Thanks very much for starting the new thread, Larry.

You are exactly right about this being the bread and butter weather trading market in the heating season!

This was the previous thread:

                NG 8/4/25-10/9/25            

                            77 responses |              

                Started by WxFollower - Aug. 4, 2025, 12:26 a.m.        

    https://www.marketforum.com/forum/topic/113735/

                Re: Re: Re: Re: Re: NG 8/4/25+            

                            By metmike - Oct. 9, 2025, 1:33 a.m.            

Near record warmth in the NorthCentral US last week where HDDs are normally higher than CDDs so the warmth there was bearish.

Plenty of warmth to go around in other places except a bit cool West/Southwest compared to average. 

https://www.cpc.ncep.noaa.gov/products/tanal/temp_analyses.php

+++++++++++++

Actual EIA number was close to expected. Storage levels will likely peak in November before the first major cold snap. 

Stocks are slightly above last year. However, last Winter was very cold compared to average. Look at the graph below and not THE BLUE LINE. This is the storage level of the last EIA report. 

Storage went from near record highs a year, to well below the 5 year average and near the bottom of the 5 year range!!!!

This shows, as Larry stated why temperatures in the HDD season are so critical to storage and prices!!  So the current surplus, even higher than last year's(at the moment) is not enough to prevent storage  from getting low by early in 2026!!!

https://ir.eia.gov/ngs/ngs.html

  ‹ See All Natural Gas Reports

Weekly Natural Gas Storage Report

 for week ending October 3, 2025   |  Released: October 9, 2025 at 10:30 a.m.   |  Next Release: October 16, 2025 

Summary

Working gas in storage was 3,641 Bcf as of Friday, October 3, 2025, according to EIA estimates. This represents  a net increase of 80 Bcf from the previous week. Stocks were 23 Bcf higher than last year at this time and 157 Bcf above the five-year average of 3,484 Bcf. At 3,641 Bcf, total working gas is  within the five-year historical range.

 For information on sampling error in this report, see Estimated Measures of Sampling Variability table below. 

This is a time of year when ng futures move higher towards a seasonal peak ahead of Winter as the market dials in cold weather risk premium and storage fills up with extra buying.  Each year is different but this average over the past 20 years gives us clear tendencies at certain times of year.

Some seasonal rules that I use:

1. Being long in April/May almost always makes money, regardless of the weather.

2. Being short in Dec/Jan during widespread unusual warmth almost always makes BIG money.

3. Extreme cold during the first half of the heating season usually has a much higher impact on speculators buying than towards the end, unless storage is very low. During the back half of the heating season, the market knows that time is running out to draw down storage.  

4. Summer heat waves don't have the same RELIABLE impact on prices as extreme cold in Winter but are tradable.

5. The market, under some conditions (early Winter-pattern change to brutal cold) can start trading that weather 2 weeks before it starts. 

Got any more, Larry?

https://charts.equityclock.com/natural-gas-futures-ng-seasonal-chart

Natural gas price charts.

https://tradingeconomics.com/commodity/natural-gas

1. 10 years: Classical symmetrical wedge with lower highs and higher lows.

2022-highs from very low storage, followed by massive injections and near record high storage-low prices in 2023/24, followed by near record draw downs from the cold Winter of 24/25-spike to a lower high early 2025. Downtrend line since then, recently touched as the apex of the triangle is approaching.

2. 1 year: 

a. 2 strong uptrend channels from last Winters cold. Major market top in March 2025.

b. Late Spring/early Summer uptrend channel from Spring positive seasonal along with widespread intense heat forecasts for the Summer of 2025.

c. Downtrend channel that accelerated in late August from huge pattern change to widespread near record cool temps(and low electricity use for AC use,  generated by burning natural gas)

d. Bottom ahead of positive Fall seasonal and late season heat returning with elevated CDDs. 

e. Prices at the very end have plunged again as late season heat has turned from bullish to bearish. Rapidly increasing average HDDs pass up rapidly dropping CDDs on October 4th but every year is different. The VERY warm forecasts this week are now BEARISH weather.

