1. Bottom Line & Directional Bias
Call: Bearish NG=F. Invalidation: a daily settle above 3.14 (R2 from the settle-based pivot set).
Three reasons. (1) Price structure: the 2026-10-02 settle was 3.04, down 5.89% over five sessions and effectively flat over 20 sessions (-0.07%), holding the lower third of the 20-day 2.9–3.4 channel (position 27%). The last completed weekly bar (2026-09-21–25) closed at 3.23, +5.98% w/w, and that gain has since been fully retraced — the current week is unfinished and cannot be used for a weekly-close conclusion. (2) Positioning: managed-money net length fell to -132,799 contracts on 2026-09-29, -7.45% of open interest, a -67,252 weekly change, with crowding at the 38th percentile of the three-year window — a large but not yet extreme short, leaving room to extend. (3) Curve: M1-M2 is -0.306 (-9.16%) with roll yield -109.9%, i.e. contango, so longs carry a roll cost and there is no prompt-tightness signal to anchor a rally.
Risk to the call is a macro-driven dollar move or a cold-weather headline; the 3.14 level is the line where the bearish structure fails.
2. Price Action & Technical Analysis
The prior session settle (2026-10-02) was 3.04, +2.29% on the day but -5.89% over five sessions and -0.07% over 20 sessions. The 20-day channel is 2.9–3.4, with price at the 27th percentile of that range — lower third, but not at the floor. The 52-week range is 2.48–7.83, so the market is trading in the bottom quarter of its annual envelope, which argues against treating this as a mid-range chop.
Volatility is the dominant tactical fact. ATR14 is 0.127, or 4.19% of price as a full daily range — not a one-sided move. RV20 is 43.2% annualized. Any stop placed inside roughly 0.13 of entry is inside normal daily noise; the 3.14 invalidation is 0.1 above the 3.04 settle, which is why the level, not a tight stop, defines the view.
Pivots from the settle-based set: P 3, R1 3.09, S1 2.95, R2 3.14, S2 2.86. The ordering matters. Price at 3.04 sits above P and below R1; the first meaningful cap is 3.09, and the bear case needs a rejection there rather than a break. Below, S1 2.95 is the first shelf and S2 2.86 is the measured downside objective — the same 2.86 that the house trade idea targets.
On the weekly frame, the last completed bar (2026-09-21–25) opened 3.04, traded 2.98–3.4 and closed 3.23, a +5.98% weekly gain. That was a strong bar, and it has not been built on: the unfinished current week (from 2026-09-28, five sessions) shows 3.04, -5.89%. The correct read is a failed follow-through, not a weekly reversal signal, because the current week has not closed. The 3.4 weekly high is the level that would have to be reclaimed to invalidate the bearish structure on a higher timeframe; the nearer, tradeable invalidation is 3.14.
View: bearish while below 3.09–3.14; first objective 2.95, then 2.86.
3. Supply-Demand Balance & Fundamental Drivers
The structural signal in this market is the shape of the curve. M1-M2 is -0.306, or -9.16%, with a roll yield of -109.9% and a slope of -0.0206. That is contango: the front month trades below the second month, so a long position pays to roll and the market is not pricing prompt scarcity. In a genuinely tight gas balance, the front spread typically moves toward flat or backwardation; here it is doing the opposite. This is the single most important fundamental fact in the report and it is consistent with the bearish call.
Positioning data corroborates a market that has been supplied rather than starved. Open interest on 2026-09-29 was 1,782,129 contracts, down from 1,837,146 the prior week — a decline in total participation alongside a sharp increase in shorts. Producer/merchant hedging sits at 27.96% of the crowding framework, essentially unchanged from 27.41% a week earlier, so the commercial side is not aggressively adding hedges at these prices; the pressure is coming from the managed-money side.
