Buy when price crosses above the 200-day SMA, exit when it crosses below. The canonical long-term trend filter applied to gold — a market known for strong secular trends.
Strategy Logic & Backtest Setup
Buy when price crosses above SMA(200). Sell when price crosses below SMA(200). 5% stop loss.
Robustness Score: 62/100. Moderate robustness with notable sensitivities. Typical return +0.0%, worst-case -9.1% (5th percentile).
Model results from 304 simulation runs across curated historical market phases and synthetic stress tests, aggregated by market regime.
Values are model estimates: average and range (±1 standard deviation) from Monte Carlo variations across historical market phases and synthetic stress tests. Not a prediction of future performance, not investment advice.
Stocks & indices climb over weeks or months — like Bull Run 2017 or Tech Rally 2021.
Market drifts directionless inside a range — like SPY 2015 or Range 2011–2012.
Low volatility, muted price action — like mid-2017 or pre-Lehman 2007.
Large daily swings, vol spikes — like the February 2018 vol shock.
Markets decline over an extended period — like Dotcom Bust 2001 or 2022 Bear Market.
Sudden sharp drawdowns, liquidity stress — like Lehman 2008 or COVID Crash 2020.
Aggregation note: Cases can contribute to multiple regimes — a crash counts as both "Market Crash" and "High Volatility", for example. Range combines within-subtype and between-subtype spread. Total 558 case contributions across 9 failure modes.
Your strategy returned -3.1% on average across sideways markets with false signals. Require signal confirmation (e.g. close above the level for N bars) to filter noise.
Model-based scenario simulation. All values are produced by computational market models — curated historical market phases and synthetic stress tests with Monte Carlo variations. They describe how the strategy behaves in the modelled scenarios, not future performance in live markets.
Not investment advice, not financial analysis under § 34b WpHG (German Securities Trading Act), not a recommendation to buy or sell. Past or simulated performance is not a reliable indicator of future results.
Strategy performance across curated market episodes — real historical periods plus synthetic stress scenarios. Each case is chosen to test a distinct failure mode of trading strategies.