Trading on Natural Gas: Analysis of Weekly Bias and its Persistence over Time in

Identifying a market bias, i.e., a recurring behavior in price trends, is one of the most straightforward approaches when trading commodities. This article focuses on Natural Gas futures (Henry Hub Natural Gas – NG) to check whether a weekly bias we analyzed more than a year ago, thanks to software developed by Unger Academy, i.e., the Bias Finder, is still present and exploitable. The events of 2022, particularly the Ukraine conflict, have strongly influenced commodity prices, so the question arises: Is this bias still working?

Let’s first take a look at what bias we are talking about. Using the Bias Finder, specifically designed to identify recurring market behaviors, we loaded the Natural Gas futures historical data from January 1, 2008, through December 31, 2021, with a 15-minute time frame.

We then asked the software to graph the average weekly price trend for 3 different periods: from 2010 to 2013, from 2014 to 2017, and from 2018 to 2021, excluding the data from 2008 to 2009 because they are the “dirtiest”, as at that time these futures were still mostly traded at the trading pits and therefore fundamentally different from what they are today.

Let’s look at the results of the average weekly trend. It is immediately noticeable that in the Thursday session, which starts on Wednesday at 6 pm and ends on Thursday at 5 pm, we see a reasonably recurring behavior: all 3 periods (i.e., the blue 2010-13 equity line, the orange 2014-17 equity line, and the yellow 2018-21 equity line) show significant declines.

Figure 1 – Average weekly Natural Gas (NG) price trend until 2021.

This bias was certainly most pronounced from 2010 to 2013, but we also saw some decent declines between 2018 and 2021. Between 2014 and 2017, on the other hand, this bias did not disappear entirely but somewhat faded, perhaps indicating that volatility in the market decreased during these years.

This recurring behavior can be linked to the publication of data on natural gas stocks, which occurs precisely every Thursday around 4-4:30 pm (Italian time). This is the change in the number of cubic feet of natural gas stored in underground storage facilities in the previous week: whether demand rises or falls directly impacts price trends. Let’s look in detail at the Bias Finder chart in the Thursday session. We can see precisely that the downward movement is triggered between 6 am and 12 pm (New York time), which perfectly coincides with the timing of the stock data release (4 pm Italian time is 10 am New York time).

But what happened from the beginning of 2022 until today?

As mentioned above, the market has experienced strong upward shocks due to geopolitical events and the energy crisis in general, and we also saw this by looking at the bills at the end of the month. Let’s ask the Bias Finder, data in hand, to plot (in purple) the period from 2022 to May 2023. We can instantly see how volatility has exploded, with average weekly price changes more significant than in previous years.

Figure 2 – Average weekly Natural Gas (NG) price trend until May 2023

If we look at Thursday in particular, we can see how the previously very clearly and precisely identified bias has shifted somewhat. The average trend of the day remains bearish; however, this movement is no longer concentrated around the time stock data is released. Instead, it is more spread over the beginning and the end of the day, with even an upward bias at the time of day that used to trigger the declines.

To test what we have found out, thanks to the Bias Finder, we can try to develop a simple strategy that exploits the bias identified up to 2021 and then check how it would have behaved in 2022 and up to the present. Using a 15-minute time frame and a fixed stop loss at $1,200, we enter short on Thursday between 6 am and 12 pm (the time frame we defined above) when some sort of breakout confirmation occurs, i.e., the break of the lows of the last 4 periods (i.e., the previous hour). We then close the positions…