1. Bottom Line & Directional Bias
Call: Bullish natural gas (NG=F, NYMEX November 2026 NGX26), with invalidation on a daily settle below 3.1553 (pivot S1).
Three reasons support the long side. First, price structure: the 2026-09-25 settle of 3.225 sits in the upper third of the 20-day 2.902–3.395 channel (65.5% position), with 5D +5.98% and 20D +5.01% — the trend of higher lows since 09-21 (2.976 low) is intact. Second, the term structure is only shallowly in contango, M1-M2 at -0.029 (-0.9%) with a slope of -0.0148; roll cost is modest, so a long position is not fighting a steep carry headwind. Third, the 09-25 -4.3% single-session drop was a rejection from the 3.395 20-day high, not a structural break: the settle held above the 09-23 close of 3.153 and above S1 at 3.1553.
The principal offset is positioning. Managed-money net was -132,799 contracts as of 2026-09-29, a -67,252 weekly change, with netPct -3.57% and crowding at the 27th percentile of the three-year window. Shorts are not crowded, so there is no mechanical squeeze to power a sharp rally; upside must come from spot tightness or weather, not from positioning.
Invalidation: a settle below 3.1553. A settle below 3.0857 (S2) would confirm failure.
2. Price Action & Technical Analysis
The prior final settlement was 3.225 on 2026-09-25, down -4.3% on the day, but still +5.98% over five sessions and +5.01% over twenty. The report-date bar (2026-09-27) is an unfinished Globex/Asia session and is not treated as a close.
The last five settled bars show a clear advance and a sharp rejection: 09-21 closed 2.993 (H 3.042 / L 2.976), 09-22 closed 3.117, 09-23 closed 3.153, 09-24 closed 3.37 after tagging 3.395, and 09-25 closed 3.225 after failing to hold above 3.3. That 09-24 spike to 3.395 is the top of the 20-day range (2.902–3.395); the 09-25 session retraced roughly 60% of the two-day advance but did not break the 09-23 close.
ATR14 is 0.1164, equal to 3.61% of price — a full expected daily range, not a one-sided allowance. RV20 is 39.2% annualized. With ATR at 3.61% of spot, any stop tighter than about 0.12 in price terms sits inside normal daily noise.
Pivots from the settle-based snapshot: P 3.2277, R1 3.2973, R2 3.3697, S1 3.1553, S2 3.0857. Price settled essentially on the pivot (3.225 vs P 3.2277), which is a neutral-to-constructive location: the market is balanced at the mid-point of the recent expansion, with R1 3.2973 as the first hurdle and the 20-day high 3.395 as the range ceiling. The 52-week range is 2.483–7.827, so the market is in the bottom third of its annual range — this is a recovery trade, not a breakout to new highs.
Weekly context: the last completed weekly bar (2026-09-14 to 2026-09-18) opened 3.014, high 3.09, low 2.984, closed 3.043, +2.01% w/w. The current week (from 2026-09-21, five sessions) is unfinished at 3.225 (+5.98%) and cannot be described as a weekly close or weekly breakout. The completed weekly bar does confirm a higher weekly low versus the 2.984 print and a close above 3, which is constructive but not decisive.
View: constructive while above 3.1553; the first objective is R1 3.2973, then the 20-day high 3.395.
3. Supply-Demand Balance & Fundamental Drivers
The most important fundamental fact in this snapshot is the shape of the curve. M1-M2 is -0.029, or -0.9%, with a slope of -0.0148 and a roll yield of -10.79% annualized. That is contango, but shallow contango: the front spread is under one cent, which means the prompt market is not signalling acute oversupply. In a genuinely loose gas balance, the front spread typically widens well beyond this, and roll yield becomes a material drag on long positions. At -0.9% for the front month, carry is a cost but not the dominant P&L driver over a one-to-two-week horizon.
What the price action tells us is consistent with a market that has absorbed a supply overhang: the 20-day range is 2.902–3.395, a 0.493-wide band, and price has moved from the lower half to the upper third over twenty sessions (+5.01%). A market in free fall on oversupply does not post a +5.98% five-day gain and hold above its pivot.
Macro transmission is indirect but relevant. The US 10-year yield is 5.184 (+0.43%) and DXY is 100.97 (-0.32%), both as of 2026-09-25. A softer dollar is a mild tailwind for dollar-denominated commodities including gas, though gas is far less dollar-sensitive than metals or crude because it is a domestically priced North American market. The more relevant macro channel is demand: the week-ahead calendar carries Core PCE at 0.3% m/m forecast and Final GDP at 1.5% q/q — a soft-growth, sticky-inflation mix that does not argue for a demand surge, but also does not argue for industrial collapse.
