Schaeffer's Trading Floor Blog

Trading 101: Getting the Call Right Is Only Half the Battle

Monday-morning-quarterbacking a recent volatility trade

by 5/15/2014 7:32 AM
Stocks quoted in this article:

I don't trade actively anymore, but one play I do like to put on from time to time is to fade something VIX-related when the CBOE Volatility Index (VIX) gets overbought.

I put on just such a trade back in January, and literally just took off the balance of it yesterday. It was a "winner," but in hindsight it was possibly one of the worst "good" trades I've ever done. And that matters. Let me try to explain.

Basically, here's what I did: When VIX went overbought way back in January, I bought some VelocityShares Daily Inverse VIX Short-Term ETN (XIV), then I bought a little more. A few days later, I sold half back, somewhere near where I bought the first half. I then sat with the balance of the XIV and waited for the inevitable iPath S&P 500 VIX Short-Term Futures ETN (VXX) implosion. That of course happened eventually … a mere three months later.

It worked and it was terrible all the same.

I correctly thought the VIX rally was a blip, and wanted to fade it. I'm a broken record on that; I always expect mean reversion. I win that trade most of the time, but occasionally get caught fading VIX way too early. So details matter -- I need to both max out the winners and defend the losers well enough. You could really say that about any play you put on, of course.

Anyway, that's what I did right. But I made two big mistakes.

No. 1: I went long XIV instead of just going short VXX. XIV is a tracker for the inverse of XIV. It captures negative 1x the one-day move of VXX. In the short term, going long XIV will essentially return the same as going short VXX. Over the longer term, results may vary.

Which leads to mistake No. 2: Making a choice that's indifferent over the short term, but more variable (probably in a bad way) over the longer term … and then holding over the longer term.

Here's how VXX looks in 2014. (Click chart to enlarge.)

Chart of VXX in 2014
Chart courtesy of TD Ameritrade

And here's XIV. (Click chart to enlarge.)

Chart of XIV in 2014
Chart courtesy of TD Ameritrade

Visually, they look like exact inverses. However, when you use actual numbers, they're not quite. As a tracker, XIV compounds the day-to-day results. Compounding adds "value" in a one-directional move, but subtracts value in a churning move. And while VXX ultimately drifted over time, it churned an awful lot on the way to inevitably get there.

Let's say on Feb. 7, you made the decision that the Volatility Fever had broken, and you wanted to short VXX until it made a new all-time low (setting aside that you really shouldn't use price targets in VXX itself). Well, you would have held it until May 2 and earned about 12%.

But what if instead of shorting VXX, you went long XIV? You would have gained 8.5%. Yes, that's a nice win, but is significantly worse than if you shorted VXX.

Time-wise, that might have made sense if you had to pay to borrow VXX. But that's not the case anymore, as VXX is easy to borrow.

You may say "win's a win," but that inherently assumes that you always win. You absolutely don't. You truly need to maximize winners, because that next play, or the one after that, might turn into a disaster. Getting the call right is only half the battle. Efficient use of capital and getting the most you can out of every play is the other half, and to win over the long haul, you need to master that too.

Disclaimer: Mr. Warner's opinions expressed above do not necessarily represent the views of Schaeffer's Investment Research.


permanent link
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT

Partner Center

© 2014 Schaeffer's Investment Research, Inc. 5151 Pfeiffer Road, Suite 250, Cincinnati, Ohio 45242 Phone: (800) 448-2080 FAX: (513) 589-3810 Int'l Callers: (513) 589-3800 Email: service@sir-inc.com

All Rights Reserved. Unauthorized reproduction of any SIR publication is strictly prohibited.

Market Data provided by QuoteMedia.com | Data delayed 15-20 minutes unless otherwise indicated.