Wouldn’t we all like to know that today is the bottom of a sharp sell-off, so we can buy at just the right time? Some would say that’s trying to catch a falling knife, and certainly there is significant risk in a volatile market that’s been collapsing. But then risk is relative to reward, and if the reward is big enough, the risk can be worth it.
“A short-term strategy based on volume” (January 2016) discusses using volume spikes to identify turning points. That’s a classic concept. Now we will take two simple concepts: Annualized volatility and the stochastic indicator, to determine when to buy a bottom. Understand that nothing is foolproof, but this combines two basic ideas and has all the indications of working across a wide selection of index markets.
There are only three calculations needed for this method:
1. Annualized volatility taken over the past 20 days, the same calculation period used for options volatility:
AV = standard deviation (returns, 20) x square root (252)
(Be sure that you use the returns, not the price. Returns are: Close(today)/Close(yesterday) – 1.)
2. Stochastic (an unsmoothed momentum calculation):
100*(Close(today) - Lowest(low,20))/
(Highest(high,20) - Lowest(low,20))
This indicator gives you the position of today’s close within the high-low range of the past 20 days. It has values from 0 to 100.
3. A 100-day moving average to determine that prices are moving down.
We’re looking for a sharp sell-off to be a buyer. That’s easily defined as a combination of downtrend, high volatility, and a low stochastic to indicate good timing. The entry will be mostly based on the volatility because the stochastic will probably be showing low values all the way down.
Once in a trade, we will exit if the volatility declines back to a normal level, or the stochastic rises reflecting a rally off the bottom. We don’t want to be too demanding of the exit values because high volatility is also high risk. There is no stop-loss, so we want to be sure to get out.
This strategy needs three conditions to be met for an entry signal:
- The closing price must be below the 100-day moving average to ensure a downtrend.
- The annualized volatility must be greater than a threshold value to indicate high volatility and it must be rising.
- The stochastic must be below a threshold value to be oversold.
Exit the trade if either of the following conditions occurs.
- The annualized volatility falls below a threshold value. Declining volatility often indicates the end of a move.
- The stochastic rises above a threshold value, indicating the price is no longer oversold.
“Catching bottoms” (below) shows the sector SPDR SPY at the top, along with the 100-day moving average. The stochastic is in the middle, and the annualized volatility in the bottom panel. The buy and sell signals are also shown at the top. The four entries occur when the annualized volatility is above 0.19 and the stochastic is below 15.
Applying techniques using volume and volatility to an index rather than an individual stock often works better. It shows that the event we are tracking is broad based and not a spike associated with single stock news, such as a scandal or an earnings surprise.
Whenever possible, we want to generalize the parameters; if we can apply the same values to many markets we consider the solution robust. For this method, we’ll always use:
- Moving average of 100
- Stochastic period of 14
- Annualized volatility period of 20
- Stochastic oversold entry threshold of 15
- Stochastic exit threshold of 60
- Annualized volatility exit threshold of 5% (0.05)
The only value that changes will be the volatility entry level because volatility can vary for each market. The volatility exit level won’t be important because most trades will exit when the stochastic rallies above 60; it responds much faster than the volatility.
“Broad profits” (below) gives a summary of selected ETF results. It is sorted by volatility entry threshold, highest to lowest. In general, the higher the threshold, the fewer trades. The Profit Factor is the gross profits divided by the gross losses, a measure of reward to risk. The annualized rate of return (AROR) is low because the time in the market is small. But then your exposure to risk is also small. The far right column shows what the returns would be if this method would have been in the market 100% of the time.
It is important to visualize the pattern of results, shown in “Profit picture” (below) for SPY, IWM, and XLE. SPY is far less active than either IWM or XLE and they all have a volatile period during 2008, even though the loss was recovered in very few days within a single trade.
Could it be better?
Of course we could optimize these rules, even find specific values for each parameter for each market. We might be able to remove the loss seen in 2008 in the equity chart. But then we would be fine-turning this to the past history of those markets, a method that has never turned out to be rewarding. The future just doesn’t quite follow the patterns of the past, and no one market contains enough patterns to give us a robust solution for the unseen future. By finding one set of parameters that works across all markets, we have essentially used more data to arrive at one solution. The results are not as good as optimizing, but they are more realistic.