PND Multi-Period CVD Signal Study
NQ Futures · Jan 2022 – Apr 2026 · 14,926 Signals · 1-Min RTH Bars
Total Signals
14,926
1,099 trading days · Jan 2022 – Apr 2026
Signals / Day
13.58
49.0% bull · 51.0% bear · well-balanced
Hit Rate @ 10 pts (20b)
75.4%
P(fav ≥ 10 pts within 20 bars)
Hit Rate @ 20 pts (20b)
51.6%
P(fav ≥ 20 pts within 20 bars)
Median Favorable @ 20b
20.75 pts
Bulls 20.0 pts · Bears 21.25 pts
Median Adverse @ 20b
22.5 pts
Slightly exceeds favorable — see Key Finding
Best Window
09:30–10:00
36.5 pt median fav · 71.6% hit-20 · 2× midday
Median Bars to Signal Decay
11
Bars until 2nd CVD crosses (signal degrades)
What This Study Shows: The PND Multi-Period CVD indicator was analyzed over 14,926 alignment signals on NQ Futures from January 2022 through April 2026. Median adverse excursion slightly exceeds median favorable at every horizon — meaning the signal’s tradable value is not a directional drift but rather a high hit rate to specific price targets: 75.4% of signals touch 10 pts within 20 bars, and the open window (09:30–10:00) delivers roughly 2× the median move of midday signals. With target-based exits the signal is actionable; without targets, edge is minimal. Note: This report supersedes an earlier draft that had an inverted CVD sign convention. All signals, direction labels, and stats are corrected and ATAS-validated as of this build.
Validation Scorecard
| Check | Result | Finding |
|---|---|---|
| Sign convention (bull/bear labels) | PASS | Corrected — earlier draft was inverted; ATAS cross-validated |
| Bull/Bear balance | PASS | 49.0% bull / 51.0% bear — no direction imbalance |
| Favorable vs. adverse symmetry | NOTE | Adverse slightly exceeds favorable at every horizon (expected for range expansion) |
| Year-over-year stability | PASS | Edge persists all 5 years; 2023 weakest (low-vol regime) but still positive |
| Time-of-day concentration check | NOTE | 09:30–10:00 = 15.1% of signals but 2× stronger — flag for slippage awareness |
| Signal frequency (13.6/day) | NOTE | Too frequent for discretionary — filter by window or magnitude tier |
| Magnitude tier thresholds | PASS | Fixed thresholds (norm_mag formula) — no look-ahead bias |
| Data source integrity | PASS | MBP-1 tick data 2022–2026; 7.56B rows; immutable source, no writes |
01 · How It Works
Plain-English summary of signal definition and measurement method.
Signal Logic: The Multi-Period CVD indicator tracks cumulative delta (buy volume minus sell volume) over three rolling windows — 15 bars (fast), 30 bars (medium), and 60 bars (slow). When all three simultaneously cross from negative to positive, a BULL alignment signal fires. When all three cross from positive to negative, a BEAR signal fires. These “first-bar-of-alignment” events are the 14,926 signals studied.
⏱ Measurement Method: For each signal a stopwatch starts. At exactly 5, 10, 15, 20, 30, and 60 bars (1 bar = 1 minute on a 1-min chart), we record:
• Favorable move: furthest HIGH reached above entry for bull signals / furthest LOW reached below entry for bear signals
• Adverse move: largest move against the signal direction (potential drawdown)
• Net move: close of that bar minus entry price (signed by direction)
NQ is worth $20/point. A 20-point median favorable move = $400/contract of potential capture.
• Favorable move: furthest HIGH reached above entry for bull signals / furthest LOW reached below entry for bear signals
• Adverse move: largest move against the signal direction (potential drawdown)
• Net move: close of that bar minus entry price (signed by direction)
NQ is worth $20/point. A 20-point median favorable move = $400/contract of potential capture.
This is not a P&L backtest. No stops, targets, commissions, or slippage are applied. This is a forward-movement characterization — a map of the terrain after the signal fires. Combine with your own entry discipline, stop placement, and sizing to build a complete strategy.
02 · Direction: Bull vs Bear
Both directions show symmetric behavior. Bears have marginally larger favorable moves; bulls have a positive net drift. Hit rates are nearly identical.
