1. Price Action & Technical Analysis
Natural gas settled at 3.115 on Mon Sep 28, down 3.41% on the day, after a 4.3% drop on Fri Sep 25 and a 6.88% spike on Thu Sep 24. The sequence matters more than any single print: the market gapped into a breakout above the prior 20-day range, failed to hold it, and has now given back essentially the entire impulse. From the Sep 24 close of 3.37 to the Sep 28 close of 3.115, the contract has lost 7.57% in two sessions. That is a textbook failed breakout, and failed breakouts in natural gas tend to resolve back toward the pre-breakout value area unless a fresh fundamental catalyst appears.
The medium-term trend is still positive on a simple rate-of-change basis. The 5-day change is +8.79% and the 20-day change is +13.54%, both measured into the Sep 28 close. But those figures are flattered by the Sep 22-24 advance (3.117 to 3.37, a 8.12% three-session gain) and now understate the deterioration in market structure. Momentum is decelerating: the last two sessions have erased 7.57%, and the close of 3.115 is below the Sep 22 close of 3.117. In other words, the entire breakout week has been round-tripped.
The 20-day range is best defined by the Sep 23 reference levels: pivot P at 3.157, first resistance R1 at 3.211, first support S1 at 3.091. The Sep 28 close of 3.115 sits between S1 and P, in the lower half of the immediate range and below the pivot. That is a bearish position within the range. The Sep 24 high close of 3.37 is the range top; the Sep 22 close of 3.117 and the Sep 28 close of 3.115 define a congested 3.11-3.12 shelf that is now the first line of defense. A daily close below 3.091 would confirm the failed breakout and open the 3 psychological level.
ATR14 is 0.1195, roughly 3.8% of the last close. That is a wide daily band and it changes how levels should be traded. A move from 3.115 to the 3.091 S1 is only 0.024, so S1 is not a meaningful stop distance on its own. Conversely, a retest of R1 at 3.211 is 0.096 away, and a stop beyond the Sep 24 close of 3.37 would be 0.255.
Volume confirms the rejection. The Sep 24 spike traded 264,747 contracts; the Sep 25 reversal traded 215,067; the Sep 28 follow-through traded only 3,230 contracts, a very thin tape that likely reflects a partial session or holiday-adjacent liquidity. Thin follow-through on a breakdown is a caution flag: it does not invalidate the failed breakout, but it argues against pressing shorts at the low. The better expression is to sell strength into the 3.2-3.22 retest zone rather than sell the 3.115 print.
Open interest on the Sep 23 snapshot was 449,516 contracts with change-in-position at 83.4%, a high churn reading that is consistent with the violent two-way trade of that week. The subsequent COT open interest figures (1,837,146 on Sep 22) are on a different reporting basis and should not be mixed with the daily OI print. The takeaway from the daily data is simply that positioning was turning over aggressively into the spike, which is exactly the condition that produces a failed breakout when the marginal buyer disappears.
2. Fundamental Drivers
The macro backdrop is not supportive of a sustained natural gas rally. The US 10-year yield is 5.184, up 0.43% on the Sep 25 reading, and the dollar index is 100.97, down 0.32%. A 5.18% 10-year yield is a restrictive real-rate environment; it raises the discount rate on all long-duration assets and, more specifically, it keeps industrial demand expectations contained. The dollar at 100.97 is off its highs but still firm. For a dollar-denominated commodity with no yield, the combination of high nominal yields and a stable-to-firm dollar is a persistent headwind.
The curve is the single most important fundamental fact in this report. Natural gas is in contango: M1-M2 is -0.029, or -0.9%, and the 12-month roll is -10.79% with a slope of -0.0148. Contango means the front month is cheaper than the back, so a long position bleeds carry every day it is held. The annualized roll cost of roughly 10.79% is a substantial hurdle. For a speculative long to be profitable, the front month must rally by more than the roll cost plus financing. That is a high bar in a market where the front spread is already negative. It also explains why rallies in this contract tend to be sharp and short-lived: they are squeezes against a carry-negative structure, not trend moves.
