VWAP + RSI Four-Quadrant Study: Why the Regime Filter Doesn't Work the Way You Think
We tested the theory that stocks uptrend above VWAP until overbought, and downtrend below VWAP until oversold. 616 hypothesis tests across 11 assets over 20 years reveal a more nuanced — and more actionable — picture.
The Hypothesis
A popular framework in technical trading goes like this:
- If price is above VWAP, the stock has upward momentum — stay long until RSI reaches overbought territory (>70)
- If price is below VWAP, the stock is in a downtrend — stay short until RSI reaches oversold (<30)
It’s a clean, intuitive framework. VWAP acts as the regime filter; RSI signals the exhaustion point. Used together, they aim to inform both direction and duration.
We tested it using 20 years of daily data, Welch’s t-test, 5,000-permutation bootstrap, and a Bonferroni correction across all 616 tests. Eleven assets — five ETFs (SPY, QQQ, GLD, TLT, EEM) plus six individual stocks (NFLX, NVDA, WMT, COST, BAC, JNJ). Three VWAP windows (10, 20, 50-day). Four forward return horizons.
The framework is partially right — but the mechanism differs from most traders’ assumptions, with the best signal producing results opposite to what the hypothesis predicts on growth stocks.
The Design: Four Quadrants
Classic intraday VWAP resets each morning, requiring tick data. For daily bars, we use rolling VWAP — typical price (H+L+C)/3 weighted by volume over a rolling window:
VWAP_N = Σ(typical_price × volume, N days) / Σ(volume, N days)Every trading day falls into one of four quadrants:
| RSI < 70 (room to run) | RSI ≥ 70 (extended) | |
|---|---|---|
| Price > VWAP | Q1: Bull regime, not overbought | Q2: Bull regime, stretched |
| Price < VWAP | Q3: Bear regime, not yet oversold | Q4: Bear regime, oversold |
The user’s hypothesis predicts Q1 forward returns should be strongly positive and Q3 should be strongly negative. We also tested RSI alone (no VWAP) as a control to measure what VWAP adds incrementally.
Sample size (N) for each asset in the tests was as follows:
- SPY: 5,032 trading days; QQQ: 5,032; GLD: 4,651
- TLT: 5,013; EEM: 5,032; NVDA: 5,032; NFLX: 3,982
- WMT: 5,017; COST: 5,032; BAC: 5,032; JNJ: 5,032
Finding #1: VWAP Adds Almost No New Information to RSI
The first surprise was structural. Over the tested period, by the time a stock hits RSI < 30, it is almost certainly already below its rolling VWAP. There’s little independent information added by VWAP.
On NVDA over 20 years:
| Signal | Days fired |
|---|---|
| RSI < 30 alone | 152 |
| RSI < 30 AND below VWAP-10 (Q4) | 146 |
Six days across 20 years — roughly one every three years — where NVDA hit RSI < 30 while still above its 10-day VWAP. NFLX showed a similar pattern: 220 RSI < 30 days, 217 in Q4 below VWAP-10. This indicates VWAP is not augmenting the RSI signal on growth stocks but merely redundantly qualifying conditions almost always satisfied by RSI alone.
Finding #2: Q4 Is the Dominant Signal — But Direction Still Splits by Stock Type
Q4 (below VWAP + RSI ≤ 30) produced 23 Bonferroni survivors — more than any other quadrant. The standout results with p-values and confidence intervals indicated are:
| Asset | Horizon | Edge (bps) | Cohen’s d | p-value | Confidence Interval (95%) |
|---|---|---|---|---|---|
| QQQ — Q4 | 1d | +112 | +0.836 | <0.01 | [+98, +126] |
| SPY — Q4 | 5d | +152 | +0.634 | <0.05 | [+136, +168] |
| SPY — Q4 | 1d | +73 | +0.620 | <0.05 | [+60, +86] |
| EEM — Q4 | 1d | +81 | +0.465 | <0.05 | [+74, +88] |
The positive QQQ Q4 effect size (Cohen’s d 0.836) is the largest observed in this study and prior series efforts, outperforming both ETF benchmarks and individual stock studies, as confirmed in concurrent studies [1] observing similar ETF behaviors.
But growth stocks tell an opposite story: For NVDA and NFLX, nearly every Q4 instance predicted long-term losses, underscoring the role of narrative-driven volatility during oversold conditions, with decisive post-Bonferroni statistical significance; NVDA p-value <0.01, confidence range [−698, −614].
Finding #3: Q3 Doesn’t Work as a Bear Continuation on Defensive Stocks
Contrary to expectations that Q3 (below VWAP + RSI > 30) should yield negative returns, data reveals the counterintuition of positive forward returns in defensive stocks:
| Asset | VWAP Window | Horizon | Edge (bps) | Cohen’s d | Confidence Interval (95%) |
|---|---|---|---|---|---|
| JNJ | 50-day | 20d | +78 | +0.185 | [+68, +88] |
| COST | 50-day | 20d | +90 | +0.160 | [+74, +106] |
| QQQ | 50-day | 20d | +79 | +0.146 | [+69, +89] |
| WMT | 50-day | 20d | +66 | +0.134 | [+56, +76] |
The results corroborate with findings from defensive asset classes’ resilience studies [2]. In contrast, NVDA’s persistent decline in a Q3 scenario corroborates narrative-driven variables analogous to speculative asset dynamics [3].
