Light volume Little Help?

A post by Babak at Trader's Narrative commented on an article by William Hester on recent market conditions. In it he concluded

But there is an important nature of volume that I don’t think he took into consideration. Everything else being equal, volume tends to be cyclical and follow a pattern. For example, it tapers off during holidays like Christmas and New Years (yellow squares on the chart). Volume is light from June to August - what is referred to usually as the summer doldrums. And it spikes for panic lows (you can see a few examples above).

But is there anything we can look to from recent low volume?

I started from using June 8th 2009 as a base; on Monday the amount of volume traded was 32% below a 60-day moving average of volume. In the prior week volume traded between 6 and 21% below the 60-day moving average of volume. So looking back to 1950 how many days in June did volume trade below the 60-day MA of volume by 30% or more?


The majority of these cases occurred in the late 50s and early 60s with large number of light volume days populating June. But there was a dramatic shift from 1979 to 2009 with only the odd day here and there and no light June trading day in the nineties.

How did the S&P perform during these periods?

I tried to group the performance into general associations. The first group of five enjoyed periods of continued strength into the end of July or beginning of August before entering a downward phase lasting into mid-September and October. Only in 1988 did the rollover occur early and bottom at the end of August.


The next group of five had sustained periods of weakness for June or mid-July, but this was followed by significant rallies where the June/July low reflected the absolute low for the following 6 months.


And then there was the odd cases which didn't really fit any pattern (although 1976 could probably muscle into the 'June low and rally' group)


So is there anything we can take from this?

Not really. There is nothing to suggest very light volume as part of a seasonal light trading period reflects underlying weakness. In the case of light June trading there is perhaps a greater case to be made in favour of a consolidation lasting to the end of June/mid-July before markets kick into a significant push higher - but it's a weak case.

However, with the exception of 2001, there was no strong evidence to a significant market meltdown and certainly no move which made a substantial undercut of June prices.

So, seller beware.

In my next article I will look at how the relative position of the S&P to its moving averages combined with light volume influence future price action.

Dr. Declan Fallon, Senior Market Technician, Zignals.com the free stock alerts, market alerts, and stock charts website
Read more!

It's 1974 again - except worse

On November 20th the S&P closed at 752. In the context of the last fifty eight years this is the worst performance the S&P has ever recorded. In 14,368 data points, the past six weeks rank 14,349th to 14,368th with respect to S&P performance relative to its moving averages (20-, 50-, and 200-day MAs). On the day it closed at 752 the S&P finished 16% away from its 20-day MA, 25% from its 50-day MA and a staggering 40% from its 200-day MA. The only other year in recent memory to come as close was October 3rd 1974. Then the S&P closed 7% away from its 20-day MA, 14% from its 50-day MA and a comparably bullish 29% away from its 200-day MA.

For the month of December, taking the post September/October routs into perspective, 1987 was the only other year coming close to matching now. Then the S&P lingered 2% from its 20-day MA, 10% from its 50-day MA and only 20% from the 200-day MA (this was for December 10th 1987).



Bears and worried bulls could look to December 1973 when the S&P traded 3% from its 20-day MA, 10% from its 50-day MA and 13% from its 200-day MA - only to see the following year perform even worse. Will this be the story for 2009?


The most likely outcome is somewhere in between as the market drifts sideways as the moving averages 'catch up' to the market. One month on from my "Obama Bottom" article the S&P has completed the Obama unwind and should be in a position to rally from here into the early part of next year. But given the new boundaries of despair the S&P has set I wouldn't be betting big on it.

Just for interest, the best the market performed was November 3rd 1982 when the S&P traded 5% above its 20-day MA, 12% above its 50-day MA and 23% above its 200-day MA. Those numbers appear a million miles away from where the S&P stands now.

Have an opinion you would like to share? Make a call on your favourite stock. Here's the call spread for Bank of Ireland (BKIR)



Dr. Declan Fallon, Senior Market Technician, Zignals.com the free stock alerts, market alerts, and stock charts website
Read more!

