How Often Do Gaps Start Multi-Day Trends?
Gaps are some of the most visible and psychologically powerful events on a price chart. When a market opens above the previous day’s high, it feels like something big is about to happen—like a trend is beginning. How often do gaps start multi-day trends?
To answer this, we ran a simple but revealing backtest on four key markets—SPY (stocks), Bitcoin, TLT (bonds), and GLD (gold)—to measure what happens after a gap. We tested whether buying at the open and holding for 1 to 10 days leads to consistent gains.
Here’s what we found:
- Bitcoin was the only asset with strong and sustained follow-through after gap-ups.
- SPY showed minimal edge—gaps tend to get digested or faded quickly.
- TLT had virtually no trend after gaps—classic mean-reverting behavior.
- GLD showed small, steady gains that could appeal to swing traders.
If you’re a short-term trader or strategy developer, these findings help you decide which gaps to trust and which to fade. Below, we break down the exact rules, the data, and the insights that came out of the test.
Let’s dive in.
Relevant article: –Gap types
Why Study Gap-Ups in the First Place?
“Gap trading” is one of the oldest intraday edges studied by discretionary and systematic traders alike. When an asset opens above yesterday’s high (a gap-up), it signals overnight buying pressure. The natural question is:
Does that pressure turn into a sustained trend you can trade?
To find out, we ran a clean, data-driven backtest on four core asset classes that most portfolios already touch:
- SPY – U.S. equities
- Bitcoin – the dominant crypto asset (we backtested the cash version trading 24/7, thus the #trades are smaller than for the other three assets)
- TLT – long-dated U.S. Treasuries
- GLD – spot gold
How Often Do Gaps Start Multi-Day Trends? Trading Rules
We made the following trading rules:
| Step | Rule |
|---|---|
| 1 | Entry: Next open above the previous day’s high → buy at the open |
| 2 | Exit: Sell exactly N trading days later (we tested N = 1 … 10) |
| 3 | Metric: Average % gain per trade over the holding window |
| 4 | Universe: SPY, BTC, TLT, GLD |
| 5 | Bias note: Only gap-ups tested (gap-downs trend less in such upward-biased assets) |
The code is fewer than 30 lines in Python or Amibroker—perfect for quick research sprints.
Backtest Results at a Glance
| Sell After (N days) | SPY Avg % | BTC Avg % | TLT Avg % | GLD Avg % |
|---|---|---|---|---|
| 1 | 0.01 | 1.24 | 0.01 | 0.06 |
| 2 | 0.06 | 1.55 | 0.01 | 0.12 |
| 3 | 0.06 | 1.50 | 0.01 | 0.14 |
| 4 | 0.08 | 1.55 | 0.01 | 0.25 |
| 5 | 0.11 | 2.42 | –0.05 | 0.27 |
| 6 | 0.17 | 3.59 | 0.02 | 0.31 |
| 7 | 0.20 | 3.35 | 0.04 | 0.32 |
| 8 | 0.23 | 3.15 | 0.02 | 0.46 |
| 9 | 0.36 | 3.62 | 0.01 | 0.48 |
| 10 | 0.35 | 2.98 | 0.07 | 0.47 |
(Data source: in-house; period identical across assets to keep comparisons fair.)
Key Takeaways
1. SPY: An Efficient Market Eats the Edge
- Stat-sized price drift: 0.01 % on day 1, only 0.35 % after 10 days.
- Interpretation: Whatever bullish information created the gap is usually priced-in quickly. Size-adjusted, the edge barely beats commissions or slippage.
- Practical angle: Filtering gaps by macro regime, VIX level, or volume may be mandatory to carve a tradable edge.
2. Bitcoin: Volatility = Opportunity
- Outlier performer: 1.24 % on day 1, compounding to ~3 % by day 10.
- Why? Crypto gaps often reflect order-book imbalances across 24-hour exchanges. Follow-through is real.
- Caveat: Higher volatility means higher risk.
3. TLT: Mean-Reverting Bond Market
- Flat as a pancake: ±0.05 % on most horizons.
- Lesson: Bond gaps most often reverse or stall; fixed-income markets absorb news swiftly, and mean reversion dominates.
- Actionable? Fade the gap rather than follow it—but test first.
4. GLD: Slow but Steady Drip
- Small, consistent gains: 0.06 % (day 1) → 0.47 % (day 10).
- Read: Gold often trends gently after a gap-up, but the move is modest compared to daily volatility (~0.9 %).
- Worth it? Possibly for swing traders who can leverage low volatility, but the edge is thin after costs.
Why Bitcoin Behaves Differently
- 24/7 trading with fragmented liquidity → News-driven gaps are “fixed” more slowly.
- Higher baseline volatility → A one-day 1 % move is noise in BTC but a signal in SPY.
- Retail dominance → Crowded sentiment can propel follow-through longer.
Practical Implications for Traders
- Edge existence ≠ tradability. Consider position sizing, frictional costs, and drawdowns.
- Add filters. Time-of-week, relative gap size, volatility regime, or macro backdrop can sharpen edges.
- Diversify gaps. Combining equity and crypto gaps may smooth the equity curve because the correlation of returns is low.
Limitations & Next Research Steps
- Sample size—gap-ups are infrequent, especially in bonds. Pooling decades of data helps, but still leaves tail-risk issues.
- No risk adjustment—raw % gains ignore volatility and drawdowns. Calculating CAGR, Sharpe, and max DD should follow.
- Only long gaps tested—Short setups (gap-down reversals) might be compelling, especially in mean-reverting assets.
The Bottom Line
- Bitcoin gaps trend; SPY gaps usually don’t; TLT gaps almost never do; GLD sits in the middle.
- Raw numbers alone won’t pay the bills—traders need filters and risk management.
- Still, gap studies remain a fantastic sandbox for learning how overnight information becomes real price action.