Subject: NH-NL Compared
Back in July 2008 I first introduced the Mechanical Investing community to the NH-NL indicator that could be used to manage market risk. There are a number of ways to use this, but the method I originally presented subtracts the new lows from the new highs in the Nasdaq Composite daily, and then applies a nine day weighted moving average on the result. The original post is here.
http://www.datahelper.com/mi/s...
An alternative method is to use a thirteen day exponential moving average on the result. In this post I'll compare the two for the last few years using the QLD / QID as the investment vehicle. During BULL signals QLD is held, and during BEAR signals QID is held.
13-Day Exponential Moving Average
Date Signal Return $100.00 % Gain
1/2/2018 BULL 24.69% $124.69 24.69%
10/5/2018 BEAR 5.70% $131.80 31.80%
1/31/2019 BULL 42.26% $187.49 87.49%
2/27/2020 BEAR -31.84% $127.80 27.80%
4/29/2020 BULL 209.83% $395.97 295.97%
11/22/2021 BEAR 47.38% $583.58 483.58%
1/13/2023 BULL 0.00% $583.58 483.58%
1/13/2023
9-Day Weighted Moving Average
Date Signal Return $100.00 % Gain
1/2/2018 BULL 27.74% $127.74 27.74%
10/4/2018 BEAR 8.45% $138.54 38.54%
1/31/2019 BULL 58.08% $219.00 119.00%
2/26/2020 BEAR -24.99% $164.28 64.28%
4/29/2020 BULL 213.61% $515.19 415.19%
11/18/2021 BEAR 60.84% $828.60 728.60%
1/9/2023 BULL 7.60% $891.54 791.54%
1/13/2023
So if you are started with $100 invested on 1/2/2018 and put it into QLD you'd end up with $583.58 using the 13-Day EMA, while you'd do somewhat better using the 9-Day WMA resulting in a final amount of $891.54.
If you held QLD LTBH - no buys or sells from 1/2/2018 your gain would be 105.71% compared to 791.54% using the 9-Day WMA. In contrast, if you bought QID at the same time your loss would be -88.26%.
In other words, this does a bit better than just a risk management approach. A way to improve on this even more is to have a tool for identifying major market bottoms. Using such a tool you would not wait for it to return to a bull signal after going bearish, but instead, you'd enter based on the tool to identify a bottom. Jim has suggested signals from his tool over the years as one option for doing that.