Better Know an Indicator: Variable Moving Average

In a series of articles I will detail some of the lesser know indicators available for use on Zignals. The first one is the Variable Moving Average.

So what is the Variable Moving Average (VMA)?

MarketScreen.com gives a description of each of the key moving averages. Based on their definition:

A variable moving average is an exponential moving average that automatically adjusts the smoothing percentage based on the volatility of the data series. The more volatile the data, the more sensitive the smoothing constant used in the moving average calculation. Sensitivity is increased by giving more weight given to the current data.

A variable moving average is designed to perform better in trading ranges - the achilles heel of moving averages.

In Zignals Stock Charts you can set the VMA for the open, high, low and closing price for any period (in days).


You can see on the chart how much more responsive it is to the volatility of the underlying instrument


But that doesn't necessarily mean it generates better signals:


However, by doing a little tinkering - switching the VMAs to a 26-period of the price high and a 12-period of the price low, you get a dramatic drop in the number of signals and an increase in the quality of those signals as it 'kicks you' out of the trade quickly if you are on the wrong side of the move:


Something the core exponential moving average doesn't do - although it's not doing to badly based on the above parameters (nicely out since November 2007):


As you can see there is plenty of opportunity to fine tune the settings, something only Zignals Stock Charts allow for free.

If you would like copies of these charts for your Zignals account please email me: declan-at-zignals.com

Dr. Declan Fallon, Senior Market Technician, Zignals.com the free stock alerts, market alerts, and stock charts website
Read more!

S&P Moving Average Behaviour revisited

Back in June I did a study on the relationship between the S&P to its 20-, 50-, and 200-day MAs. The conclusion was not pretty reading:

The take home lesson is for the next month or two further downside is not just likely but probable; only in 2003 did a rally develop soon after the match. For the other five of the six matches the S&P lost between 7% and 30% of its value before it finally turned around.

Well, since then the S&P has trimmed off 14% and even this might yet not be enough. What is the current state of play for the moving average relationships?

As of Friday's close the S&P was 9% off its 20-day MA, 13% from its 50-day MA, and 21% from its 200-day MA. Have there been situations in the past which closely mirrored this - and if so, what happened next?

There were five periods since 1951 which matched this set-up in the S&P


How did markets perform after these periods?

In 1962 the moving average relationship was not an immediate marker for a bottom, but a major bottom did occur 2 weeks later. It was followed by a short term bounce and eventual retest 4 months later. Once the retest completed a new bull market began:


A similar theme played out in 1970. The sharp decline was followed by a relief rally and retest before a substantial bull market followed. As before, the actual bottom occurred 2 weeks later although the retest only required 2 months to complete.


In 1974 there was another match very similar to the previous two except this time the bottom took 2 months instead of two weeks to occur. The retest took another two months before the next bull market kicked off:


The greatest number of matches came in 2001. This time the relationship between the moving averages and the S&P came within a week of the lows, although the attempted retest failed and new lows were posted in 2002. It was in 2002 that another matched relationship was made between the MAs and a positive retest was posted 3 months later. This eventually led to the cyclical bull market completed in 2007:


What do these historical comparisons tell us?

We are likely a couple of weeks from a bottom, but it is not impossible for this to take longer

During this period the market will see sharp losses, perhaps trimming 10-20% off where the markets lie now (Monday will be the start)

The subsequent rally will be short lived and will morph into a retest of the low

The retest will be the time to buy heavy

A significant bull market has a good chance of emerging from the quagmire - remember markets lead economic news.

Use Zignals Alerts to notify of your favorite stocks taking 10, 15 or 20% trim from Friday's close, or % move from a moving average as illustrated below:



Dr. Declan Fallon, Senior Market Technician, Zignals.com the free stock alerts, market alerts, and stock charts website
Read more!