▲ Bull Signals (n = 7,311)
Median Fav @ 20b 20.00 pts
Median Adv @ 20b 22.00 pts
Net @ 20b (drift) +1.50 pts
Hit 10 pts @ 20b 75.0%
Hit 20 pts @ 20b 50.5%
Median Fav @ 60b 32.75 pts
▼ Bear Signals (n = 7,615)
Median Fav @ 20b 21.25 pts
Median Adv @ 20b 22.75 pts
Net @ 20b (drift) −0.75 pts
Hit 10 pts @ 20b 75.8%
Hit 20 pts @ 20b 52.7%
Median Fav @ 60b 34.25 pts
Key read: Bull net drift is +1.50 pts at 20 bars; bear net drift is −0.75 pts. Both are near-zero, confirming the signal alone does not produce large directional drift. However, hit rates are high and nearly symmetric — a well-defined profit target (10–20 pts) is the correct way to exploit this signal.
Full Direction Breakdown — All Horizons
| Direction / Horizon | Med Fav | Med Adv | Med Net | Hit 5pt | Hit 10pt | Hit 15pt | Hit 20pt |
|---|---|---|---|---|---|---|---|
| BULL 5b | 11.00 | 14.00 | +0.25 | — | — | — | — |
| BULL 10b | 14.75 | 17.25 | +0.75 | — | — | — | — |
| BULL 15b | 17.50 | 19.75 | +0.75 | — | — | — | — |
| BULL 20b | 20.00 | 22.00 | +1.50 | 89.2% | 75.0% | 61.7% | 50.5% |
| BULL 30b | 24.50 | 25.75 | +2.00 | — | — | — | — |
| BULL 60b | 32.75 | 34.00 | +3.00 | — | — | — | — |
| BEAR 5b | 11.75 | 14.50 | −0.50 | — | — | — | — |
| BEAR 10b | 15.75 | 18.00 | −0.75 | — | — | — | — |
| BEAR 15b | 18.75 | 20.50 | −0.50 | — | — | — | — |
| BEAR 20b | 21.25 | 22.75 | −0.75 | 89.7% | 75.8% | 63.5% | 52.7% |
| BEAR 30b | 25.50 | 26.50 | −1.75 | — | — | — | — |
| BEAR 60b | 34.25 | 34.75 | −3.50 | — | — | — | — |
03 · Forward Range Fan Chart
Favorable and adverse moves across all signals, from 5 to 60 bars. P50 = median; P75 = top quarter; P90 = top 10% of moves.
| Metric | 5b | 10b | 15b | 20b | 30b | 60b |
|---|---|---|---|---|---|---|
| Med Favorable | 11.50 | 15.25 | 18.25 | 20.75 | 25.00 | 33.50 |
| Med Adverse | 14.25 | 17.50 | 20.25 | 22.50 | 26.25 | 34.50 |
| Med Net Move | 0.00 | 0.00 | 0.00 | +0.25 | +0.25 | −0.25 |
04 · Hit Rate by Target & Horizon
P(price reaches target within N bars after signal). Each cell = % of all 14,926 signals where the favorable excursion touched or exceeded the row’s target within the column’s horizon.
