable drivers, channels and asset betas
V2 · Geoconflict-Trade Tester V2current
The current interface: backtest from 1996, paper accounts, the calibrated simulator and the live lab.
- Backtest with trend signals, sizing, costs, placebo and walk-forward tests
- Paper accounts for the frozen rules
- Scenario simulator calibrated to 11 real episodes
- Live & lab: weekly memo, decay bands, review lines and the trial ledger
- Gauntlet: each version of the theory against history
V3 · Geoconflict V3preview
Views and simulators: data visuals and conflict scenarios that read the frozen record.
- Crisis replay, escalation ladder, paper-account race and more views
Findings so far
- Out of sample (1996–2010, never used to choose the rule): the frozen S&P + conflict sprint beat the S&P by 5.4 points a year, ahead in 12 of 15 years.
- 2011–2026, the window the rules were chosen on: the sprint beats the S&P by about 3.4 points a year with realistic costs. Without 2022 it trails the S&P by about 0.5 a year.
- Defense swings about 16% a year around the S&P, so a backtest of this length cannot prove a few-point edge. The live record since 2026-10-19 is the real test.
This is research, not investment advice.