Seasonality is the swing factor. The next 20 sessions from this calendar start have a 15-year mean of -0.46% and a median of -1.85%, with only 7 of 15 years higher. That is a mild statistical headwind — the shoulder season between cooling demand and heating demand is historically soft. The bull case therefore cannot rest on seasonality; it must rest on the market having already priced the shoulder-season weakness and on the curve staying shallow.
View: balance is neutral-to-constructive; shallow contango plus a recovering 20-day trend argues against a fresh leg lower, but the seasonal window caps upside enthusiasm.
4. Positioning & Fund Flows
CFTC managed-money positioning is the clearest bearish input in this report. As of 2026-09-29, open interest was 1,782,129 contracts, with longs 220,522, shorts 353,321 and net -132,799, a weekly change of -67,252. The prior week (2026-09-22) showed net -65,547 with a +34,658 weekly change; 2026-09-15 showed net -100,205; 2026-09-08 showed net -96,742. The latest report is 15 days old relative to the report date and should be read as of its own date, not as current-week positioning.
The crowding table is more informative than the raw net. NetPct was -3.57% on 2026-09-22 with a crowding percentile of 27 on the three-year window, CTA at 74 and hedge at 27.41%. The prior three weeks were netPct -5.51% (crowd 33.17), -5.31% (crowd 32.48) and -4.95% (crowd 31.29). Two things follow. First, the net short is not crowded — a 27th percentile reading means positioning is far from an extreme, so there is no fuel for a violent short squeeze. Second, the CTA reading of 74 is high and unchanged across all four weeks, indicating trend-following proxies are already positioned for the move; that is a source of two-way risk if price stalls, because a trend signal reversal would force selling.
There is no divergence to flag: price rose over the 5D window while the most recent positioning data (as of 2026-09-29) showed a larger net short, but the positioning print is 15 days stale and the two observations are not contemporaneous, so this is a data-timing artefact rather than a genuine price/positioning divergence.
Implied volatility context: ^OVX (WTI implied vol) is 55.09, at the 58th percentile of one year, while ^VIX is 14.87 at the 9th percentile. Energy optionality is priced at a premium to equity optionality, and NG realized vol (RV20 39.2%) is elevated in absolute terms. With no NG-specific implied vol print available, the read is that energy event risk is being paid for broadly, which argues for selling premium rather than buying it, and for using wider stops on outright longs.
View: positioning is a headwind, not a trigger; the 27th percentile crowding means the short base can absorb rallies without being forced to cover.
5. Cross-Asset Relative Value
Cross-asset ratios available in this snapshot are limited to the dollar and rates complex, plus the energy implied-vol complex. DXY at 100.97 (-0.32%, 2026-09-25) and ^TNX at 5.184 (+0.43%, 2026-09-25) describe a market with a firm nominal yield but a softening dollar — a combination that historically supports real assets at the margin without providing a strong directional impulse to North American gas.
The more useful relative-value observation is within energy volatility. ^OVX at 55.09 (58th percentile) versus ^VIX at 14.87 (9th percentile) is a wide spread: the market is pricing substantial turbulence in crude while pricing near-complacency in equities. Natural gas, as a close substitute in the energy complex, trades in the same risk regime; when energy implied vol is at a mid-to-high percentile while broad risk vol is at a low percentile, the energy complex is where the event risk is concentrated. That argues for expressing gas views with defined risk rather than naked directional size.
^GVZ at 22.44 (13th percentile) and ^VXSLV at 35.99 confirm that the precious-metals complex is priced for calm, reinforcing that the current volatility premium is an energy-specific phenomenon, not a broad macro one.
View: relative value favours defined-risk long gas exposure over outright leverage; the energy vol premium is the dominant cross-asset signal available.
6. Historical & Seasonal Patterns
Seasonality for the same calendar start over the next 20 sessions, measured across the last 15 years: mean -0.46%, median -1.85%, with the market higher in 7 of 15 years. The best instance was 2020 at +43.93% and the worst was 2015 at -19.55%. The sample is small and the distribution is fat-tailed — the mean is dragged by the 2020 outlier, while the median is the more honest central estimate at -1.85%.