3. 1 month: Early seasonal high???? This usually doesn't happen until well into November but varies from year to year. My guess is that a pattern change to widespread MUCH colder temps in the high population centers that has staying power will make new highs. 

The straight up line in the center of the graph was the price adjustment from the lower priced October contract rollover to the higher priced November contract that is the current front month.

Warm weather in late October is bearish.

Warm weather and depressed heating demand are not bearish natural gas.

This was the last European Ensemble model average. HDDs are on the left. Green dashed line shows how quickly the seasonal average is increasing. We will be well below that the next 2 weeks. Purple was the last 0z run. The tan line was the previous 12z run from 12 hours earlier.

Although the CDDs on the right are elevated, by the end of the 2 week period below, the average, climatology/historical HDDs are 8 times higher than the average CDDs. 

This futures market trades on weather patterns coming up and how they impact heating and cooling demand as far as 2+ weeks from now. The NG market is interpreting warm=bearish until, at least April.  However, ng stops trading temperature forecasts closely towards the end of Winter(unless storage is unusually low) because its impossible to draw down storage from robust levels when more than half of Winter is over.

This temperature product is no longer available, starting October 9, 2025(so I have the last date for data on October 8th below). Not sure if this is related to the government shutdown since its automated and all the other NOAA sites are working.

https://www.cpc.ncep.noaa.gov/products/tanal/temp_analyses.php

+++++++++++++++++++++++

The injection of 80 Bcf was close to the middle of average guesses. Maybe a tad bullish with the price also going slightly higher for a very short while after the release. 

https://ir.eia.gov/ngs/ngs.html

  ‹ See All Natural Gas Reports

Weekly Natural Gas Storage Report

 for week ending October 10, 2025   |  Released: October 16, 2025 at 10:30 a.m.   |  Next Release: October 23, 2025 

Summary

Working gas in storage was 3,721 Bcf as of Friday, October 10, 2025, according to EIA estimates. This represents  a net increase of 80 Bcf from the previous week. Stocks were 26 Bcf higher than last year at this time and 154 Bcf above the five-year average of 3,567 Bcf. At 3,721 Bcf, total working gas is  within the five-year historical range.

 For information on sampling error in this report, see Estimated Measures of Sampling Variability table below. 

Note: The shaded area indicates the range between the historical minimum and maximum values for the weekly series from 2020 through 2024. The dashed vertical lines indicate current and year-ago weekly periods.

The just out 12z European Ensemble model, as well as the other models continues with above average temperatures for the rest of October. HDDs on the left, last run is always purple.

++++++++++

https://tradingeconomics.com/commodity/natural-gas

1. 1 month: Clear double top. Downtrend in October from very mild temps. The huge blip straight up in the center was when the October contract rolled over the the higher priced and current November contract.

2. 1 year: Now closing below key $3 support.

3. 10 years: Threatening to break out to the downside of a long lived symmetrical triangle (lower highs and higher lows). 

A Trump recession will reduce INDUSTRIAL DEMAND and be bearish for natural gas prices. As long as temperatures remain mild, the downside is favored, despite this being the time of year with positive price seasonals. Any sign of a pattern shift to sharply colder should trigger a big bounce up. 

From the previous post:

Any sign of a pattern shift to sharply colder should trigger a big bounce up.

We just had a HUGE gap higher!

The models are looking colder(less mild) at the end of the 2 week period which, especially in the first half of Winter matters a great deal. The purple line on the left was the last HDD forecast from the last 12z European Ensemble model. CDDs on the right don't matter anymore.

My bad on the November outlook that I posted at first. That was a surface pressure anomaly map.

This is the correct temperature anomaly map for November from the last European weekly model run. 

However, late in the month/early December we have a significant +PNA/-NAO featuring a ridge west/trough east couplet which means cold in the East! This is a high population area that consumes a great deal of fuel to heat their homes.