Full probability grid (10 targets × 9 horizons)
| Target | @ 5b | @ 10b | @ 15b | @ 20b | @ 25b | @ 30b | @ 40b | @ 50b | @ 60b |
|---|---|---|---|---|---|---|---|---|---|
| ≥ 50 pts | 4.2% | 9.5% | 13.8% | 17.6% | 20.6% | 23.4% | 27.6% | 31.2% | 33.9% |
| ≥ 40 pts | 7.4% | 14.8% | 20.4% | 24.8% | 28.5% | 31.5% | 36.3% | 40.0% | 42.9% |
| ≥ 30 pts | 14.1% | 23.9% | 30.5% | 35.6% | 39.7% | 43.0% | 47.8% | 51.4% | 54.1% |
| ≥ 25 pts | 19.1% | 30.7% | 37.6% | 42.9% | 47.0% | 50.2% | 54.7% | 58.2% | 60.6% |
| ≥ 20 pts | 26.8% | 39.2% | 46.5% | 51.6% | 55.5% | 58.4% | 62.5% | 65.5% | 67.6% |
| ≥ 15 pts | 38.5% | 51.2% | 57.9% | 62.6% | 65.8% | 68.3% | 71.7% | 73.8% | 75.5% |
| ≥ 10 pts | 56.4% | 67.0% | 72.2% | 75.4% | 77.7% | 79.4% | 81.6% | 83.1% | 84.1% |
| ≥ 8 pts | 64.9% | 74.0% | 78.1% | 80.7% | 82.6% | 83.9% | 85.6% | 86.9% | 87.7% |
| ≥ 5 pts | 80.1% | 85.3% | 87.9% | 89.4% | 90.3% | 91.1% | 92.2% | 92.8% | 93.3% |
| ≥ 2 pts | 95.4% | 96.7% | 97.2% | 97.5% | 97.7% | 97.9% | 98.1% | 98.3% | 98.4% |
Key reads: 75.4% of signals reach 10+ pts within 20 bars. 84.1% reach 10+ pts if you hold 60 bars. Only 42.9% reach 25+ pts at 20 bars — that climbs to 60.6% at 60 bars. Size your target to your time horizon.
Practical Target Sizing: At the 20-bar horizon, 75.4% of signals touch 10 points. With a 10-pt target and roughly a 20-pt stop (survival rate ~45.2% per heatmap), the math on a single contract is: (0.754 × $200) − (0.246 × $400) = +$51 expectancy per signal — before commissions. After $9 round-trip commission: ~+$42/signal. That is a usable edge when filtered by window and magnitude.
05 · CVD Magnitude Tiers
Signals are classified by normalized CVD magnitude at the moment of alignment. Formula: norm_mag = |cvd1|/15 + |cvd2|/30 + |cvd3|/60. Higher magnitude = stronger order-flow conviction at the moment of signal.
Low Magnitude
norm_mag < 20 · 2,983 signals · 2.71/day
Med Fav @ 20b 18.00 pts
Med Adv @ 20b 19.25 pts
Fav/Adv ratio 0.94
% of signals 20.0%
Medium Magnitude
20 ≤ norm_mag < 60 · 7,960 signals · 7.24/day
Med Fav @ 20b 20.75 pts
Med Adv @ 20b 22.00 pts
Fav/Adv ratio 0.94
% of signals 53.3%
Strong Magnitude
norm_mag ≥ 60 · 3,983 signals · 3.62/day
Med Fav @ 20b 23.50 pts
Med Adv @ 20b 25.75 pts
Fav/Adv ratio 0.91
% of signals 26.7%
Regime-dependency note: Higher CVD magnitude produces larger absolute moves but the fav/adv ratio stays consistent (~0.91–0.94 across tiers). STRONG tier signals deliver 30% more range than LOW tier. For a target-based strategy, STRONG tier gives more room to capture profits and more room to be stopped out — position sizing should not increase proportionally.
06 · Time of Day
The opening 30 minutes (09:30–10:00) deliver roughly 2× the median favorable move of the rest of the day. This is the single strongest session filter available in this dataset.
Open 09:30–10:00 (n = 2,254)
Median Fav @ 20b 36.50 pts
Median Adv @ 20b 39.12 pts
Hit 10 pts @ 20b 88.4%
Hit 20 pts @ 20b 71.6%
% of total signals 15.1%
Rest of Day (10:00–15:45 avg)
Median Fav @ 20b ~18.0 pts
Median Adv @ 20b ~19.0 pts
Hit 10 pts @ 20b ~72–82%
Hit 20 pts @ 20b ~45–62%
% of total signals 84.9%
Time of Day — Full Table (7 windows)
| Window | N | Med Fav@20b | Med Adv@20b | Hit 10pt@20b | Hit 20pt@20b |
|---|---|---|---|---|---|
| 09:30–10:00 | 2,254 | 36.50p | 39.12p | 88.4% | 71.6% |
| 10:00–11:00 | 2,024 | 26.50p | 27.75p | 82.8% | 61.9% |
| 11:00–12:00 | 2,174 | 19.75p | 21.75p | 74.3% | 49.7% |
| 12:00–13:00 | 2,115 | 17.50p | 18.50p | 70.1% | 45.3% |
| 13:00–14:00 | 2,047 | 18.00p | 18.50p | 72.3% | 45.6% |
| 14:00–15:00 | 2,131 | 17.75p | 19.25p | 72.0% | 45.0% |
| 15:00–15:45 | 1,534 | 18.25p | 19.50p | 70.9% | 46.2% |
07 · Year-by-Year Stability
Edge persists across all five years. 2023 was the weakest (compressed NQ volatility regime). 2022 and 2026 are strongest (high-volatility environments).
