Unusual trading signals in individual stocks have historically been able to forecast a market that is about to fall. These warnings also provide a heads-up for opportunistic entry points. This is very valuable for those on the sidelines waiting for a pullback in this multi-year bull market. I wrote about this ultra-rare signal in late January of this year, which screamed that a drop was around the corner for stocks. The markets cracked within hours of that article. Because the signals are so rare, we pay close attention when they happen.
Today's discussion is a little different. I am going to show you how Mapsignals uses this data to forecast these rare events. The way our unusual trading signal works is how we apply many measurements to a stock's trade behavior on a given day. Simply put, we look for stocks with rising prices on outsized volumes. The same goes for stocks that are decreasing in price on large volumes. We use many moving averages – from short to medium term – along with other technical measures. We are searching specifically for big demand for stocks.
We then group all of these signals and turn the data into usable investing information. Sometimes, we get unusual stock trading signals that turn out to be noise. The key is strength in numbers; by looking at thousands of stocks, we try to see if groups of stocks are showing the same patterns at once. We ask ourselves questions like: "Is there too much exuberance, indicating that a sell-off is near? Or is there too much pessimism, suggesting that a rally is near?"
What do all of these signals look like in a given week? Below is a chart of all unusual trading signals by market cap for the week of Oct. 1 through Oct. 5, 2018. Notice all the red all the way to the right. There were more than double the number of sell signals vs. buy signals (563 sells vs. 216 buys).
Our usual observation is 1.5 to 1 in favor of buying on a given week going back through 30 years of data. So this week was out of the ordinary. The selling wasn't just apparent the first week of October – it began showing up the week prior (Sept. 24 through Sept. 28, 2018). By looking to the right, we can see that selling also eclipsed buying that week (329 sells vs. 225 buys).
But how can we use this information to give us an edge in trading and investing? Mapsignals created a ratio that tracks this buying and selling on a 25-day moving average. Since markets zig and zag, we want to smooth out the buying and selling over five weeks to show if an extreme pattern is emerging ... i.e., a drop could be near for the market.
In the chart below, we display the ratio as a 25-day moving average of our buying/selling signals, overlaid with the Russell 2000 with a range between 0% and 100%. When the ratio goes above 80% (green), we get overbought and typically expect a sell-off soon after. (This happened in late January 2018.) When it falls below 25% (red), we get oversold and usually see a massive rally shortly after. Historically, the ratio has been very accurate. On Friday, Oct. 5, 2018, the ratio dropped to below 50%, meaning that selling is larger than buying over the past five weeks ... this usually means that a market is about to fall. Look at Friday's closing data below:
The ratio falling below 45.6% shows us lower markets afterwards like clockwork. To show you how accurate and timely this ratio can be, look at this update we gave on Wednesday morning:
Currently, the MAP ratio sits at 45.6% and is at a level that has historical precedent: 45.6%. This morning, I went through our data to see what the forward returns were for the iShares Russell 2000 ETF (IWM) in each instance where the ratio descended and fell below 45.6%. Because we are at the 45.6% level, we should break below this level today. We do expect to see lots of selling signals once today comes to a close. But again, bottoms are made when selling exhausts itself.
Below are stats on the 45.6% level and why we feel it is important. We have had only eight instances in which the ratio has been in a downtrend and then fell below 45.6%. This is out of 1,579 trading days of data, so this is a very rare event! A few observations on market returns after breaching below the 45.6% level:
- The market traded lower in all instances.
- The average trading days until reaching the local bottom = 13. The shortest was three days, and the longest was 27.
- The average return of IWM following each of the eight instances until the local bottom = -5.29%.
The table below details the eight instances in which the ratio descended and fell below 45.6%:
Below are the historical instances highlighted by a red circle with yellow fill:
The Bottom Line
Our data suggests that a falling ratio reading below 45.6% should lead to further downside in the market. However, this isn't a bad thing because readings reaching oversold (25%) levels have been great chances to scoop up great stocks. Our long-term view is bullish on the market, and we believe that pullbacks like these are buying opportunities.
To learn more about Mapsignals' institutional signals, please visit our "About Us" page.
Disclosure: The author holds no position in IWM at the time of publication.