My favorite short term indicator the DVB is not having a happy time of it, doggedly signalling long, but the market has been stubbornly refusing to listen….
Trading the indicator via the basic method of long below 0.5 and short above 0.5 is still up 6.14% for the last 250 days on the SPY, as the equity curve below shows, it has fallen from grace somewhat.
I don’t trade this indicator blindly and am a profound Eurosceptic so have kept positions light since things kicked off. It’s not exactly been a vintage month though.
There are many factors at play here such as an increasing trend following bias, but I wanted to break down the indicator’s performance based on the number of days in a trade to see if there was a point where a strong selloff became a strong sell signal not a strong rebound signal.
Method:
Long with DVB reading of 0.5, short above 0.5.
Trading SPY dividend adjusted. No commission/ slippage etc.
Long trades:
The table below shows returns from the method depending in the consecutive number of days in the trade. 1 day means the trade lasted just 1 day, days means it lasted 2 days (no more). For “2 days” there was a consecutive buy signal over 2 days (no more).
CAGR = Compound Annual Growth Rate.
MAR = CAGR divided by drawdown
Interestingly, in many ways, the strategy needs to be given a few days to breath. Returns are inferior if the trade lasts just a couple of days. The optimal risk/ reward appears to come from 3 days, though this does also seem to be the peak.
Short trades:
Shorting shows a slightly different pattern, here 2 days is the optimal risk/ reward with a market decrease in performance until we get back out to 8 days and beyond. This asymmetry makes sense considering the market’s inherent upwards bias. You’d know a lot quicker if a short trade is going to fail than a long trade.
Summary:
There’s no smoking gun here. Consecutive signals over multiple days does not necessarily mean you should reverse the signal. However, for both long and short trades, risk/ reward does appear to peak after 2-3 days. If combined with some sort of abnormal market filter, there may not be too much harm in cutting trades loose after this time period.