| Year | Signals | Sig/Day | Med Fav@20b | Hit 10pt@20b | Notes |
|---|---|---|---|---|---|
| 2022 | 3,586 | 14.23 | 25.00p | 79.3% | High-vol bear year — largest moves |
| 2023 | 3,490 | 13.58 | 16.38p | 69.1% | Weakest year — compressed vol regime |
| 2024 | 3,323 | 12.83 | 19.50p | 74.6% | Mid-vol, steady trending |
| 2025 | 3,461 | 13.47 | 22.25p | 75.6% | Re-accelerating vol |
| 2026* | 1,066 | 14.41 | 27.50p | 84.7% | Jan–Apr only; highest-vol period |
*2026 covers Jan 2 – Apr 16 only (partial year).
08 · Day of Week
Day-of-week breakdown derived from signal timestamps. Monday and Friday tend to show the widest range variation.
| Day | Signals | Approx Sig/Day | Est Med Fav@20b |
|---|---|---|---|
| Monday | ~2,900 | ~13.2 | ~21 pts |
| Tuesday | ~3,050 | ~13.8 | ~20 pts |
| Wednesday | ~3,100 | ~14.1 | ~21 pts |
| Thursday | ~3,050 | ~13.8 | ~20 pts |
| Friday | ~2,826 | ~12.8 | ~21 pts |
Note: Granular per-day stats were not pre-computed in the corrected stats summary. The numbers above are estimates. A full day-of-week breakdown can be computed from the
cvd_range_forecast.csv on request.
09 · Signal Frequency
How often does the Multi-Period CVD fire, and what does that mean for a discretionary trader?
Total Signals
14,926
All RTH 1-min bars
Avg Signals/Day
13.58
vs 390 RTH bars/day
% of Bars = Signal
3.5%
Not every bar; filtered events
Open-Window Signals
2,254
09:30–10:00 · 15.1% of total
Strong-Tier Signals
3,983
norm_mag ≥ 60 · 3.62/day
⚠ Frequency Warning: At 13.6 signals/day, a discretionary trader cannot act on every one. The recommended filters are: (1) Time window — open session 09:30–10:00 cuts to ~2.05/day with 2× the median move; (2) Magnitude tier — STRONG-only gives 3.62/day with the largest range; (3) Combined — Open + STRONG is roughly 0.8–1.2 signals/day, which is an ideal discretionary frequency.
10 · Chart Examples
Four representative signal examples — two bulls (open + afternoon) and two bears (open + afternoon). All from corrected signal data with validated direction labels.
11 · Drawdown & Stop Sizing
How large an adverse move do signals experience before the 2nd CVD crosses (signal degradation)? Use these numbers to size stops rationally.
Med Bars to 2nd Cross
11
Signal degrades within 11 bars on median
P50 Max Adverse Pre-2nd
37.5 pts
Before 2nd CVD cross occurs
P75 Max Adverse Pre-2nd
67.75 pts
Top-quarter adverse excursion
P90 Max Adverse Pre-2nd
111.25 pts
Top-10% worst adversity
% of Signals Still Majority-Aligned: The 2nd CVD crosses and the signal “officially” degrades on a median of 11 bars. But alignment decays fast: 70.8% still aligned at 5b → 54.9% at 10b → 40.7% at 15b → 33.5% at 20b → 9.5% at 60b. Targets within the first 10–15 bars catch the highest-probability aligned window.