That's a LONG ways off and LOW skill!!!

Temps to start December:

This was the huge +$1,500/contract higher open and gap up to start this week's trading. We closed last week close to long term $3 support after putting in what looks like a significant low from near record mild temps and low seasonal heating demand to start the HDD season this month. 

This is likely an upside break away gap.

https://tradingeconomics.com/commodity/natural-gas

What do you think, Larry?

8pm update:

These are prices the last year. The last month+ has set up a clear symmetrical wedge pattern with lower highs since the big March top(from last Winters cold drawing down storage-a 2nd  lower, early Summer high-then a 3rd, even lower high to start October, with very mild temps the first half of October pressuring prices.

We also saw a major late August/early September low from unusually cool temps at the end of Summer reducing cooling demand, followed by 2nd HIGHER low and last week, after prices had dropped from the mild October, a return to colder forecasts put in a 3rd, higher yet low. 

The big gap higher tonight from this formation is an upside breakaway gap that's a pretty bullish signal. 

If weather forecasts keep turning colder this week, it should hold. However, in weather markets THE WEATHER causes the price signatures. When the weather changes, it usually causes the price pattern prediction on the chart to be violated. 

The weather pattern change usually comes first. If it looks like the market reacted BEFORE the weather to most people, its almost always because the traders with the best weather information had that information before everybody else and they were getting in, moving the market. 

The early trading bird catches the money worms in weather markets!!!!

Getting in too early, trying to pick a bottom or top can cause a big drawdown temporarily if you are right, before it turns to a profit.

If you are wrong about the weather in a weather market.......you lose money. 

Too be honest, I was caught by surprise at how big the gap higher was tonight, just looking at the weather.

As Larry knows, the natural gas market, more than any others has big gaps higher and lower on the Sunday Night open. This is often because there were 2 days for the weather forecast to change and an additional 2 NEW days at the end of the forecast compared to the one Friday. Those NEW days are the ones that are most prone to have changes because of the increased uncertainty with time in a weather forecast. 

The symmetrical wedge(triangle) is the 3rd one down on the far right below.

Understanding lower highs and higherlows in trading: a comprehensive guide

https://itez.com/en/blog/education/lower-highs-higher-lows-trading-guide

        Natural Gas Futures Surging as Weather Models Shift Dramatically Colder     

                                        Daily Gas Price Index                            Markets Report                            Natural Gas Prices            

Having trended substantially colder over the weekend, weather forecasts were in the driver’s seat Monday morning as natural gas futures mounted substantial early gains.

+++++++++++++++

Record warmth in the middle of the country last week dominated the weather picture. 

However, there are not that many high population centers in the Plains but there are a few in the Midwest,  where the warmest air compared to average was located. The natural gas market dialed this in with lower prices EARLY in October WELL BEFORE it happened.

The East Coast was close to average. The West Coast was chilly. 

https://www.cpc.ncep.noaa.gov/products/tanal/temp_analyses.php

+++++++++++

This source does not show price gaps so I put in the red lines that define the close Friday and the open Sunday Night. Very large gap higher on the open with very solid follow thru that confirms the initial interpretation after the open last evening of it as an upside break away gap.

Coming from a solid test of the uptrend with higher lows last week,  this is extremely bullish.

https://www.cpc.ncep.noaa.gov/products/tanal/temp_analyses.php

This was the last 500 mb from the latest 12z European Ensemble model which show the colder pattern in early November evolving.

This is a ridge/west tough/east couplet which will features very warm temps in the West and chilly temps in the East. We don't yet have a connection in Canada that would bring down Arctic air or air masses from higher latitudes. Canada will be fairly mild at this time so air masses that cross the Canadian border south into the US will have a modest brand of cold. This is quite a bit different than the recent very mild pattern.

Heating demand  and HDDs with this scenario will be ABOVE average compared to MUCH BELOW average thru the first half of October. 