Inventories and central-bank flows: the calendar this week is dominated by US inflation and labor data. Core PCE m/m for August is forecast at 0.3% versus 0.2% prior, Personal Spending at 0.8% versus 0.2%, and Final GDP q/q at 1.5%. On Fri Oct 2, Non-Farm Employment Change is forecast at 98K versus 162K prior, with the unemployment rate steady at 4.1% and average hourly earnings at 0.3% m/m. A 98K payroll print would be a sharp slowdown and would normally be bullish for rate-cut expectations, which would weaken the dollar and support commodities. But the same print would also signal weaker industrial activity, which is bearish for gas demand. The net effect is ambiguous, and that ambiguity is itself a reason not to hold large directional risk into Friday.
Geopolitics is running hot in energy-adjacent headlines. Oil jumped higher at the Globex open for the new week; Iran's Hormuz proposal was called a cynical bid for upfront concessions; a Hormuz missile strike raised risk before the open; and Riyadh schools shifted to remote learning after Houthis claimed a downed drone. These are oil-market headlines, but they matter for gas through two channels. First, a broader energy risk premium can lift all hydrocarbons, including gas, on a beta basis. Second, if Middle East risk escalates, LNG shipping and European gas supply routes become part of the story, which can tighten the global gas balance. The absence of any direct natural gas headline in the last 48 hours means the gas market is currently trading on its own curve and positioning, not on a gas-specific catalyst. That is important: without a gas-specific supply or demand shock, the contango will dominate.
On the demand side, the Chinese PMI data on Tue Sep 29 (New York) / Wed Sep 30 (Beijing) is the key macro input. Manufacturing PMI is forecast at 50.1 versus 49.8 prior, and Non-Manufacturing at 49.3 versus 49.0. A move above 50 in manufacturing would be the first expansionary print in some time and would support industrial demand expectations, including for LNG. The RatingDog Manufacturing PMI is forecast at 51.7 versus 51.5, already in expansion. If the official manufacturing PMI confirms the 50.1 forecast, the reflation narrative gets a modest boost, which is a mild positive for gas. If it misses, the demand side weakens further and the contango case strengthens.
3. Positioning & Fund Flows
The COT data through Sep 22 shows a meaningful but incomplete short-covering episode. Speculative longs rose to 266,563 from 264,362, while shorts fell to 332,110 from 364,567. Net positioning improved to -65,547 from -100,205, a 34,658-contract reduction in net shorts in one week. That is a large weekly delta and it explains the Sep 22-24 price spike: shorts were covering into a thin market. But the net is still negative, and the prior three weeks show the short base was being built, not dismantled: net was -89,481 on Sep 1, -96,742 on Sep 8, and -100,205 on Sep 15. The Sep 22 improvement is the first meaningful reversal, but it leaves the market still net short by a wide margin.
Crowding is the more nuanced signal. Net positioning as a percentage of open interest is -3.57%, up from -5.51% the prior week. The crowding score is 27, up from 33.17. In this framework, a lower crowding score means less crowded positioning. So the market has moved from a moderately crowded short to a less crowded short. CTA positioning is unchanged at 74 across all four weeks, and hedge positioning is 27.41%, up from 26.69%. The stable CTA reading suggests trend-following flows are not the marginal driver here; the change is coming from discretionary and hedging activity. The hedge ratio rising while net shorts fall is consistent with producers adding hedges into the spike, which caps rallies.
The combination of a still-negative net position and a low crowding score is a two-sided signal. On one hand, the short base is large enough that further covering could produce another squeeze. On the other hand, the crowding score of 27 says the short is no longer extreme, so the marginal squeeze fuel is less potent than it was two weeks ago. This is exactly the setup where a failed breakout is more likely than a sustained trend: the easy shorts have covered, and the remaining shorts are more committed. Momentum arguments should be labeled explicitly here: the price momentum is still positive on a 5-day and 20-day basis, but the positioning momentum (the rate of short covering) has already peaked. When price momentum and positioning momentum diverge, the positioning signal usually wins over a 1-2 week horizon.
Options and volatility: the only volatility proxy in the data is the CBOE VIX at 14.87, down 5.11% on Sep 25. A VIX below 15 is a low-volatility regime. For natural gas, a low macro-vol backdrop typically compresses implied volatility in energy options as well, which reduces the cost of optionality but also signals complacency. In a contango market with a failed breakout, low vol favors selling rallies over buying breakouts. There is no gas-specific implied volatility reading in the data, so we do not infer one; the VIX is used only as a macro risk-appetite gauge.
4. Cross-Asset Relative Value
The cross-asset data set is limited to the dollar index, the 10-year yield, and the VIX, plus the headline flow across oil, copper, and macro. We do not have gold, silver, or copper price levels in the data, so we cannot compute the gold-silver, oil-gold, or copper-gold ratios or their percentiles. We will not fabricate them. What we can say is directional and qualitative, anchored to the numbers we do have.
