Why We Don't Divide Decay by 252

Every leveraged ETF decays. The question we get asked most is how this tool projects the rate, and the short answer is that it doesn't use a rate at all.

The lazy version of decay projection is to take an average annual decay figure and divide it by 252 trading days. Anyone can do that, and it's misleading, because it treats every day as identical.

When UVXY is at a 50% spike it behaves nothing like UVXY sitting at all-time lows. One flat decay rate can't describe both.

Here's the size of the difference.

Sitting on the low, UVXY loses a median 2.57% over the next week. Stretched 50% or more above the low, it loses 7.72%. Same 5 sessions, three times the damage.

If the spike level triples the loss over the week ahead, the projection has to start from the spike level.

So the tool looks up every past day the ETF traded at the same percentage above its all-time low as it does today.

That's where the price cone comes from. The downward slope in it is UVXY's own history.

Here's how it works, in two steps.

Step 1: Finding the matches

Say UVXY is sitting 62% above its all-time low. The tool goes back through the entire price record looking for days when it first hit 62% above whatever the all-time low was at the time, not today's low.

Only one match per all-time low. Once a low has produced a 62% day, it's done, and the next 62% match has to wait for a new low.

Those dates become the sample size.

Step 2: Replaying what happened next

From the close on each of those days, the tool tracks where price went over the following weeks.

Stack all those paths on top of one another and you get the cone. It shows how UVXY has actually behaved after spikes of this size.