Expanding the view all the way north shows mostly mild zonal flow in Canada with the dip east in southeast Canada so the source region for these air masses will not be very cold.

However mild air for Canada, when plunging south will be chilly for the regions south, especially farthest south. 

The 0z European Ensemble model was +6.5 HDDs vs the previous run very early this morning which caused a huge spike up at that time.

1:50 pm:

This last 12z GEFS run in purple/left was the coldest yet(pattern changes tend to gradually get colder).

This helped NG close towards the top of the daily trading bar!

The teleconnection couplet today looks colder with the positive height anomaly greater over the higher latitudes in North/Central Canada(instead of the Western US where its modest) and the negative height anomaly in the Northeastern US.

https://www.psl.noaa.gov/map/images/ens/ens.html#nh

10-22-25: I note that the map above updates every day. That’s good for the maps on our weather page.

However, the solution today is different than the one described yesterday which will be different than tomorrow/Thursday, so please keep that in mind.

The extreme short covering bounce from the weather pattern change to much cooler over the weekend may have been dialed in and run its course for now as it bumps up to resistance.

OR, if the forecasts get colder then we could be having a major upside break out!

This was the last 12z EE in purple on the left:

https://tradingeconomics.com/commodity/natural-gas

1. 1 month: Note the upside gap higher to start the week on Sunday Night. We peaked today, well over +$5,500 from the close on Friday! Will this resistance hold?

2. 1 year: Symmetrical triangle with lower highs and higher lows(the last 2 months).

3. 10 years: Are we having an upside break out of this longer term symmetrical triangle/wedge??

As always, the weather forecasts CONTROL the technical pattern/formations. If we go back to unseasonably mild for a couple of weeks, the price will go lower. If the forecasts keep getting colder, we will go higher. And it will happen with wild gyrations that are often triggered from individual much colder or much warmer weather model runs.

These reactions usually start early in the model run as traders with that data first are buying at the market on cold or selling at the market on warm. It happens just like that 100+ times every Winter!

I get data extremely fast from a wonderful,  paid data service that I use 24 hours/day  but traders with in house models are sometimes even faster. Seconds faster when the price is spiking up or down means MISSING THE TRADE. 

And it can mean trading on a model run thru 7 days, when somebody else just got day 9 and has a slight edge of a  minute or less!  Larry understands how this works exactly! 

That describes the overnight session. The day session features the big traders that have the huge edge with in house model forecast data that can run up to an hour faster than the government global models. It's government data/model runs but they get the data and, I suspect run it on in house models using AI(been around for many years in the world of weather). 

See more below!

This article is pay walled!

March 11, 2025 at 9:30 AM CDT

Updated on March 11, 2025 at 10:28 PM CDT

Takeaways  by Bloomberg AI

When Alex Goldstein was just 7 years old, he says, he was drawn to extreme weather. Within 10 years, he convinced his father to take him storm chasing and hasn’t stopped since.

Now, at 35, he leads a small group of data scientists and meteorologists who help teams of traders at one of the world’s largest hedge funds position themselves in commodities markets.

+++++++++++++++++++++++

For Commodities Traders, Better S2S Forecasts Could be the Most Valuable Commodity of All

https://www.salientpredictions.com/blog/for-commodities-traders-better-s2s-forecasts-could-be-the-most-valuable-commodity-of-all

Hey MM,

 It seems like NG wx trading keeps getting more and more complicated, which is one reason I’ve yet to even try it again. I’ve been too busy, regardless, and don’t feel like getting back to staying up much of the night. But I keep an open mind for the future, mainly for the less unreliable late fall/winter period. If I felt I could do it successfully in the daytime, especially around the 12Z model releases, that might one day get me into it again if not too busy though probably with a smaller volume of contracts or perhaps mainly through going long options.

 I hope you’ve been doing well. Have you been having a decent level of success?