The dollar index at 100.97, down 0.32%, is the primary cross-asset transmission channel. A softer dollar is a mild tailwind for dollar-denominated commodities, but a 0.32% move is small relative to the 3.41% daily drop in gas. The gas move is idiosyncratic, not a macro-dollar story. The 10-year yield at 5.184, up 0.43%, is the more important cross-asset signal: rising long yields tighten financial conditions and weigh on the more industrial, demand-sensitive commodities. Gas is partly industrial, partly weather-driven, so it sits in the middle. The fact that gas fell 3.41% on a day when the dollar fell 0.32% tells you the sell-off was driven by curve and positioning, not by macro.
The oil headlines are the most relevant cross-asset input. Oil jumped at the Globex open on Middle East risk, and the Hormuz headlines are escalating. Historically, gas and oil can decouple for long periods, but in acute geopolitical energy shocks they correlate higher. If oil sustains a risk premium, gas can catch a sympathy bid, particularly in the front month, which would be the mechanism for another squeeze attempt. The counterargument is that the gas curve is in contango while oil is often in backwardation during supply shocks; the carry structures are different, so the sympathy bid in gas is less durable. A high crack spread during a supply shock is not automatically strong demand for the whole complex; it can be a refinery constraint signal. We apply the same discipline to gas: a geopolitical spike is not the same as structural demand growth.
Copper headlines (Zimbabwe lithium, Donlin mine spill) are supply-side mining stories with limited direct read-through to gas. The BOJ minutes and the Bessent comments on Fed rates are macro inputs that affect the dollar and real yields, which we have already covered. Net: the cross-asset backdrop is mixed-to-negative for gas. The dollar is slightly softer, which helps, but high long yields and a contango curve dominate. Without a gas-specific catalyst, relative value does not justify a long.
5. Sentiment & News Monitor
There are no natural gas-specific headlines in the last 48 hours. The sentiment read is therefore derived from price action and the cross-asset tape. Price sentiment is bearish: two consecutive down sessions totaling 7.57% from the Sep 24 close, a failed breakout, and a close below the pivot. The 48-hour headline bias is neutral-to-bearish for gas specifically, with the energy complex headlines skewed to oil geopolitics (Hormuz, Houthis, Iran) that could produce a sympathy bid but are not gas fundamentals. Macro headlines are mixed: BOJ minutes and a September rate hike reference, Bessent urging the Fed to keep an open mind on rates citing AI productivity, and a soft dollar. The VIX at 14.87, down 5.11%, signals calm macro risk appetite, which is not the environment that produces sustained commodity squeezes. Overall sentiment score: mildly bearish, with a squeeze-risk caveat from the still-net-short positioning.
6. Historical & Seasonal Patterns
We do not have multi-year seasonal price data in the data set, so we will not invent seasonal averages or five-year ranges. What we can do is place today's price using the historical percentiles and spread Z-scores that are available. The available historical anchors are the 20-day change of +13.54% and the 5-day change of +8.79%, both measured into the Sep 28 close. A 20-day gain of 13.54% is a strong short-term move, but it has been substantially retraced in the last two sessions. We will not fabricate that percentile.
The spread Z-scores: the M1-M2 spread is -0.029, or -0.9%, with a slope of -0.0148 and a 12-month roll of -10.79%. The contango is persistent and the annualized roll cost is the dominant historical pattern here. In a contango market, the base rate for long positions is negative carry; the base rate for short positions is positive carry. That is the single most important historical/structural fact for positioning. The COT net percentile is -3.57% of open interest, and the crowding score is 27. A net percentile at a negative value means the numerator (net speculative length) is below zero, i.e., the market is net short; mean reversion from a net-short extreme favors the upside, but the crowding score of 27 says the short is no longer extreme. The conflict between the still-negative net percentile and the reduced crowding is resolved by the price action: the mean-reversion fuel has already been partly spent in the Sep 22-24 squeeze, and the failed breakout suggests the remaining fuel is insufficient to sustain a trend.
7. Bull/Bear Scenario Analysis
Bull case (≥4):
- Short covering resumes: net positioning is still -65,547 contracts and netPct is -3.57%; a fresh catalyst could force another 30-50K of short covering, similar to the 34,658 delta seen in the Sep 22 week.
