TRoy Research: one trend rule, many markets
Our own research. The rule is simple: hold a market while its price is above its 200-day average, step aside when it falls below. One market on its own has long quiet stretches and sharp drops. Doing the same on many different markets at once spreads the risk: when one market falls, others are usually still rising. Here are the results, how we tested them, and a live version you can change yourself.
1. Test results (run offline, daily data, 2001 to 2026)
| Per year | Ups and downs | Sharpe | Worst drop | |
|---|---|---|---|---|
| Whole period: hold all | 8.3% | 15% | 0.62 | 43% |
| Whole period: trend rule on each | 5.4% | 8% | 0.72 | 12% |
| Whole period: S&P 500 ETF (SPY) | 8.7% | 19% | 0.54 | 56% |
| Last 40%: hold all | 11.3% | 15% | 0.79 | 30% |
| Last 40%: trend rule on each | 7.3% | 8% | 0.92 | 10% |
| Last 40%: S&P 500 ETF (SPY) | 12.0% | 19% | 0.71 | 34% |
| Per year | Ups and downs | Sharpe | Worst drop | |
|---|---|---|---|---|
| Whole period: hold all | 7.8% | 18% | 0.51 | 47% |
| Whole period: trend rule on each | 9.5% | 13% | 0.78 | 26% |
| Last 40%: hold all | 21.2% | 28% | 0.82 | 47% |
| Last 40%: trend rule on each | 19.3% | 19% | 1.01 | 26% |
What it shows: spreading the same rule across many markets cut the worst drop sharply (from 43% to 12% in the ETF test, from 47% to 26% in the crypto and currency test) and improved the risk-adjusted return (Sharpe) in both tests. In the ETF test it earned less per year than holding everything, in exchange for much smaller drops. In the crypto and currency test it earned as much or nearly as much as holding, with smaller drops.
2. One market at a time: popular indicators compared
| Method | Sharpe | Per year | Worst drop |
|---|---|---|---|
| Buy and hold | 0.57 | 13.8% | 45% |
| 200-day trend rule | 0.52 | 11.0% | 30% |
| Vote of nine trend signals | 0.50 | 8.0% | 25% |
| RSI(14) above 50 | 0.49 | 9.4% | 31% |
| Squeeze Momentum | 0.48 | 8.2% | 34% |
| 12-month momentum | 0.46 | 9.7% | 35% |
| MACD cross | 0.45 | 6.9% | 31% |
| Ichimoku cloud / Supertrend | 0.43 | 7.7% | 29% / 33% |
| Breakout (Donchian 55/20) | 0.36 | 5.8% | 28% |
| UT Bot | 0.25 | 4.4% | 40% |
| Bollinger Bands / RSI mean reversion | 0.20 to 0.23 | 0.8% to 2.2% | 31% to 36% |
The plain 200-day rule was the strongest single-market method we tried. Adding momentum filters, volatility sizing, stops or machine learning on top did not improve it.
3. Live: the trend portfolio on 14 markets, calculated in your browser
Loading...
Markets: Bitcoin, Ethereum, BNB, Solana, XRP, Cardano and a gold-backed token (long or cash), plus seven currency pairs (long or short). Equal weights among the markets that have enough history. A signal at the close is acted on from the next day. The live numbers differ a little from the offline table above because the calendar and the data window are not identical.
How we tested, and what to keep in mind
- Signals are read at each day's close and acted on the next day. Costs are included: 0.05% per side for ETFs, 0.10% for crypto, 0.02% for currencies.
- “Last 40%” rows cover only the final 40% of each history. The 200-day length is a standard choice; nearby lengths (150 to 250 days) gave similar results, and we compared it with many popular alternatives.
- Test periods include long bull markets, which favour holding. Different periods can give different results.
- Currencies use daily reference rates (one per weekday), not tradable quotes. Crypto candles are public market data. Share and ETF data is not shown live here.
- Past results do not promise future results. This is research information, not advice, and TRoyTRADE does not trade for customers.
Charts: Lightweight Charts, an open-source library by TradingView, Inc. (Apache 2.0).