I did ok early in the Summer trading C and S but have been back on the sidelines since then. 

I had some small NG trades last Winter but only overnight. 

Unfortunately, during the later morning, the market appears to react an hour before the GEFS comes out, so there must be big traders with in house AI models that are way out ahead of everybody else. Its been that way for awhile.

They are not there overnight most of the time. 

Sometimes the early afternoon is trade-able off the 12z EE model. 

Mike said:

“Unfortunately, during the later morning, the market appears to react an hour before the GEFS comes out, so there must be big traders with in house AI models that are way out ahead of everybody else.”

———

Hey Mike,

This is still another reason to keep me away at least for now.

1. Are you sure they’re not reacting to the 12Z GFS since it comes out an hour before the GEFS? It sure used to react to it strongly, often even more than the 12Z GEFS when that goes in the same direction as the GFS.

2. How often has that reaction prior to the GEFS ended up being consistent with the actual GEFS HDD change? And if wrong, does it later go the other way as the GEFS is released? If so, that may mean even better  opportunities for those reacting as the actual 12Z GEFS is released.

3. What do you mean by in-house AI models and why should NG react to them if they’re not GEFS or Euro ens?

This is still another reason to keep me away at least for now.

Thanks, Larry.

1. I'm not sure of anything since I'm not one of their meteorologists. No, I don't think it's from the GFS because the reaction is not very consistent with the new 12z GFS's.

2. My "1 hour" is a rough estimate of the time frame and yes, it's almost always in the right direction as the GEFS and the magnitude of the move up or down, well before the 12z GEFS government model run comes out online is very consistent with the latest 12z GEFS. 

3.  In house AI models, unlike those of the GLOBAL models can use the raw data that the GLOBAL models start with and only run it for the United States and maybe Canada for instance and have solutions that come out much quicker. I speculated this for several years now, just based on many dozens+ obserations of the market reactions ~1 hour BEFORE the actual 12z GEFS comes out. Not just most of the time but almost all of the time big moves in 1 direction, correctly predicted what the 12z GEFS was going to show. 

Here's more for you:

Tech giants unleash AI on weather forecasts: are they any good?

https://www.bbc.com/weather/articles/cwy6ykp7049o

20 June 2025

Traditional models take hours to run on hugely expensive supercomputers, however this new breed of models can take less than a minute to run on a standard laptop. They don't need to know all the 'burdensome'  laws of physics, but are instead trained on 40 years of past data to make their predictions.

++++++++++++++++=

I know they have changed in recent years, being updated with better technology but this has been happening for over a decade in the mornings BEFORE the 12z GEFS comes out. I am 100% certain of this. What exactly they have been using(which has evolved) I can't say but it clearly is a product that uses the FRESH data from weather balloons and other sources to FRONT RUN the 12z GEFS.

This never happens at other time frames to the same extent.

At 0z, for instance you can see the reactions often being a couple of minutes faster than Weather Bell's products show the changes and often simultaneously which suggests we are just seeing government weather models that everybody gets having an impact. 

I have not traded natural gas in the last several months so things could have evolved, however, it's really easy to just look as a price chart later in the day or next day and see the spike and their times to observe the impact as long as you know what the GEFS and EE showed too.

I think the product being used to front run the 12z models only has 1 big impact BEFORE the models start coming out, including the GFS. And it's a gradual impact, over the course of an hour or so, suggesting the data might not all come out for the entire run in 1 blob or else, if it does than other traders, who know what the front runners are doing are following their lead. 

FYI the weather community has been using different forms of AI for almost 2 decades. 

I joke all the time with my wife about how lazy its made operational meteorologists at local NWS offices. They often just use the AI product for the worded forecast which breaks the forecast down hour by hour. In doing so, the verbiage often suggests impossible preciseness in the timing of forecasts.

I think I gave an example here earlier this year that I'll try to find.

++++++++++

Actually, the examples were sent to my wife/son via email earlier this year. I'll see if I can copy those here.