Double 7’s Strategy

In the book, Short Term Trading Strategies that Work, which Larry Connors and I published in early 2008, we wrote about a simple strategy called “Double 7’s Strategy.” Through the years people often ask about this strategy. Does something that simple really work? How does it do in a portfolio? Does the concept work on stocks? Today, we will be answering these questions.

The Original Rules

Buy

  • Close is above 200-day moving average
  • Close is a 7 day low of closes
  • Buy on Close

Sell

  • Close is a 7 day high of closes
  • Sell on close

A clarification on what it means to be a “7 day low of closes.” There are two ways one can understand this.

Method 1. Today’s close is less than or equal to the previous 6 day closes. The AmiBroker code for this is “Close<= LLV(Close, 7).”

Method 2. Today’s close is less than the previous 6 day closes. The AmiBroker code for this is “Close < Ref(LLV(Close, 6), -1).”

As we originally tested and I will test today is method 1. In the spreadsheet, I also include the results of Method 2 when tested on stocks. Short story is there is little difference.

Changes from the original test. I don’t like testing with execution at the close. I will then change that to execute at the next open. The original test was a points test. These tests will be a portfolio test.

New Rules

Buy

  • Close is above 200-day moving average
  • Close is a 7 day low of closes. Method 1.
  • Buy on next Open using 100% of cash

Sell

  • Close is a 7 day high of closes
  • Sell on next open

 

Results on SPY & QQQ (2000 to 2007)

First we will see how the rules did during the time period that we originally tested.

160210a

The CAR numbers are not ‘go open up a fund’ numbers. But they beat buy and hold with only ~26% exposure and substantially better drawdown. But how has it done since then? Has this simple mean reversion edge shrunk or disappeared?

Results on SPY & QQQ (2008 to 2015)

160210b

The Double 7’s Strategy has not held up as well since 2008. Both CARs now are less than buy and hold. The SPY’s drawdown is much lower but the QQQ’s is just as bad. I got curious about this and it all comes from one bad trade in 2008.

What about stocks?

I wanted to know how this has done on stocks. Could something so simple work on stocks?

Stock Rules

Buy

  • Stock is member of index (tested both SPX and NDX)
  • The Index close is above the index 200 day moving average (added because I strongly believe in these. Especially for mean reversion strategies)
  • Close is above 200 day moving average
  • Close is a 7 day low of closes
  • Buy on next Open using.
  • 10% per position for a maximum of 10 positions

Sell

  • Close is a 7 day high of closes
  • Sell on next open

Stock Results

160210c

The returns are only a little better than buy and hold. The drawdowns about one-third better. Not great numbers. What I found interesting was the results the last couple of years. It looks like 2013 and 2014 where great mean reversion years while 2015 was not.

Spreadsheet

Fill the form below to get the spreadsheet with lots of additional information. This includes yearly breakdown, worst 5 drawdowns and additional statistics.

Final Thoughts

There is ‘simple’ and then there is ‘too simple.’ The Double 7’s Strategy may have worked in the past because mean reversion trading was not as popular. Now the strategy may be too simple so there is no to little edge there anymore.

Good Quant Trading,

Fill in for free spreadsheet:

spreadsheeticon

Click Here to Leave a Comment Below

Nick Wienholt - February 10, 2016 Reply

Hi Cesar – it’s very interesting that 2015 was a poor year for this strategy. I’m trading a slightly modified version of Three Day Lows and it hit the ball out of the park in 2015 – 45.5% return. A lot of that was in the first two months of the year, and came from extending the strategy out to the full universe of ETFs with sufficient liquidity. Results at https://www.dropbox.com/sh/1hvy634vd4lnayk/AAAaMMwpNCoNFoyVmqOtxfQ-a?dl=0

    Cesar Alvarez - February 10, 2016 Reply

    Nick, this probably has to do with the choice of ETFs to trade. It looks like you have both long and short versions and are doing a scaling in. This does show that the concept of MR is working but it depends on the strategy.

Leave a Reply: