Many traders approach the market with the belief that good trading ideas must be complicated. Traders dive right in with arcane theories, advanced mathematical approaches, or highly subjective approaches that only an eye with natural talent and 10,000 hours of screen time can identify (take your pick).
The NAS Trading perspective is that simplicity is often the best approach. Simplicity can work for a simple reason: The vast majority of investors have an inordinate tendency to do the wrong thing at the wrong time. These emotionally driven capital flows show up in the form of market patterns.
It is common to hear that most investors have bad timing. What is the evidence for this assertion? I believe it can be found in the difference between time-weighted and money-weighted returns. If you are unfamiliar with this concept, check out this article.
Where does all the money from investors' bad "bad timing" go? If you can consider this question, you can understand the opportunity more clearly. The goal of active trading or investing is to be on the other side of the great mass of market participants that consistently make bad decisions. How do we identify these opportunities?
This is where research comes in. This study will use a similar methodology to what I used in my previous article on the VIX ETN (VXX) and VIX futures. If you have not read that article yet, it might be a good warm-up for this article.
Lets examine OHLC (Open, high, low, close) price dynamics in the SPDR S&P 500 ETF Trust (SPY). Our goal is to find some of the locations that the "wrong way investors" are dumping their money when they create negative dollar-weighted returns relative to the index. (Note: studies conducted using SPDR dividend-adjusted data 1/3/2005-9/19/2012)
The first idea we will look at in this series is extremely simple: Has the position of today's close within today's range historically effected the close-to-close return (today's close to tomorrow's close) in the SPDR S&P 500 ETF Trust ? Just to make sure we are on the same page, the following bar chart illustrates what I mean:
I have divided the above price bar into quartiles such that if I say, "Today price closed in the bottom 50% of the trading range" it would mean that price closed below the .50 level identified in the chart above. As you can see, in the chart above priced closed within the top 75% of the range. For those of you familiar with technical indicators, this is equivalent to the "%R" indicator measured over one day.
So, lets start by establishing a benchmark. What is the close-to-close return for all days within our test period?
The average daily close-to-close return in the SPDR S&P500 trust has been .0003, or three basis points (A basis point is 1% of 1%). about 55% of days have had a positive return, the average gain was about .008, and the average loss was -.0096.
Next, lets look at the forward close-to-close return when today's close is in the bottom 50% of today's range:
Looking at the second column, we see that the average close-to-close return has jumped up to .0015 (15 basis points), percent positive return is 57%, and both average win and average loss are both about .009. The average return has been significantly higher when today's close is in the bottom 50% of today's range.
The following chart is presented as "building block" material, not a trading system. We are seeking to do two things. First, We are seeking to gain an understanding of market dynamics that when refined can be capitalized upon. Second, it is a demonstration of how small advantages can add up over time to something very significant. Given these caveats, lets check out how this seemly small (3 vs. 15 basis points per day return) adds up over the test period using $100,000 as our starting capital (1/1/2005 - 9/19/2012):
Compounded since 2005, the difference is quite dramatic. The, "buy if today's close is in the bottom 50% of today's range" rule has more than doubled the total return. If you have felt fearful of buying when the market is closing in the lower half of its range, you have been putting yourself at a disadvantage. One interesting feature of the above chart is that the "buy if the market closes in the bottom 50% of today's range" was up in 2008. Starting 2008 equity was $143,098.16, and ending equity was $162,439.18, for a gain of about +13.5%, vs. the markets return of about -37%. However I want to once again make clear: The above is not in itself a trading system. It is simply a starting point that we will build from when analyzing day-to-day price action.
Next, let's examine the chart and data for when today's price closes in the top 50% of today's range:
Doesn't it feel good to buy on days when the market is closing well, and "everyone" is happy? Well, for traders and investors who have felt this way, the joke has been on them. Close-to-close returns have on average been negative after the market closes in the top 50% of its trading range. The results hit the floor in 2008 and, amazingly, never rebounded during the 2009-2012 bull market.
Lets divide the data into quintiles and see how it looks:
Starting at the left and working right, you will see the basis point return for all days, the return when the close is in the bottom 20%, the return when the close is in the 20% to 40% quintile, etc. I have added some additional statistics in the rows below.
It looks like the most significant result has occurred when the market closes within the bottom 20% of the day's range. Does a bottom 20% close suggest selling has on average become overly emotional? I note that when I consider this data, I can easily recall many times when the market closed lower and then continued falling the next day. Buying at those points does feel more psychologically risky than buying after a solid bullish day that closes in the top of the day's trading range. Yet, the data shows that on average, a bottom 20% close today has led to the best performance over the next 24 hours, and a top 20% close has led to the worst return over the next 24 hours.
Does it look like chasing the current day's performance is one place wrong way traders have been dumping their money over the past seven years? It sure looks that way to me. The ideas examined here are building blocks. It is also important to understand that even persistent patterns change or degrade over time. Consequently, it is important to keep pattern studies updated as new data comes in, and to be continually re-evaluating and improving.
In coming articles we will be examining other "simple" factors that provide insight into market dynamics and can be used to create a trading edge. We will also look at combining price-based signals with your fundamental/value based investing ideas.