Shorting UVXY Wins 100% of the Time. But your account will go to Zero

There is a trade that looks like the closest thing to free money in all of markets: shorting UVXY. The fund is built to lose. It tracks one and a half times the daily move of short-term VIX futures, and between the leverage and the cost of rolling those futures it bleeds almost without pause. Since it launched in 2011 it has fallen effectively 100%, ten thousand dollars down to nothing, and along the way it has printed a new all-time low 748 times, roughly once a week. So the idea writes itself: wait for it to spike, short it, and collect the decay. We backtested exactly that, shorting every time UVXY rose 50% above its most recent all-time low, across the entire history. The win rate was 100%. The account still went to zero.

On paper, it never loses

Because UVXY always grinds back to a new low, a short held until that new low is all but guaranteed to win. The numbers look flawless. Every one of the 27 shorts triggered at a 50% spike since 2011 was eventually profitable, a median gain of 38% in a median of 23 days. Shorting at bigger spikes paid even more. And most of the time it is calm: half of those 50% shorts never went more than 4% against you before the decay took over.

Shorting each spike and holding until UVXY prints its next all-time low, 2011 to 2026. Held that way, every short eventually won. The last two columns are the warning.

Short trigger	Trades	Eventually won	Median gain	Typical worst point	Deepest worst point
+25% spike	46	100%	+27%	-10%	-716%
+50% spike	27	100%	+38%	-4%	-517%
+100% spike	7	100%	+57%	-22%	-410%
+200% spike	3	100%	+69%	0%	-231%

This is the surface that makes the trade so seductive, and so dangerous. A 100% win rate and a median trade that barely moves against you is exactly the kind of record that talks a person into betting bigger. The damage is hiding in that last column.

The path runs through a steamroller

The catch is the word "eventually." Between your short and that guaranteed win sits whatever the spike does next, and VIX spikes are not bounded. One short in nine at the 50% trigger was drawn more than 100% underwater before the decay arrived, a complete wipeout if you had committed your capital. The worst of them, a 50% short into late February 2020, watched UVXY run another 617% higher. The position was down 517%, more than five times the money put up, before it finally decayed back to a 41% win two hundred days later. No account survives a 517% loss to collect that win.

The worst draw-downs a 50% short suffered before the decay rescued it. Every one of these eventually turned a profit. Surviving to collect it was the problem.

You shorted a 50% spike in...	Loss at the worst point	Eventual gain	Days to the win
Feb 2020 (COVID)	-517%	+41%	200
Aug 2015 (China devaluation)	-109%	+43%	143
Feb 2018 (Volmageddon)	-103%	+38%	127
Mar 2025	-72%	+38%	83
Aug 2024 (yen carry unwind)	-58%	+46%	69
Oct 2018	-49%	+40%	124

A real account does not get to wait

That 100% win rate quietly assumes something no broker allows: that you can sit through any drawdown with unlimited capital. In reality a naked short that goes 100% against you is a margin call, and you are bought back in near the top of the spike, locking in the very loss the backtest says you would have recovered from. To see what that does to real money, hold the short for a fixed month instead of waiting for the low. The median is still a tidy 28%, and 78% of months come out green, but the worst single month lost 279%. Stretch it to three months and patience smooths the typical result to a 45% gain with 88% green, while the worst case still nearly wipes you out at -93%. The middle is lovely. The tail is fatal.

Your account, simulated

Put it together. Start with ten thousand dollars, short a slice of it at every 50% spike, pocket the decay, and repeat. The ending is the same no matter how careful you think you are being.

Compounding the strategy from $10,000, covering each short at the next all-time low, with the account marked ruined when one trade's worst drawdown met the whole balance (where a broker margin-calls you out).

You short this much of your account at each spike	What happens
100% of equity	Wiped to $0 on the 12th trade (Aug 2015)
50% of equity	Wiped to $0 on the 19th trade (Feb 2020)
25% of equity	Wiped to $0 on the 19th trade (Feb 2020)

Even committing just a quarter of your account to each short, you string together eighteen winners, watch the balance climb, and then one ordinary-looking spike in February 2020 erases all of it. To have survived that single day you would have had to keep each short under roughly a fifth of your account, and even then a future spike larger than any yet seen, and UVXY has run as much as 946% off a low, would eventually finish the job. At sizes small enough to survive, the profits are too small to matter. It is the same trap we found shorting BOIL every winter: a strategy defined not by the many times it wins but by the one time it does not.

How we tested it

• Data. Split-adjusted daily UVXY prices from its October 2011 inception to today, about 3,700 trading days.
• The spike. "Spike %" is the price above the running all-time low, the same definition used by our Spike Ladder and Spike Analyzer tools. A trade is the first day a fresh cycle crosses the trigger (+25, 50, 100, or 200%).
• Hold to the low. The main test covers each short at the next new all-time low, which is always below the entry, so the trade is a paper win by construction. That exit assumes unlimited capital, the exact assumption the article stress-tests.
• Worst point. For each trade we measure the max adverse excursion, how far the short went against you before it was covered. This is where margin calls happen.
• Fixed horizons. A cleaner read on real P&L: cover after 21 and 63 trading days instead of waiting for the low.
• The account. Compound from $10,000, shorting a fixed fraction of equity at each fresh 50% spike, and mark the account ruined when one trade's worst drawdown met or exceeded the whole balance. UVXY tracks 1.5x short-term VIX futures (it was 2x before February 2018), which only deepens the spikes in the earlier years.