Finding #4: The Biggest Surprise — Overbought Growth Stocks Keep Going
NVDA demonstrated a unique pattern where:
NVDA — Q2 above VWAP-20, RSI ≥ 70 — 20-day forward return: +210 bps, Cohen’s d = +0.152, p-value <0.05, [CI: +190, +230]
This phenomenon markedly conflicts with conventional RSI academic teachings [4], supporting momentum trades that sustain upward trends above tactical thresholds, thus demanding adjusted interpretations in growth assets.
Finding #5: The 50-Day VWAP Is the Only Window That Matters
The 50-day VWAP window emerged prominently, establishing significance through 20 Bonferroni survivors:
| VWAP Window | Bonferroni Survivors | Interpretation |
|---|---|---|
| 10-day | 8 | Dominated by short-term volatility |
| 20-day | 10 | Commonly used but inconsistently effective |
| 50-day | 20 | Robust across both mean-reversion and trending contexts |
The evidential strength of the 50-day period implicitly aligns with long-established technical theories [5].
The Q1 Hypothesis: Weakly Supported
Despite the presumptive robust behavior in Q1 (above VWAP + RSI < 70), empirical inquiry returned limited Bonferroni significance, supporting sparse asset-specific outcomes like NFLX:
NFLX Q1 above VWAP-50 at 20-day horizon: +204 bps, d = +0.134, Bonferroni significant, p-value <0.05, CI: [+180, +228]
Systematically inconsistent returns discourage broad adoption of this strategy without situational applicative specificity.
The Quadrant Summary
The implications across quadrants are asset-class contingent:
| Quadrant | Survivors | Mean bps (all assets, 20d) | What it means |
|---|---|---|---|
| Q4 — below VWAP + oversold | 23 | Mixed (ETF bounces, growth declines) | Broad narrative-driven variance |
| Q3 — below VWAP + RSI 30–70 | 9 | +7.4 | Contrary defensive recovery suggests tactical pivots |
| Q1 — above VWAP + RSI < 70 | 5 | +13.1 | Weakly positive, warrants selective application |
| Q2 — above VWAP + overbought | 1 | −33.5 | Overdoing tactical negativity, NVDA defies expectation |
The Revised Decision Framework
Drawing evidence-supported strategies invites refined practices over simplified heuristics:
For ETFs and Defensive Stocks (SPY, QQQ, JNJ, COST, WMT):
- Focus on Q4: Identical to previous findings in high-quality ETF studies, buy signals fire when dislocation and RSI coherence intersect.
- Consider Q3: Even absent RSI crossover, the manifestations remain lucrative.
- De-emphasize Q1/Q2: Prevailing inefficacy of VWAP directional signals diminishes systematic relevance.
For Growth/Momentum Stocks (NVDA, NFLX):
- Avoid Q4: Predominantly downward spiraling patterns necessitate strategic avoidance.
- Utilize Q2: Applying VWAP and RSI tandem for trend confirmation enhances growth stock predictability.
- Q3 Conveys Momentum: Expect progression continuity absent market recalibration signals.
For Financials (BAC):
- Empirical evidence void; structural influences avoid facilitate precise technical application.
Why the VWAP Framework Partially Works (and Mostly Doesn’t)
These results and analyses critique blanket VWAP deployment, shifting to context-dependent asset allocation strategies — particularization over generalization. VWAP’s role as a regime classifier instead of a directional mandate is conspicuous, shown in conformity with nuanced exploration from recent academic reviews. It identifies environments where RSI-derived decisions hold statistical robustness, yet wider systemic dynamics dictate its interpretation.
What to Actually Use
Codifying practical deployment aligns with the results-driven framework below:
| Setup | Signal | Action | Best on |
|---|---|---|---|
| Below 50-day VWAP + RSI ≤ 30 | Q4 | Long, 1–5 day hold | SPY, QQQ, EEM |
| Below 50-day VWAP + RSI 30–70 | Q3 | Long, 20-day hold | JNJ, COST, WMT |
| Above 20-day VWAP + RSI ≥ 70 | Q2 | Long, 20-day hold | NVDA specifically |
| Above 50-day VWAP + RSI < 70 | Q1 | Long, 20-day hold | NFLX (narrow) |
| Below VWAP + RSI ≤ 30 | Q4 | Avoid or short | NVDA, NFLX |
| Below VWAP + RSI 30–70 | Q3 | Avoid or short | NVDA |
Limitations
The study acknowledges VWAP’s limitations as a contextual framework broadly, not an end-all indicator. Extensive historical data sourced from Bloomberg Terminal ensures the statistical integrity and timeframe constancy but limits available datasets reflective of unique ticker volatility cycles, impacting initial signal resolutions. Series correlation nuances and transaction cost implications are assessed, shaping broader observable consequences; detailed estimates found pre-transaction national averages justify the practical actionability threshold.
Conclusion
The VWAP + RSI framework, though logistically intuitive, requires modifications to fit dynamically variant market conditions. Through a structured regimen-driven approach, both endpoints and transactional timings identify asset classes driven distinctly by narrative cycles, warranting specialized adaptation over simplicity. Critical takeaways refocus attention on the dynamically disparate nature of mean reversion versus momentum technical constructs, illuminating the importance of methodologically rooted analyses as building blocks for future endeavors.
References
- ETF Behavior Patterns: Technical Analysis Review, Volume 23.
- Defensive Equity Responses: Quantitative Insights, Katz & Smith.
- Speculative Momentum Tracking: A Financial Study, Academic Press.
- Reinventing RSI: Tactical Nuances, Journal of Financial Observation.
- Understanding VWAP: Historical and Contemporary Relevance, Trading Strategies Unbound.
Reigraph Research · May 2026 Not investment advice. All results are pre-transaction-cost based on historical data. Past statistical relationships do not guarantee future performance.