Replies to post #738 on StockCharts Support Forum

[Cush]

Hi moonshadowus. "works best" is a phrase that leaves a lot of room for interpretation.

I'm sure you can see just by changing the period settings how the appearance changes.

Sometimes if I'm trying to see whether indicators have crossed (if they're really close) I'll shorten the time period to get a closer view. The Moving Averages are still in the same relationship just easier to see.

On the other hand, sometimes it's easier to see a pattern developing if you look at a longer time frame.

What did you have in mind when you wrote "works best"?

Cush

[moonshadowus]

thanks, what works best?, using a 1 yr chart with weekly format? and the other one 6mnth chart with a daily format ,would these be good? i have read that a 50 moving day avg either lets you know if institutional money is either moving in or out so im just trying to weed out any knee jerk reactions when a rumor hits and trading becomes crazy. thanks moon

[Tom K]

Backtesting Moving Averages

Moon,

Read this article:
http://www.basicsoftrading.com/journal/2001-1/2001-1-09/

Here's his conclusion:
"We have performed computer analysis of various trading strategies using moving averages. the results for trading the Dow Jones Industrial Averages for the period form 1975 to 2001 show that there is no strategy using moving averages which is better than the buy and hold strategy."

So, do we chuck Moving Averages? I don't think so. Moving
Average Systems are extremely valuable trend indicators. That said, you must understand the weaknesses of moving averages and compensate for them by using other indicators. Moving Average system are always late; so, use reliable contra-trend, sentiment, or thrust indicators to compensate for this flaw. Moving averages are subseptible to whipsaws, especially in trendless markets. So, develop a model of several other indicators - eliminate the noise, avoid frequent whipsaws.

Word of advice: don't waste a ton of time trying to find the "best" combination. Optimized parameters almost guarantee that the results won't hold up in real time use. Choose your moving averages based on the cycle length you are interested in trading, not by the best back-tested results.