Macro transmission is indirect but real. The US 10-year yield at 5.28% and DXY at 101.93 (-0.17%) matter to gas mainly through the industrial demand channel and through the cost of capital for producers. A firm dollar and elevated long rates are, at the margin, a headwind to industrial load and to the marginal drilling economics that set future supply. Neither is decisive this week; both lean mildly against a demand-led rally.
What is missing from the picture is a hard inventory number versus the five-year average. Without it, the honest fundamental read rests on the curve and on positioning: contango says the market is not short of prompt supply, and the short build says participants agree. The burden of proof is on the bulls to show a tightening front spread before this view changes.
View: structurally bearish; the trigger to revisit would be M1-M2 moving toward flat.
4. Positioning & Fund Flows
The COT sequence is the cleanest bearish evidence in the dataset. Managed-money net went from -96,742 (2026-09-08) to -100,205 (2026-09-15) to -65,547 (2026-09-22) and then to -132,799 (2026-09-29). The 2026-09-22 week saw a +34,658 short-covering bounce — and price responded, with the last completed weekly bar closing 3.23, +5.98%. That bounce was then sold: the following week net length collapsed by -67,252 contracts, the largest weekly swing in the four-week window, and price gave back the entire move to 3.04.
The critical nuance is crowding. Net as a percentage of open interest is -7.45%, but the three-year crowding percentile is only 38.4 — up from 27.0 the prior week, yet still below the midpoint. A short position this large that is not yet in the top half of its historical distribution is not a crowded trade. That matters for two reasons: it argues against an imminent short-squeeze, and it leaves room for further selling without positioning becoming an obstacle.
The CTA trend proxy is steady at 38 across all four weeks, and hedge participation is stable near 27–28%. Neither shows the capitulation or the momentum chase that typically marks a tradeable low. Open interest falling from 1,837,146 to 1,782,129 while shorts increase means the marginal seller is not being offset by new longs — a one-sided flow.
On volatility, RV20 is 43.2% annualized. The dataset does not carry a natural-gas implied-vol index, so no IV-versus-RV conclusion is drawn here; the practical implication is simply that realized movement is large and position sizing must reflect it.
View: short positioning is building, not exhausted; the 38th-percentile crowding reading supports continuation rather than reversal.
5. Cross-Asset Relative Value
Cross-asset context is thin for gas specifically, but two readings frame the macro backdrop. ^VIX at 15.31 (-1.08 points) sits at the 15th percentile of its one-year range, and ^GVZ at 23.23 is also at the 15th percentile — broad risk appetite is calm and gold volatility is subdued. ^OVX at 51 is at the 49th percentile, mid-range. The read-through is that there is no macro-volatility regime forcing commodity de-risking; gas is trading on its own balance, which is consistent with the curve and positioning signals rather than a macro story.
DXY at 101.93, -0.17% on the day, is a mild tailwind for dollar-denominated commodities at the margin, but a single 0.17% move is not a driver. The 10-year at 5.28% is the more relevant cross-asset input: it keeps financing costs elevated for shale producers and, at the margin, restrains the drilling response that would otherwise cap rallies. That is a slow-moving support for price, not a reason to be long today.
Relative to the broader commodity complex, gas is the outlier in this dataset: oil implied vol is mid-range and equity vol is cheap, while gas realized vol is 43.2%. That dispersion is the tradeable feature — gas is where the movement is, and the movement has been down.
View: no cross-asset signal strong enough to override the domestic curve and positioning case; neutral-to-mildly-supportive macro, bearish instrument-specific setup.
6. Historical & Seasonal Patterns
Seasonality for the same calendar start over the next 20 sessions, measured across the last 15 years, is positive: mean +3.37%, median +2.27%, and up in 9 of 15 years. The best instance was 2025 at +27.08% and the worst was 2022 at -16.43%. This is the one block in the dataset that argues against the bearish call, and it deserves to be stated plainly rather than buried.
