We pulled the common intraday playbook into one test: opening range breakouts, VWAP crosses and rejections, EMA pullbacks, gap-and-go, gap fades, prior-day level breaks, momentum continuation, lunch range breaks, and liquidity sweeps. Then we converted each setup into deterministic 5-minute rules and ran them across four liquid symbols: SPY, QQQ, NVDA, and TSLA.
The data came from Alpaca historical bars. The test window was January 2, 2026 through June 5, 2026. The result was 5,604 simulated trades across 28 strategy variants. This is not a production trading system; it is a first-pass way to separate setups that deserve more refinement from setups that mostly sound good in isolation.
What worked best
The strongest group was not generic VWAP trading. It was key-level continuation: prior-day high breakouts, prior-day low breakdowns, 30-minute opening range breakouts, and long EMA pullbacks above VWAP. Those setups had positive average R and enough sample size to be worth another pass.
| Strategy | Trades | Win Rate | Avg R | Total R |
|---|---|---|---|---|
| Prior day high breakout | 153 | 47.7% | 0.20 | 30.16 |
| Intraday momentum breakdown short | 130 | 39.2% | 0.15 | 20.13 |
| Gap fade long | 46 | 45.7% | 0.14 | 6.50 |
| 30m ORB breakout | 286 | 49.3% | 0.13 | 37.92 |
| Prior day low breakdown | 128 | 41.4% | 0.13 | 16.89 |
| EMA pullback long | 354 | 41.2% | 0.12 | 41.90 |
| 15m ORB breakout | 259 | 42.9% | 0.06 | 14.98 |
| VWAP cross short | 354 | 37.0% | 0.04 | 15.07 |
| VWAP cross long | 360 | 35.8% | 0.03 | 9.12 |
| Liquidity sweep reclaim long | 139 | 33.8% | 0.02 | 3.06 |
| Opening drive long | 71 | 46.5% | 0.02 | 1.33 |
| EMA 9/21 cross long | 218 | 43.6% | 0.01 | 1.71 |
Recreated chart setups
Each setup below is one of the best historical examples from the backtest. The chart uses the same Alpaca 5-minute candles used by the test, with VWAP, EMA9, EMA21, entry, stop, target, and exit marked on the session. Use the selector to inspect the setup date before judging the signal quality.
TSLA 2026-06-03 - LONG prior-day level breakout
price broke prior session high. close reclaimed prior day high.
Entry Jun 3, 10:00 AM ET. Exit Jun 3, 10:50 AM ET via target.
The main read
Prior-day high breakouts led the table at 0.20 average R across 153 trades. Intraday momentum breakdown shorts came next at 0.15 average R, followed by gap fades after gap-down VWAP reclaims. The best opening-range variant was the 30-minute ORB breakout: 286 trades, 49.3% win rate, and 0.13 average R.
The surprise was how weak some popular setups were without additional filtering. VWAP cross long and VWAP cross short were only slightly positive. VWAP rejection long, lunch range breakout long, EMA pullback short, and gap-and-go short were negative in this sample. That does not mean those setups never work; it means the naive version is not enough.
What to refine next
- Add market regime filters. ORB and prior-day breaks likely need different treatment on high-volatility trend days versus compressed range days.
- Separate longs and shorts by symbol. SPY/QQQ index behavior is not the same as NVDA/TSLA single-name behavior.
- Improve VWAP rejection context. The simple trend-side wick test underperformed; it probably needs day-type, distance-from-open, and higher-timeframe filters.
- Retest exits. This pass used fixed R targets and structural stops. Partial exits and time-based exits may change the ranking.
The usable conclusion is narrow but useful: start with key-level breakouts and long trend pullbacks, then add filters. Do not assume that a named intraday setup has an edge just because it is popular.