The idea is environmental rather than predictive. Instead of forecasting which regime comes next, you hold assets that each do well in at least one of them: equities for rising growth, long bonds for falling growth, commodities and inflation-linked bonds for rising inflation, nominal bonds for falling inflation.
Because no forecast is required, the portfolio accepts that one or two sleeves will always be losing. The trade is a lower peak return in a strong equity decade in exchange for shallower drawdowns when that decade ends.
It is not a guarantee. Every sleeve can lose at once when real interest rates jump, and the approach usually needs leverage on the bond side to make the risk contributions comparable. Describe it as a framework for thinking about macro exposure, not a portfolio that cannot lose money.
Related: risk-parity, asset-allocation, diversification, sixty-forty-portfolio, risk-on-risk-off