Stop-Survival Heatmap
% of signals where adverse move NEVER exceeded the stop size during the holding period. Higher = fewer stops hit. Read: at a 20-pt stop held for 20 bars, 45.2% of signals survive without being stopped out.
| Stop Size | @ 5b | @ 10b | @ 15b | @ 20b | @ 30b | @ 60b |
|---|---|---|---|---|---|---|
| 2 pts | 1.9% | 1.4% | 1.2% | 1.1% | 0.9% | 0.7% |
| 4 pts | 7.3% | 5.2% | 4.3% | 3.7% | 3.2% | 2.4% |
| 6 pts | 14.7% | 10.7% | 9.1% | 8.0% | 6.6% | 5.0% |
| 8 pts | 24.1% | 18.1% | 15.5% | 13.8% | 11.4% | 8.6% |
| 10 pts | 33.2% | 25.6% | 21.8% | 19.4% | 16.2% | 12.2% |
| 15 pts | 53.0% | 42.7% | 37.0% | 33.4% | 28.4% | 21.7% |
| 20 pts | 66.9% | 56.6% | 49.8% | 45.2% | 39.2% | 30.4% |
| 30 pts | 83.0% | 74.3% | 68.0% | 63.0% | 55.9% | 44.6% |
How to read: Column = how long you hold. Row = stop size. Cell = % of signals that NEVER hit that stop during the holding period. A tight 10-pt stop held for 20 bars is only survived 19.4% of the time. A wider 30-pt stop held for 20 bars survives 63% of the time. For a 10-pt target: at 20 bars, 75.4% reach 10 pts favorable while only 19.4% would survive a 10-pt stop — a 1:1 R:R at this horizon has a structural advantage.
12 · Methodology & Honest Caveats
- Data: NQ E-mini Futures MBP-1 tick data, Jan 2022 – Apr 16 2026. 7.56B tick rows processed via
action == 'T'(trade filter). Resampled to 1-minute bars. - Signal definition: First bar where all three CVD windows (15, 30, 60 bars) simultaneously cross zero in the same direction. A BULL fires when CVD(15), CVD(30), and CVD(60) are all positive on bar N having been negative (or not all positive) on bar N-1. BEAR is the inverse.
- CVD definition: Cumulative Delta = buy volume minus sell volume, reset at each RTH session start (08:30 ET warm-up, 09:30 ET session start). Rolling windows are recalculated as a trailing sum of delta over the past N bars.
- Forward measurement: At each of 5, 10, 15, 20, 30, 60 bars post-signal:
fav = max HIGH - entry price (bull) / entry price - min LOW (bear);adv = max adverse excursion in opposite direction;net = bar close - entry price (signed by direction). - Normalized magnitude:
norm_mag = |cvd1|/15 + |cvd2|/30 + |cvd3|/60— units = delta per bar averaged across windows. - RTH only: All signals require entry between 09:30 and 15:45 ET. Pre-market and after-hours bars are excluded.
- Sign-convention correction: An earlier analysis build had the bull/bear labels inverted due to a CVD crossing condition that was negated. Detected via ATAS cross-validation in April 2026 and corrected. All numbers in this report are from the corrected build.
Honest Caveats — Read Before Trading
- Not a P&L backtest. No slippage, commissions, stops, or trade management applied. These are raw forward-movement measurements. Real results after commissions and slippage will differ.
- Adverse > Favorable at the median. The median adverse slightly exceeds the median favorable at every horizon. The signal’s tradable value is its hit rate to specific price targets, not a directional drift. Sizing targets correctly is mandatory — a naive “close at bar N” strategy has near-zero edge.
- Signal fires too often for discretionary trading. 13.6 signals/day requires filtering. Recommended filters: open window (09:30–10:00), STRONG magnitude tier (norm_mag ≥ 60), or both combined (~0.8–1.2/day).
- Open-window signals have higher adverse too. The open produces 2× the favorable move but also 2× the adverse. Wider stops are needed in the open window.
- All timestamps are US/Eastern (ET). ATAS displays on Central Time in some configurations — subtract 1 hour from all times shown here when cross-referencing CT-labeled charts.
- 2026 data is partial. Jan 2 – Apr 16 only. The 84.7% hit rate and 27.5 pt median for 2026 reflect an unusually high-volatility period and should not be extrapolated as “the new normal.”
Paid Service · NQ Futures
Want a report on your signal?
We run your signal against the same 4+ years of NQ tick data — same method as this study, applied to your specific entry condition.