Is Shorting BOIL Every Winter Really Free Money?

The pitch is seductive, and you have probably heard it. BOIL is a 2x leveraged natural gas ETF, and leveraged ETFs bleed. Natural gas demand peaks in winter, and the seasonality, the story goes, makes winter the worst stretch of the year for it. So: short BOIL every December, cover in spring, collect a year's worth of decay in three months, repeat forever. We ran fourteen winters of split-adjusted data to find out whether it is as easy as it sounds. The short answer: the edge is real, and the word "easy" is a lie.

First, BOIL really is a melting ice cube

Start with the thing that makes the whole idea plausible. Since it launched in 2011, BOIL has lost essentially 100% of its value on a split-adjusted basis. Reverse split after reverse split, it has ground toward zero. Two times daily leverage on one of the most volatile commodities there is, plus the cost of rolling natural gas futures, is a recipe for relentless decay. Shorting it at a random moment and holding three months made money 69% of the time, and shorting it for a full year averaged +46% (median +60%), profitable in 11 of 14 years. So shorting BOIL is, on average, like shorting an ice cube on a warm day. That part is true any time of year.

Winter is genuinely the worst season for it

The seasonal claim holds up too. Sorted by calendar month, BOIL's two worst months by a wide margin are the winter bookends, when the cold-weather premium that built up in the fall deflates.

Average BOIL return by calendar month since 2011, and what a short earned (the negative of it). December and February stand out; April and August are the months to avoid shorting.

Month	Average BOIL return	Shorting it earned
December	-19%	+19%
February	-12%	+12%
October	-5%	+5%
April	+5%	-5%
August	+5%	-5%

December and February alone gave back nearly a third of BOIL's value on average. The winter short is not arbitrary, it lands right on the two months where the decay is fiercest. (Note that the deep-winter month, January, is the weak link at roughly flat, because January is when cold snaps actually hit. More on that in a moment.)

So how good is the winter short? On the median, very

Shorting BOIL over calendar winter (December through February) each year, 2012 to 2026, against two benchmarks: a short started on a random day and held three months, and a short held the entire year.

	Dec-Feb winter short	A random 3-month short	Short BOIL all year
Average return	+31%	+10%	+46%
Median return	+49%	+17%	+60%
Win rate	87%	69%	79%

The Dec-Feb short won 13 of the last 15 winters, with a median return of +49% in three months. That is roughly a full year's worth of the year-round decay, earned in one quarter, and about three times what you would have made shorting a random quarter. On the median, "a winter's short is a year's profit" is basically true, and the seasonality is real, not imagined. If the story ended here it would be free money. It does not end here.

The catch: the one winter in seven that ends the game

Look back at that table. The average winter return (+31%) sits well below the median (+49%). That gap is the whole story, and it is hiding in two winters that were not small losses but catastrophes. Here is every winter, as a short P&L:

The two red winters are not bad luck, they are the trade's actual risk showing up. 2014 was the polar vortex. 2025 was a hard cold snap. Both sent natural gas vertical, and BOIL, at two times leverage, doubled the move and rocketed up, handing the short a 45% and a 70% loss. Widen the window slightly to November through March and winter 2025 lost 130%, more than the entire stake. You are, after all, shorting a leveraged long-volatility product into the exact season it is built to explode. The same leverage that melts BOIL in a mild winter detonates it in a cold one.

And a short that loses more than 100% is not a drawdown you recover from, it is a funeral. We compounded the strategy, reinvesting each winter's profit into the next. The disciplined Dec-Feb version turned 1 dollar into about 17 over fourteen winters, which sounds wonderful. But the slightly greedier November-March version, riding one extra month at each end, blew the account to zero in winter 2025, when a single winter lost more than everything in it. There is no fifteenth winter to bounce back in. One cold winter is terminal, and you cannot know in advance which winter is cold.

How we tested it

• Data. Daily, split-adjusted BOIL prices from its 2011 launch to today, about fourteen and a half years. Split adjustment matters enormously here: BOIL has done many reverse splits, and the raw price would be meaningless.
• The trade. A short held over a fixed window each year, with profit-and-loss measured as the negative of BOIL's return over the window (short gains when BOIL falls, loses when it rises).
• Seasonality. The average BOIL return for each calendar month, to locate where the decay actually concentrates.
• The windows and the benchmarks. Several winter spans (December-February, November-March, and others) measured against a short started on a random day and held three months, and a short held the full year.
• The survival test. Compounding a fresh short each winter, reinvesting the profit, to see whether a single bad winter ends the account.
• One real-world caveat the gross numbers ignore: BOIL is frequently hard to borrow, with steep and variable borrow fees, and a borrow can be recalled at the worst moment. Real returns run below these, and in a squeeze you may not be able to hold the position at all.