In the Weblog, I recently noted that the average trading price for the ES futures has been in a narrow range over the past three trading sessions. Specifically, we've had average prices of 1419, 1418.5, and 1420. When we see successive average trading prices that are so similar, this suggests that we're in a trading range, in which prices are oscillating around their average. Such markets can be quite profitable for short-term trader who can identify the range and average price as they are setting up. I recently illustrated the notion of framing trade ideas as "returns to the average price" in a morning session tracking the market.
So what tends to happen after a three-day trading range with a tight array of average prices? Going back to 2004 (N = 755 trading days), I calculated the average trading price for each day in the S&P 500 Index (SPY) simply as the average of the day's open, high, low, and close prices. I then calculated the three-day high-low range of these average prices.
When we've had a narrow range of average prices of less than .30% (as at present, indicating a three-day trading range; N = 144), the next five days in SPY have averaged a loss of -.02% (76 up, 68 down). By contrast, when we've had a wide range of average prices of greater than 1.0% (N = 151), the next five days in SPY have averaged a gain of .24% (90 up, 61 down). For the entire sample, the average five-day change in SPY has been .16% (431 up, 324 down).
It thus appears that returns have been subnormal following a three-day trading range. Such a range represents a short-term consensus regarding value in the market. When we have widely spread average prices, such consensus is missing. In general, markets tend to reward holding stocks through uncertainty, not through consensus.