Mark McRae; A Unique Moving Average Cross

Over at Trader's Blog there is a guest article by Mark McRae on a moving average crossover method he uses. The basis of the method is using a moving average of open price in conjunction with a moving average of closing price and trading the crossovers. A MACD or ADX is applied as a filter to reduce whipsaw trades. With Zignals Stock Charts you can create moving average overlays using Open, Close, High and Low prices, and moving averages can be simple, exponential, weighted, time series and variable. A MACD or ADX indicator can be added as suggested for a filter. Below are some examples

Mark McRae uses a 5-period exponential moving average of the close and a 6-period exponential moving average of the open:


A sample chart of Mark's output would look like this:


I looked at a wighted moving average crossover with a 3-period moving average based on the High and a 5-period moving average based on the Open; there are fewer whipsaw signals but I'm not sure the quality of the triggers has improved?


Dr. Declan Fallon, Senior Market Technician, Zignals.com the free stock alerts, market alerts, and stock charts website
Read more!

Strategy lab: Buy cross of 200-day MA by 50-day MA - Sell 15% gain

Following the success of last week's strategy developed around a stock alert which bought a 2% gap, I looked at this week's stock alert: buy cross of 200-day MA by 50-day MA and tested a trading strategy for it.


Within Zignals, the stock alert historical test looks like this:


The historical testing showed a steady return over time with an average gain of 8% over 90 trading days, and an average maximum gain of 15% for US stocks. The win percentage was well into the 60% range, another bonus:


Test period: A complete bull-bear cycle defined by the S&P (March 20th 2000 to October 8th 2007).

I tested on two groups of stocks:

US stocks (Active Trader): AAPL BA C CAT CSCO DIS GM HPQ IBM INTC IP JPM KO MSFT SBUX T WMT

European ADRs: ALU BHP BP SAP DT ASML STM BCS UN TOT ELN AZN DEO RYAAY LUX

Invested: $5,000 per trade

Commission: $9.95

Trades: Round-trip only; partial trades were excluded.

The test bought the closing price of the day of the moving average crossover and sold at the closing price of the day a 15% gain was registered - in real terms this could generate considerable slippage.

The performance of this strategy showed considerable contrast to the 2% gap buy. The base strategy with a $9.95 commission was not profitable irrespective of the stop strategy employed. Unlike prior strategy labs where commission based trading was not profitable because of the large number trades generated (in some cases up to 1,000 trades over 7 years for 17 stocks), the 50-day MA cross of the 200-day MA generated a lowly 93 trades for the 17 stocks over 7 years. This was well below the 300+ trades of the 2% gap buy which was profitable when a commission was included. So the strategy looks to have a core weakness based on the use of a 3-8% protective stop loss, something the Zignals historical test does not account for (it looks for a straight return without the use of a stop).

The 'best' return was a -$2,052 loss using a 3% stop with a 34% win percentage. The worst return was -$6,133 with a 7% stop on a 26% win percentage.

In the absence of a commission there was no vast improvement in the return. The European ADRs did better using a loser stop, whereas the US stocks performed better with a tighter stop.


How did the strategy perform over three random test years?

The three randomised 1-year periods were March 2002/03 (bearish), October 2007/present day (bearish!!), December 2005/06 (neutral-bullish).

Not surprisingly, the October 2007 to the present day showed the worst losses; the tighter the stop the lower the loss. However, the low number of trades (7 for the 17 stocks) for this period would have kept most people out of the market. The same was true for the 2002/03 bear market. Even the neutral 2005/06 market recorded only 19 trades for 17 stocks. In the latter case the 3% stop strategy recorded a $1,328 profit on a 53% win percentage. The average performance is given below:


The strategy required more breathing room than an 8% stop allowed given the disparity in the Zignals historical test and strategy lab results. Because a 50-day MA / 200-day MA crossover tends to heavily lag a bullish reversal it is perhaps not the best entry strategy for a short term gain (such as a looking for a fixed point or percentage gain) where the probability of a pullback - and a stop hit - is high.

Got a favourite stock alert strategy? We would love to hear from you. Contributions welcome (email: contributions@zignals.com).

Dr. Declan Fallon, Senior Market Technician, Zignals.com the free stock alerts, market alerts, and stock charts website
Read more!