FiClaw
-
Financial Agent Platform
中文

Momentum Rotation

A-share momentum rotation: turning trend strength into a monthly stock process

Momentum rotation tests whether relatively strong stocks continue to outperform for a limited period. It is not a rule to chase price: the universe, lookback, skip period, portfolio size, rebalance date and exits must be fixed before a result is meaningful.

Typical holding period

Usually weekly to monthly; choose through robustness review, not one backtest.

Rebalance and trigger

Use a fixed weekly or monthly schedule and a stated executable price.

Research universe

A liquid index constituent universe or a clearly defined all-market universe.

Research hypothesis

Write the hypothesis before reading the backtest

After short-term reversal, liquid stocks with stronger medium-term returns may retain trend exposure. The relationship can change with market regime, crowding and trading cost.

This guide is for research education and backtest design. It is not investment advice or a return promise; results depend on sample, cost, regime and execution.

Strategy rules

Turn the idea into zero-discretion rules

Medium-term return

Rank 3 to 12-month returns, optionally skipping the latest month.

Measures trend while separating short-term reversal.

Tradability filter

Set thresholds for trading value, suspension days and listing age.

Keeps untradeable names out of an idealised portfolio.

Risk constraints

Cap single-name, sector and high-volatility exposure.

Stops a small group of winners from dominating the book.

Implementation

From signal to portfolio

  1. 1Fix windows, skip period, portfolio count and rebalance date before testing.
  2. 2Specify equal weight, risk weight or another deterministic sizing rule.
  3. 3Preserve actual holdings when limit-up buys, limit-down sells or suspensions block a rebalance.
  4. 4Report turnover, unfilled orders and sector exposure beside the equity curve.

A-share constraints

Execution belongs in the test

  • Price limits and suspensions can make a correct signal impossible to trade at the assumed close.
  • Use historical index membership, delisting and ST status to avoid survivorship bias.
  • Inspect sharp reversals separately because they can combine high turnover with large drawdowns.

Validation checklist

Try to break the strategy first

Prefer stable regions that survive costs, regime changes and parameter perturbation to a single maximum return.

Compare nearby lookback, skip-period and portfolio-count settings; seek plateaus rather than peaks.
Split returns, drawdown, win rate and turnover across bull, range-bound and reversal regimes.
Raise slippage and impact assumptions by 1.5 to 2 times before judging viability.
Use rolling out-of-sample or walk-forward checks against an equal-weight universe and benchmark.

Risk map

How the strategy can fail

Trend reversal and crowdingTurnover and cost dragSector concentrationPrice-limit execution gaps

Research prompt

A first backtest specification

“In CSI 300 constituents, test a monthly momentum rotation strategy using six-month returns with a 20-trading-day skip, holding the top 20 names equally. Use historical membership and model fees, slippage, price limits and suspensions. Report annual results, turnover, unfilled orders, sector exposure, out-of-sample performance and nearby parameter results.”

FAQ

Common questions

Is a longer momentum window always better?

No. Longer windows can be smoother but slower, while shorter windows can increase turnover. Compare nearby windows on a net-of-cost basis.

Why use a skip period?

A skip period may reduce overlap with short-term reversal. Its length is a testable assumption, not a universal default.

Next research step

Fix research conventions before the backtest

Use the related factors, data boundaries and out-of-sample review to turn a strategy idea into an auditable research record.

Open A-share factor libraryReview A-share data boundaries