Backtested research example
🗓 Backtest period: 2020-01-01..2026-07-12
Spec ID: spec-multitimeframe-momentum-us-stocks-1783167355_v7 · Generated: 2026-07-10 08:19 UTC
Cluster: Momentum · Sub Cluster: Weekly-Daily Trend Alignment
Long-only US large-cap momentum strategy that enters stocks only when both daily and completed-week trend filters are bullish. Positions are ATR risk-sized, capped per name, and exited on trend failure, trailing stop, or maximum holding period.
Trend persistence can arise from slow information diffusion, institutional rebalancing, benchmark flows, and behavioral underreaction. Requiring both daily and weekly trend confirmation attempts to filter short-lived rallies and participate only when momentum is aligned across horizons.
The structure is deliberately simple: it favors liquid large-cap stocks, uses price-only signals, and relies on volatility-adjusted sizing so that higher-volatility names receive smaller allocations. The trade exists if multi-horizon trend confirmation improves signal quality enough to offset delayed entries, whipsaws, and transaction costs.
[code omitted from public view]
| Param | Value | Notes |
|---|---|---|
| Rule type | MULTI_TIMEFRAME_TREND_MOMENTUM |
Long-only multi-timeframe trend-following rule |
| Universe | Top 20 US stocks by capitalization | Provided symbol set includes AAPL, AMZN, BRK-A, BRK-B, GOOG, GOOGL, JNJ, JPM, MSFT, NVDA, TSLA, V, WMT, XOM and others |
| Daily trend | 20-day SMA above 50-day SMA; close above 20-day SMA; 20-day SMA slope over 5 days positive | Daily confirmation layer |
| Weekly trend | 20-week SMA above 50-week SMA; weekly close above 20-week SMA; 20-week SMA slope over 4 weeks positive | Uses completed Friday weeks only |
| Entry liquidity filter | 20-day average dollar volume >= $10,000,000 | Avoids illiquid entries |
| Rebalance | Daily at close | Signals and executions evaluated on daily bars |
| Position sizing | ATR volatility-adjusted | 14-day ATR; target 1% risk per position; max 10% portfolio per stock |
| Exit rules | Trend failure, 3.0x ATR trailing stop, or 126-day max hold | Exits if either daily or weekly trend turns non-bullish |
| Max leverage | 4.0x | Portfolio-level leverage cap |
| Backtest window | 2020-01-01 to 2026-07-12 | As specified in the backtest plan |
| Available data period | 2018-01-01 to 2023-01-01 | Indicator warmup/data availability context |
| Costs | $0.004/share commission; $1 minimum/order; 0.01% max commission; 0 bps slippage | Slippage assumption is optimistic |
| # | Concern | Status |
|---|---|---|
| 1 | Weekly signal timing | Uses completed weekly bars only, reducing same-week look-ahead risk |
| 2 | Daily close execution | Signals are evaluated at the close; execution at the same close may be optimistic unless modeled as market-on-close availability |
| 3 | Universe construction | Top-cap universe may embed survivorship or point-in-time membership bias if not reconstructed historically |
| 4 | Indicator warmup | SMA, slope, ATR, and volume filters require sufficient prior data; missing inputs are skipped |
| 5 | Corporate actions | Results depend on split/dividend-adjusted prices and consistent volume-dollar calculations |
| 6 | Costs and liquidity | Commissions included, but 0 bps slippage likely understates implementation friction |
The backtest produced a 64.88% total return with a Sharpe ratio of 0.80, Sortino of 0.92, and maximum drawdown of -13.57%. Profit factor was 1.55 across 1,330 trades, with a relatively low 36.09% win rate offset by larger average wins than losses. Overall, the profile is consistent with a trend-following approach: moderate risk-adjusted returns, meaningful turnover, and dependence on capturing extended upside trends while limiting losses during failed breakouts.
| Metric | Value |
|---|---|
| Total Return | 64.88% |
| Sharpe | 0.80 |
| Sortino | 0.92 |
| Calmar | 0.59 |
| Max Drawdown | -13.57% |
| Volatility | 12.43% |
| Win Rate | 36.09% |
| Profit Factor | 1.55 |
| Total Trades | 1330 |
| Symbols | 20 (AAPL, ACON, AMZN, BETR, BRK-A, BRK-B, GOODO, GOOG, GOOGL, HOLO, +10 more) |
Backtests are historical simulations for research purposes only. They are not investment advice and do not guarantee future performance.