Answering Readers' Questions: Shorting the 5-10-20 Trading Strategy

I received some good questions re: the 5-10-20 trading strategy that I shared on Monday: (1) trading long & short vs long-only, (2) is the 5-day EMA really necessary, and (3) the impact of transaction costs. I’ve done my best to respond below.

Trading Long and Short vs. Long Only

The strategy I shared traded long-only. What would have been the effect of going short rather than moving to cash when the 5 and 10-day EMA crossed below the 20-day EMA? See chart below (long-only in blue, long/short in red, 1972 – present):

click to enlarge

[logarithmically-scaled]

click to enlarge

Historically, going long and short would have increased absolute returns, but would have also pretty significantly increased downside volatility. I think a long/short approach with this strategy would be reasonable, but for my money, I’d much rather use that time in cash to find better opportunities.

Is the 5-Day EMA Really Necessary?

Savvy readers noted that because of how an EMA is calculated the 5-day EMA is not usually necessary. The 5-day is almost always going to be above the 20-day when the 10-day is above, and almost always below the 20-day when the 10-day is below.

That is absolutely correct. I shared it with the 5-day EMA because that’s how I originally found it at the dk Report, but as the stats below show, it doesn’t really impact the signal.

click to enlarge

Transaction Costs

This is more of a general statement about all of the strategies you’ll find on this blog. I almost always test frictionless (meaning no transactions costs or slippage) for two reasons:

1. I only trade leveraged funds from ProFunds, Rydex, and Direxion, where there are no transactions costs or slippage. So the results here could be duplicated without extra trading frictions.
2. As I’ve discussed before, I wouldn’t trade any strategy by itself that I discuss on this blog. For my own money, I employ strategies that are taking many signals into account. Readers can see a free and very much un-optimized version of this “strategy combining” in action on the State of the Market report. The goal of course is to create a more robust signal that is taking many different indicators (short-term, intermediate, and long-term) into account.

Thanks for all of the reader comments here and at Greenfaucet and Seeking Alpha. In a follow up post, I’ll continue to expand on the 5-10-20 strategy and apply it to the Nasdaq 100 (which traders would be much more likely to trade than the full composite) as well as the S&P 500.

Comments

[John Lounsbury:]

Michael - - -

Thanks for addressing the questions. One question you didn't cover is whether T+3 delays for ETFs would impact results. If you haven't done the analysis, can you at least give your sense of feeling on the matter. If T+3 has significant negative impact, two strategies would have to be considered:

1. Maintain a sufficient money market balance to cover the T+3 delay. Since this would be 50% of the account balance, returns would be reduced to 1/2 on the stock portion plus money market return. To mitigate this reduced return, do you think a strategy could be devised that would enter and exit positions in 50% legs, attempting to be 100% invested most of the time, but making a 50% move when the approach to a signal is anticipated?

2. Trade in a margin account so that margin loans would cover the three days. This would not impact the size of the index investment that would otherwise have been made (non-margin account).

Another question relates to ETFs. Aside from friction (transaction costs) and T+3 considerations, are there any other factors to be considered for applying the strategy to ETFs?

2008 Dec 04 10:25 AM

[MILESCFA:]

"I only trade leveraged funds from ProFunds, Rydex, and Direxion, where there are no transactions costs or slippage"...

So you BUY at the bid and SELL at the ask? You must be a market maker. I'm not.

If you simply put out a limit order, say a buy at the current bid (or vice versa), this is still NOT frictionless. It's simply the prudent use of limit orders. The ONLY WAY THERE IS NO SLIPPAGE IS IF THE BID-ASK SPREAD IS ALWAYS ZERO. It's not!

2008 Dec 05 04:15 PM

[Directoy:]

Thanks for answering all our Questions!

2008 Dec 31 03:03 PM

[Michael S:]

RE to MILESCFA: there are no transaction costs or slippage because these are not intraday traded vehicle. They are mutual funds that (a) are leveraged and (b) do not penalize the trader for active trading. These are the only vehicles I trade. That's not to say that I think other vehicles are inferior, but they work well with my swing trading style. michael

Jan 07 01:11 PM