Wednesday’s close found $VIX above 20 for the first time in several months. What has happened in the past when $VIX is rising and crossed above 20?
Rules:
- When $VIX crosses above 20, buy SPY at the close
- Sell X days later
- No commissions or slippage included
- All SPY history used
Results:
Summary of Results:
I included 2 other tests with added variables for comparison:
- I added the requirement that $VIX be higher than its 50 day moving average
- I used $VIX>20 AND $VIX >MA50 AND SPY < bottom Bollinger Band
- By itself, $VIX crossing above 20 presents a rather bearish setup. There were 33 occurrences.
- However, when we require $VIX to be above its 50 day moving average, results are more neutral to bullish. There were 23 occurrences of $VIX>20 and higher than its 50 day moving average. I have written about this setup before here.
- Requiring SPY to have closed beneath its lower Bollinger Band did not change the results much, and there were only 6 occurrences.
Bottom Line: Recent market action is still suggesting more downside or consolidation over the next 20 days.
I should be more clear. There were 111 occurrences of the basic VIX above 20 setup-33 if each trade is held for 50 days.
There were 63 occurrences of the basic setup plus VIX above its MA50- 23 if the trade is held for the full 50 days.
Have to say, fan of all things Woodshedder.
Wanted to ask if you’ve come across any reliable system that uses the VIX to provide trading signals. I am a breakout trader and am always looking for a filter to cut down on the number of signals I get. I know that every second sentence by Pete Najarian on Fast Money starts with the phrase Volatility Index, but I’ve never found it to be anything other than a coincident indicator
ie, the market goes up, the VIX goes down and vice versa. As for the level, back in ’08 it was in the 80s and before that spent months, if not years in the low teens.
As always, your thoughts and insights are appreciated.
beano, can you share your preferred breakout setup? If it is something I can quantify, I can run it against the VIX to see if anything jumps out at us.
You are right in that the VIX can stay in a range for a long time.
I think what has to happen with volatility is that it has to be ranked, meaning that you have to rank the current level against levels in the recent past. For example, you would rank today’s VIX level against all VIX level for the past 252 days. I believe that is the way Jeff Pietsch does it over at ETF Prophet http://etfprophet.com/
Woodshedder, thank you for the quick response. I trade two systems, a ten day B/O with a trailing 3 day low stop, and a longer term 55 day breakout with a 20 day low stop. I only trade S&P 500 stocks and S&P TSX 60 stocks; I short occasionally (not consistently enough) and my one filter is that the volume > 30 day average volume on the day of the breakout.
The one VIX filter I’ve tried to incorporate is Larry Connors’ VIX 5% Rule, where you buy only if the VIX is 5% above its ten day average. I only see noise when I try to incorporate this, so I dropped it.
Thanks for the insights; your systematic trading ideas and quantifications are unique in the blogosphere.
Beano, those BOs should be easy enough to code. I can run the tests over the S&P500 current constituents. I do not have a survivor-free S&P500 database. I can however run the tests over my de-listed stocks database, and using a volume and price filter, probably get something close to stocks that could be included in the S&P500.
A little more information will be helpful. Are you buying the open or the close? Market or limit orders? How are your stops triggered? I’ll try and set it up to mimic as close as possible how you are trading.
I love these studies. Thanks for doing them.
Hey AB! Thanks!
Woodshedder, beano here. I buy the open the next day and take whatever fill I get. Then when I get home from work at night I place the stop at the lowest low of the past 20 days ( or 3 days if it’s a 10 / 3 trade). As you can imagine, gaps are a bitch with this type of trading, but I don’t have access to the markets throughout the day. Also, with my five digit piker account, it’s white knuckle time when I’m getting buy signals on stocks like AAPL, GOOG, IBM etc since I have to buy 100 share lots in order to place automated stops.
Beano, cool, that will be easy to mimic.
One more question- does your stop ever get altered? Meaning, do you adjust if say 30 days pass and a new 20 day lowest low is set? And, by lowest low, you do mean the lowest intra-day low, and not the lowest close, correct?
What is your commission structure?
Shoot, one more thing…if you don’t get stopped out, when do you sell?
Sheesh Woodshedder, I didn’t expect this much effort or detail on your part but it certainly is appreciated.
1) stop never gets altered. I will eventually sell at the 20 day intra day low
2) $4.95 a side
3) If I don’t get stopped out when do I sell? Never!!! (this never seems to happen though)
I have read literally hundreds of articles telling me when / how to buy stocks, but I’ve never came across one that tells me when to sell. I wrestle with this all the time, but I just don’t think I trade in big enough size to scale out. Also, commissions are an issue.
Have to say that for my own sanity, this needs to get rectified. If I buy a stock at $20 and it goes straight to $15 and I get stopped out it doesn’t bother me one bit – that’s trading. However, if I buy a stock at $20 and it goes to $30 and then traces back to $25 and my stop gets hit, it drives me nuts.
The only thing that would be more frustrating would be if I sold at $30 and it went to $50 – hence my dilemma.
Anyway, I digress – thanks for the effort and interest.
@Beano – have you ever used/played around with Chadelier stops? Various places talk about them – essentially using a trailing stop based off of the highest high/close reached during the u/t – particularly useful when a market goes parabolic or has a giant outside candle that traps subsequent candles. Typically people use some sort of ATR factor to determine how far of a pullback they should all – like 14 ema or 2 x that or however much you feel comfortable with.
And yes – this is a very hard part of trading – when to get out
Cheesetrader,
thanks for the suggestion, I will look into this. The only other type of stop I’ve tried is a trailing 3 ATR stop from the close. It seemed to lead to scratch after scratch so I went back to my original.
cheers
Cheesetrader, as per my comment below, I think we are talking about the same methodology, since my first google search mentioned Van Tharp’s book which is where I heard of the methodology. Thanks again.
It referenced testing the stop method on a futures portfolio. One question I’d like to throw out there to all systems traders is that if you’ve been following my posts above with Mr. Woodshedder, you probably recognize my methodology as the Turtle Method. As a general question, do trading systems developed in the futures arena work for trading stocks?????