Defensive Trend Pullback
Looks for RSI rebound behavior inside a longer-term SMA200 uptrend, prioritizing controlled mean reversion over aggressive entry.
Methodology / Assumptions
- Signals are produced from completed daily bars; entries fill at the next open with modeled costs.
- Headline view defaults to the strongest symbol for this strategy, and the selection-bias notice documents that assumption.
- No intraday fills, live order execution, shorts, options, or margin are modeled.
- AI memo prose is generated from the deterministic backtest payload or a template fallback; metrics are engine-derived.
Defensive Trend Pullback
mean reversion thesis: treat AAPL as research evidence only when factor evidence, trend structure, and risk-adjusted backtest behavior align. The current model reads the setup as medium-confidence research evidence, not a live trading instruction.
The ranking model weighs annualized return, Sharpe, drawdown containment, trade sample size, data quality, and cost-aware execution. This strategy is strongest when its signal frequency is sufficient and drawdown remains inside the radar threshold.
Deterministic template memo. Set DEEPSEEK_API_KEY to enable LLM-written narrative.
Benchmark and excess are measured against a passive hold of AAPL over the same window. A long-only, risk-managed rule that trades infrequently is expected to trail a strong benchmark tape — the goal here is risk-adjusted behavior (Sharpe, drawdown), not beating the index.
Deterministic synthesis from the backtest, stress diagnostics, and live regime.
Adequate evidence, contained downside, and a stress-adjusted score that holds up for simulation review.
Behavioral tilt from the strategy rules and realized statistics.
Last ~63 sessions · strategy return split into market vs selection.
Suggested next research test. Stress-test position sizing and slippage, then add to the paper-observation queue.
SMA200 uptrend with RSI pullback rebound; signal date 2026-06-15; executed next open
Defensive Trend Pullback on AAPL produced a adjusted real-data backtest annualized return of +2.7%, Sharpe 1.17, and max drawdown -2.1%.
The return and drawdown profile is relatively balanced, suggesting the rule captures trend windows while retaining some defensive behavior in weaker tape.
No major risk flag is currently triggered, but continued validation is still required.
Keep it on the radar and expand the evidence set across more symbols and market regimes.
Stable under a risk-on tape
Suitability for paper observation during selloff conditions: suitable.
Stress status tightens automatically when the broad regime is risk-off. Drawdown metrics are deterministic from the backtest equity curve; historical backtests do not represent future returns. Research only — not investment advice.
Calendar split at 2025-08-15 (70% in-sample). Backtest is unchanged — the engine simply re-derives metrics for each window from the same equity curve. Numbers are not refit; this is an honesty check.
- Annualized
- +1.3%
- vs benchmark
- -46.6%
- Sharpe
- 0.71
- Max drawdown
- -1.5%
- Trades
- 4
- Win rate
- 100.0%
- Annualized
- +6.0%
- vs benchmark
- -13.3%
- Sharpe
- 1.91
- Max drawdown
- -2.1%
- Trades
- 5
- Win rate
- 60.0%
Strategy daily return regressed on [Market, Momentum, Low-vol]. Alpha is what's left after these known factors are accounted for — the part the simple factors cannot explain.
A high R² with small alpha means the factors explain most of the return — be skeptical of “edge”. Significant alpha (|t| ≥ 2) with low R² is more interesting evidence that something beyond Market/Momentum/Low-vol is driving the strategy.
AI Research Memo: Market Selloff Review
Constructive tape: 21/28 names hold their 200-day trend with contained volatility (30% annualized). Risk signals are quiet, but drawdown discipline still applies. Composite stress score is 7/100 (risk-on). Breadth is broad and momentum is leading; short-horizon volatility is running 6% above its 60-day baseline. Conditions support continuation research, but keep stops and position sizing documented in case the regime turns.
Breadth is broad (21/28 above SMA200), momentum is leading at 2.7% avg 20-day, and volatility is elevated (30% annualized). Factor signals are not flashing broad stress, but drawdown discipline still governs observation admission.
3 of 5 screened strategies read as resilient and 0 as under stress. Strategies with positive historical returns but high downside sensitivity should be moved from "candidate" to "watch" until volatility normalizes.
Stress-adjusted scoring repriced 0 of 5 strategies versus their base score, penalizing high drawdown and downside volatility while rewarding benchmark-relative resilience and smoother equity. 2 strategies still clear the radar-candidate gate.
1 simulated observation remain live. Observation continues under stress so the desk can study how admitted rules behave through the drawdown — no orders are routed.
- Re-run the radar shortlist with the stress-adjusted score as the primary sort key and compare observation admissions.
- Measure each candidate's worst 5-day return and benchmark-relative drawdown against the current regime.
- Stress-test stop placement against overnight gaps rather than intraday fills.
- Track whether breadth broadens or volatility keeps expanding before changing candidate priorities.
Generated from deterministic engine metrics (regime, breadth, volatility, drawdown, radar scoring). No order instructions. Historical backtests do not represent future returns. Research only — not investment advice.
- Generate signals only from completed daily bars.
- Fill entries at the next bar open with modeled slippage and fees.
- Exit via stop loss, trailing stop, holding-period expiry, or the strategy's predefined exit rule.
- No major model risk flag is triggered, but validation is still required.
- Expand the universe beyond the current sector-diversified watchlist and rerun the same rule set.
- Test a tighter slippage and fee sensitivity range to understand implementation drag.
- Compare 20%, 10%, and volatility-scaled position sizing across the same signals.
Entry signals are generated from completed bars and filled at the next bar open. Stop and trailing-stop exits are evaluated on marked daily closes.
| Entry | Exit | Hold | Return | P/L | Fees | Reason |
|---|---|---|---|---|---|---|
| 2024-08-01 | 2024-08-26 | 17d | +1.1% | $212 | $2 | SMA200 uptrend with RSI pullback rebound |
| 2024-09-12 | 2024-10-01 | 13d | +3.0% | $605 | $2 | SMA200 uptrend with RSI pullback rebound |
| 2024-11-06 | 2024-11-26 | 14d | +4.8% | $969 | $2 | SMA200 uptrend with RSI pullback rebound |
| 2025-01-16 | 2025-02-25 | 26d | +4.5% | $911 | $2 | SMA200 uptrend with RSI pullback rebound |
| 2025-10-14 | 2025-10-31 | 13d | +12.2% | $2,492 | $2 | SMA200 uptrend with RSI pullback rebound |
| 2025-12-24 | 2026-01-20 | 16d | -9.5% | -$1,993 | $2 | stop loss |
| 2026-01-22 | 2026-02-06 | 11d | +11.1% | $2,262 | $2 | SMA200 uptrend with RSI pullback rebound |
| 2026-03-20 | 2026-03-31 | 7d | -0.1% | -$29 | $2 | SMA200 uptrend with RSI pullback rebound |
| 2026-04-01 | 2026-04-20 | 12d | +6.3% | $1,324 | $2 | SMA200 uptrend with RSI pullback rebound |