Research note
Sector rotation: separate policies, separate evidence
A–E, index credit risk control and crisis studies have different bases. No blanket rotation PASS or per-policy frozen result is inventoried.
Backtest verdict: Unknown — no frozen result recorded in the inventory. There is no single page-wide verdict for rotation A–E, the index credit/bonds reference or the separate crisis study. Credit risk control is not proof of sector-selection alpha.
Open this backtest in Backtesting
The rule
The five-stage Stovall model maps each stage to textbook leader sectors. A tests stage/sector forward returns; B holds an equal-weight top-three leader basket and rebalances on stage-label changes; C compares five-stage and six-regime classifications. D compares complete rotation policies; E tests rebalance/overlay variants, including conditional and experimental cases.
The rotation workbench opens at five years and 5 bps each side. Policy controls include acceleration, momentum lookback, composite gate/exposure thresholds, range overextension and optional macro optics; E has its own relative-strength lookback. The retained Lab configuration fixes top-three and thirteen-week momentum with its own gate/exposure defaults. These variants must not be collapsed into one rule. Each native policy defines cash/bond/short behavior and timing; no universal stop/target is imposed.
The index credit/bonds reference uses a twenty-year tradeable total-return basis. The separate long-history crisis experiment uses its fixed 10 bps per side, not the workbench slider or an assumed cost for that reference.
Methodology
Evidence as of: the INVENTORY review dated November 5, 2026; this is the consolidation planning baseline, not a new observation window. Publication does not extend any study's sample.
Current stage, macro-credit and holdings context are live reads, separate from study results. Preserve the price-only in-app comparisons separately from total-return index evidence. Publication lags, revised macro series, ETF availability, proxies and policy selection matter. Historical label strings stay as recorded.
D13 inventories a credit/bonds Sharpe/drawdown comparison, but its exact dates, cost and lag are unresolved in that row. This article withholds those headline values pending a provenance-complete transcription; it does not substitute the crisis experiment's cost. Per-policy returns, trades and a blanket PASS are not recorded. The crisis view keeps all thirteen episodes, calm controls, overall results and timing/cost notes; episode selection is not independent OOS evidence.
Results
| Study / variant | Inventory evidence | Window / basis | Strategy return / CAGR | Benchmark | Max drawdown | Closed trades / observation N | Significance / IC | Native verdict |
|---|---|---|---|---|---|---|---|---|
| A / B / C | D13 / F21–F23 / H26–H28 | not recorded | not recorded | not recorded | not recorded | not recorded | not recorded | Unknown — no frozen result |
| D policies / E variants | D13 / F24–F25 / H29–H30 | not recorded | not recorded | not recorded | not recorded | not recorded | not recorded | Unknown — no frozen result |
| Index credit/bonds reference | D13 / F53 | 20y tradeable total return; exact dates/cost/lag unresolved | not recorded | not recorded | not recorded | not recorded | not recorded | Unknown — risk control, no blanket PASS |
| Crisis / calm / overall | D13 / F26 / H32 | separate long-history study; 10 bps/side | not recorded | not recorded | not recorded | not recorded | not recorded | Unknown — native per-case results not transcribed |
The linked workspace retains the complete native output and notes when available. Missing evidence is not a zero return, a successful empty result or a failed strategy.
Verdict and limitations
Unknown — no page-wide result/verdict recorded in this inventory. The index credit finding concerns risk control, not sector-selection alpha. Price and total-return series cannot be combined into one ranking. A conditional, skipped, unavailable or experimental policy stays labelled as such; no new live rule or vote follows from moving its display.
Sources and live backtest
INVENTORY D13, F20–F26/F53, H26–H32/H58/H61. Sources: frontend/components/backtest/StovallBoards.tsx, CreditGateCrisisChart.tsx, frontend/lib/strategy-pages.ts, backend/app/scoring/stovall_backtest.py, rotation_strategies.py, stovall_variants_backtest.py and credit_gate_crisis.py.
Read the rotation/credit research, long-history study and out-of-sample credit report. Open all A–E policies and crisis/calm detail. Open evidence question: exact dates/cost/lag/basis for the twenty-year headline and each per-policy payload; no fresh study is authorized to fill it.