Definition
From the same regression as for alpha:
R_strategy = alpha + beta × R_market + residual
Beta is the slope: how much the strategy moves when the market moves by 1%.
| Beta | Meaning |
|---|---|
| ≈ 1 | moves 1:1 with the market (effectively buy & hold) |
| 0.3–0.7 | dampened market exposure |
| ≈ 0 | market-neutral — return independent of direction |
| < 0 | inverse — rises when the market falls (hedge / crisis alpha) |
Why this matters for strategy evaluation
Beta is the lie detector for "alpha". A strategy with beta near 1 is usually just more expensive buy & hold — the same ride as BTC, but with fees and complexity on top. The test for a genuine edge: does any return remain once you strip out the beta (⇒ alpha)? For a portfolio, beta is also the diversification measure: only low/zero-correlation (low-beta) sleeves truly reduce overall risk.
How Botty uses it
The edge-correlation audit (backtesting/edge_correlation_audit.py) measures each strategy's BTC beta. The architectural lesson from it: a market-neutral core (funding carry + VRP, beta ≈ 0) plus a deliberately sized beta/trend sleeve (PTJ trend, beta 0.9) — instead of naively mixing the two. Treating PTJ trend, highly correlated with buy & hold, as an independent edge would be a fallacy; it is beta and should be sized accordingly.
Limits
- Beta is regime-dependent. It can be low in calm phases and jump to 1 in crashes (correlations go toward 1 under stress) — exactly when diversification matters most.
- Linear and single-period. Does not capture nonlinear (option-like) exposures.
- The choice of market proxy decides. Beta against BTC is not beta against total crypto or global liquidity.