The Q-PI + ML chip is the Fear & Greed forecast's default selection. It combines the two strongest ideas in the switcher: the quantum path-integral sampler's multi-well dynamics (which can produce two-scenario, bimodal forecasts) and the machine-learning drift trained on the index's own sub-indicators.
It is the default because it ranked best on MAPE and RMSE in the walk-forward backtest at the time it was made default — not because we assert it is permanently the best. That distinction matters: a walk-forward ranking can shift as new data arrives. So this choice is presented as something you can verify or refute yourself, not a claim to take on trust. The "Model performance" table under the Fear & Greed forecast re-runs that exact backtest on current data and shows every model's coverage, MAPE, RMSE, Brier and hit-rate side by side. If another model is beating it today, the table will show you.
The path-integral half gives the forecast a realistic shape — including the ability to split into two scenarios when the market faces a fork — while the ML half lets the sub-indicators tilt the drift when they turn early. Neither guarantees accuracy; combining them is a hypothesis the backtest table is there to test. When in doubt, the plain OU baseline is the honest yardstick to compare against.
Because it ranked best on MAPE and RMSE in the walk-forward backtest at the time it was made default — not because we claim it is permanently best. You can verify or refute that in the 'Model performance' table under the chart.
Possibly — a walk-forward ranking can shift as new data arrives. That's exactly why the live 'Model performance' table re-runs the backtest on current data and shows every model side by side, so you can check rather than assume.
The path-integral half gives the forecast a realistic shape (including two-scenario bimodal forecasts); the ML half lets the sub-indicators tilt the drift when they turn early. Combining them is a hypothesis the backtest table is there to test.
Browse all S&P 500 tickers to see this metric applied to individual companies.
Educational research only — not investment advice.