"Congratulations on your superb write-up regarding 200 DMA, in yesterday's commentary.

"While it is easy to imagine (and has been observed, time and again), that the concept of mean reversion would apply to charts individual commodities, or countries, or even an asset class specific fund such as Blackrock World Mining Trust etc., I have always wondered whether the concept of mean reversion would also apply to the chart of a fund which has a mix of asset classes including equities in various countries, bonds of various countries, some investments in precious metals and even some allocation to cash, put options etc. etc. Such a fund would clearly not be traded actively and hence would not have millions of traders looking at, and trading off, the chart and its 200 DMA etc.

"I would be grateful for your thoughts on this conceptual matter. If you can think of some charts that would clarify this issue, then that would certainly help.

"Thanking you in advance."

And

"I have one more question regarding your note on 200 DMA in yesterday's commentary.

"What would constitute a 'significant diversion' from the mean?

"The common sense approach that I have generally adopted is to measure the maximum divergence from the 200 dma in the previous 5/10 years, and to be cautious once that level of divergence is about to be approached.

"Would you say that is a fair approach? Also, would one expect the maximum divergence of the previous 5/10 years to be exceeded, in the event that the extension is from a lengthy base? What constitutes a lengthy base - a few months or a few years?

"I eagerly await your comments."

Eoin Treacy's view Thank you for your kind words and these two interesting questions. The issue of whether a multi-asset class fund is liable to the same process of mean reversion as individual assets is debatable. I believe it would largely depend on what the constituents of the fund are.

In an investing environment where so many equities and commodities have tended to move more or less in line with one another, I question whether a fund made up of these assets alone could avoid mean reversion in sympathy with their respective markets. On the other hand if a fund had a small position in equities or commodities relative to income producing assets such as bonds, commercial property and potentially income trusts/MLPs, then it might be less prone to mean reversion. That is of course unless income producing assets were also the subject of widespread selling.

How the fund is structured as you mention will also be important. An investment trust trades like an equity and is quite capable of mean reversion based on the vacillating emotions of investors as they buy and sell.

Your method of comparing historic overextensions relative to the MA with current trading action makes sense to me. Every instrument has a different rhythm and corresponding set of consistency characteristics. Apprising ourselves of what these are gives us a better understanding of how that instrument trades and what is "normal" for it. I prepared just such an indicator for the S&P500 in October 2008 as an additional way to illustrate how overextended the Index had become relative to its 200-day MA. We are looking at adding this indicator the Chart Library.

What we have occasionally referred to as Brobdingnagian bases are those that have lasted for years and possibly decades. Following the credit crisis, a number of assets formed bases over a considerably shorter timeframe, helped by loose monetary policy.

As I pointed out on Wednesday, a moving average is most useful in a consistently trending market. During the powerful upward breaks we associate with base formation completion, the persistence of the breakout is much more important than the overextension relative to the MA. Within a base, the MA generally trades horizontally. Following a base formation completion it is not unreasonable to expect the overextension to be one of the larger posted, compared to those within the base formation.