What this means for the current setup: the seasonal window is a mild headwind, and the median outcome is a modest decline. A bullish call therefore requires the market to outperform its seasonal baseline, which is exactly what the last five settled sessions have done (+5.98% versus a seasonal median of -1.85% for a comparable window). The risk is mean reversion to the seasonal path: the 09-25 -4.3% session is a reminder that shoulder-season rallies in gas are frequently retraced.
View: seasonality argues for smaller size and tighter discipline on the long side, not for abandoning the trade; the burden of proof sits with the bulls to keep posting higher lows.
7. Scenario Analysis (Base / Bull / Bear)
Base case — 50%: grind higher toward R1 3.2973, then the 20-day high 3.395. Trigger: price holds above pivot P 3.2277 on a settle basis and the front spread stays near -0.029. Target: 3.2973 initially, 3.395 on extension. Action: hold the long from the 3.16 area, trail the stop to 3.04, take partial profit into 3.37–3.4. This is the base case and it agrees with the section 1 call.
Bull case — 25%: breakout above 3.395 on a settle basis, opening a run toward the 52-week range midpoint. Trigger: a settle above 3.395 with the front spread flattening toward zero or flipping to backwardation, which would signal genuine prompt tightness rather than a short-covering bounce. Target: 3.5–3.6, with 3.3697 (R2) as the immediate gate. Action: add on a confirmed settle above 3.395, move the stop to 3.22, and let the position run while the curve stays flat. Note that the 27th-percentile crowding means this path requires a fundamental catalyst, not positioning.
Bear case — 25%: rejection from the 3.395 high extends, and price settles below 3.1553. Trigger: a daily settle below S1 3.1553, ideally accompanied by a widening front contango beyond -0.029 and a CTA trend signal turning negative. Target: 3.0857 (S2), then 2.976 (the 09-21 low). Action: exit the long on the S1 settle, stand aside, and re-engage only on a reclaim of 3.2277. A settle below 3.0857 would confirm trend failure and shift the bias to neutral/bearish.
Probabilities sum to 100%. The base case is the highest-probability path because price settled essentially on the pivot with a shallow curve and an intact sequence of higher lows; the bull and bear tails are roughly symmetric, with the bear tail slightly better supported by the seasonal median of -1.85%.
8. Trading Strategies & Risk Management
Strategy 1 — Core long (conviction 6). Entry 3.16, stop 3.04, target 3.4, horizon 1–2 weeks. The entry sits just below the pivot P 3.2277 and above S1 3.1553, so a fill requires only a modest pullback into support. The stop at 3.04 is below S2 3.0857 and roughly 0.12 below entry, which is approximately one ATR14 (0.1164) — outside normal daily noise. Target 3.4 is just above the 20-day high of 3.395. Size at half normal risk budget given the seasonal headwind and the stale, non-crowded short base.
Strategy 2 — Add on strength (conviction 5). Entry 3.1, stop 2.98, target 3.37, horizon 1–3 weeks. This is a deeper pullback entry near the 09-21 low of 2.976 and below S2, for accounts willing to hold through a retest of the lower half of the 20-day channel. The stop at 2.98 sits just below the 09-21 low and about 0.12 below entry, again roughly one ATR14. Target 3.37 is just below R2 3.3697.
Risk management: both strategies are long-only and consistent with the section 1 call. Do not add to either position on a settle below 3.1553. If the front spread widens materially beyond -0.029, reduce size regardless of price, because a widening contango would signal that the balance has deteriorated. Given RV20 at 39.2% and the energy vol premium (^OVX 55.09, 58th percentile), position sizing should assume daily swings of 3–4% are routine.
9. This Week's Data Calendar
BJT 09-29 22:00 | ET 09-29 10:00 — JOLTS Job Openings (F 7.23M, P 7.27M; surprise outside F±0.04M). BJT 09-30 09:30 | ET 09-29 21:30 — China Manufacturing PMI (F 50.1, P 49.8) and Non-Manufacturing PMI (F 49.2, P 49.0). BJT 09-30 20:30 | ET 09-30 08:30 — US Core PCE m/m (F 0.3%, P 0.2%), Final GDP q/q (F 1.5%), Personal Spending (F 0.8%). BJT 09-30 22:30 | ET 09-30 10:30 — EIA Crude and Gasoline Stocks. BJT 10-01 22:00 | ET 10-01 10:00 — ISM Manufacturing PMI (F 54.8, P 54.6).
This report is generated automatically from public quantitative and macro data for research and market tracking only. It does not constitute investment advice or a recommendation to trade.