The correct treatment is to weight it as context, not as a signal. A 9-of-15 hit rate is 60% — better than a coin flip but far from a reliable edge, and the sample is small. The dispersion is enormous: a +27.08% best case and a -16.43% worst case in the same seasonal window means the distribution is wide enough that the mean carries little information for a single occurrence. The 2022 analogue is a reminder that when the balance is loose, seasonality does not save the longs.
What would make seasonality actionable is confirmation from the other blocks — a flattening front spread, or short positioning reaching an extreme percentile. Neither is present: the curve is in contango and crowding is at 38.4. Until those change, the seasonal tailwind is a reason to respect the 3.14 invalidation and size conservatively, not a reason to be long.
View: seasonal bias is positive but unconfirmed; it raises the bar for adding to shorts rather than reversing the call.
7. Scenario Analysis (Base / Bull / Bear)
Base case — 50%: grind lower toward 2.95, then 2.86. Trigger: price holds below R1 3.09 and the front spread stays in contango. The 2026-10-02 settle of 3.04 is already below R1, and the 5D -5.89% shows the bounce failing. Target 2.95 (S1) first, with 2.86 (S2) as the extension. Action: maintain the short bias, trail risk above 3.09. This scenario agrees with the section 1 call.
Bull case — 25%: reclaim 3.09 and squeeze toward 3.14, with 3.4 as the tail. Trigger: a daily settle above 3.09 on rising open interest, or a front-spread move toward flat. The seasonal block (mean +3.37%, up 9 of 15 years) is the supporting evidence, and the 2026-09-22 short-covering week (+34,658 contracts, price to 3.23) shows how fast this market can move when shorts cover. Action: stand aside or reduce; a settle above 3.14 invalidates the bearish view outright and shifts the tactical frame to the 3.4 weekly high.
Bear case — 25%: break 2.95 and extend to 2.86 and below. Trigger: a daily settle below 2.95 accompanied by another week of managed-money net selling beyond -132,799. The 52-week low at 2.48 is the longer-horizon reference if 2.86 gives way. Action: add on the break with risk back above 3.04, and take partial profit into 2.86 given ATR14 of 0.127 means the move can happen in two sessions.
Probabilities sum to 100%. The base case carries the call; the bull case is the defined invalidation path, not a second conclusion.
8. Trading Strategies & Risk Management
Strategy 1 — Short NGX26 on strength (primary). Entry 3.04 (the 2026-10-02 settle), stop 3.14 (R2, the invalidation level), target 2.86 (S2), horizon 1–5 days, conviction 7/10. Size so that a 0.1 adverse move is a normal loss: with ATR14 at 0.127, a stop 0.1 away is inside one day's expected range, so use half-normal size and add only on a rejection at 3.09. This is the same structure as the house trade idea.
Strategy 2 — Scale-out short into 2.95. Entry 3.04, stop 3.14, first target 2.95 (S1), runner to 2.86, horizon 1–5 days, conviction 6/10. Take half off at 2.95 and trail the remainder above 3.04. This monetizes the base-case path while keeping exposure to the bear-case extension toward the 2.48 52-week low.
Risk management: no long exposure while price is below 3.09 and the curve is in contango. The seasonal tailwind (mean +3.37% over the next 20 sessions) is the reason to keep stops at the level rather than tighter, and the reason conviction is 7 rather than higher. A daily settle above 3.14 ends both strategies.
9. This Week's Data Calendar
- BJT 10-05 22:00 | ET 10-05 10:00 — ISM Services PMI (SEP), forecast 54 vs prior 55.4; surprise if outside 54 ± 1.4. Affects DXY and the industrial-demand read-through.
- BJT 10-07 04:30 | ET 10-06 16:30 — API Crude Oil Stock Change (OCT/02).
- BJT 10-07 22:30 | ET 10-07 10:30 — EIA Crude Oil and Gasoline Stocks Change (OCT/02).
- BJT 10-08 02:00 | ET 10-07 14:00 — FOMC Minutes. Affects rates, DXY and the broader commodity complex.
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.