- Middle East escalation: Hormuz missile strike and Houthi drone activity could lift the entire energy complex, and gas would catch a sympathy bid, especially if LNG shipping risk is repriced.
- Chinese reflation: official Manufacturing PMI forecast at 50.1 versus 49.8 prior, and RatingDog Manufacturing at 51.7; a beat would support industrial and LNG demand expectations.
- Soft dollar: DXY at 100.97, down 0.32%, and a weak 98K payroll print on Fri Oct 2 could push the dollar lower and lift dollar-denominated commodities.
- Low VIX at 14.87: calm macro vol can precede a positioning-driven squeeze when a market is still net short.
Bear case (≥4):
- Contango carry: M1-M2 at -0.029 (-0.9%) and a 12-month roll of -10.79% mean longs pay to hold; this caps rallies and rewards shorts.
- Failed breakout: the Sep 24 spike to 3.37 has been fully retraced to 3.115; failed breakouts typically resolve toward the pre-breakout range, with S1 at 3.091 the first target.
- High long yields: 10-year at 5.184, up 0.43%, tightens financial conditions and weighs on industrial demand expectations.
- Crowding reduction: crowding at 27 versus 33.17 the prior week means the short base is less extreme; the marginal squeeze fuel is diminished.
- Producer hedging: hedge ratio at 27.41%, up from 26.69%, suggests producers are selling rallies, capping upside.
- Thin follow-through volume: the Sep 28 session traded only 3,230 contracts, so the breakdown lacks conviction, but the absence of buying is itself a bearish tell.
Near-term balance (1-2 weeks): bearish, with squeeze risk. The curve and the failed breakout dominate. Medium-term balance (1-3 months): neutral-to-bearish unless a gas-specific supply shock or a sustained Chinese demand recovery appears. The contango structure means the path of least resistance is lower, but the still-net-short positioning means rallies can be violent. Trade the range, not the trend.
8. Trading Strategies & Risk Management
Primary plan: sell a retest of the 3.2-3.22 resistance zone. Entry at 3.21. Stop at 3.34. Target at 3.05, below the S1 level of 3.091 and the pre-breakout congestion. Reward/risk: 0.16 reward versus 0.13 risk, about 1.23. That is below the 1.5 threshold, so we adjust: entry at 3.22, stop at 3.34, target at 3.04 (0.18 reward), reward/risk 1.5. Entry is 0.105 above the last close, well within range. Horizon: 1-5 days. Size: half of normal risk budget given the wide ATR and thin liquidity. Invalidation: a daily close above 3.34, which would signal the failed breakout has been negated and the squeeze is back on; in that case, stand aside and reassess. Do not chase the short at 3.115; the thin Sep 28 volume and the proximity to S1 at 3.091 make the risk/reward poor at the low. If price never retests 3.22 and instead breaks 3.091 on rising volume, the plan is void and we stand aside rather than sell into a stretched move.
9. This Week's Data Calendar
| Time (ET / Beijing) | Event | Forecast | Previous |
|---|
| Sun Sep 27, 19:50 / Mon Sep 28, 07:50 | BoJ Minutes | - | - |
| Tue Sep 29, 00:30 / Tue Sep 29, 12:30 | RBA Cash Rate | 4.6% | 4.35% |
| Tue Sep 29, 21:30 / Wed Sep 30, 09:30 | China Manufacturing PMI | 50.1 | 49.8 |
| Tue Sep 29, 21:30 / Wed Sep 30, 09:30 | China Non-Manufacturing PMI | 49.3 | 49.0 |
| Tue Sep 29, 21:45 / Wed Sep 30, 09:45 | RatingDog Manufacturing PMI | 51.7 | 51.5 |
| Wed Sep 30, 08:30 / Wed Sep 30, 20:30 | US Core PCE m/m | 0.3% | 0.2% |
| Wed Sep 30, 08:30 / Wed Sep 30, 20:30 | US Final GDP q/q | 1.5% | 1.5% |
| Fri Oct 2, 08:30 / Fri Oct 2, 20:30 | US Non-Farm Payrolls | 98K | 162K |
| Fri Oct 2, 08:30 / Fri Oct 2, 20:30 | US Unemployment Rate | 4.1% | 4.1% |
| Mon Oct 5, 10:00 / Mon Oct 5, 22:00 | ISM Services PMI | 54 | 55.4 |
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.