• 
• 
• 

• Performance
• Services
• Company
• Learn About
• My Attain

What's Driving Day Trading Profits in August?

August 30, 2004

 

So much for the dog days of summer. August has been a month to remember for many day trading systems, as Helix, Compass, Blue Wave Zones, and several more systems have posted impressive double digit gains. Coming into the last two days of August, 80%, or eight out of ten, systems day trading the full size S&P were positive for the month. If that doesn't impress you, consider that the previous three months saw just 10%, 33%, and 40% of systems profitable. August has indeed been a breath of fresh air for many investors, and came at a time when things were looking rather bleak for a few of the aforementioned systems.

The natural question many want to know is: What's driving these profits? Why now? Why August? What fundamental/technical shift, if any, led to a more favorable day trading environment? The answers to these questions are extremely difficult to come by, and many so called answers, such as "systems underperform due to low volatiltiy" have little in the way of statistical confirmation. But the reward of finding such answers could be enormous, as it would allow investors to avoid non-favorable environments and invest (or invest more) during more favorable day trading environments, and so the search continues.

What are we looking for when trying to find an answer to why day trading systems tend to perform better in some environments rather than others? We're trying to find some market measure which can explain the shift in environment, such as an increase in trading volume, decrease in volatility, rising prices, falling prices, narrowing credit spreads, etc.

In order to test these market factors, we first need something to test them against, and we chose the average monthly percentage gain for three systems: Compass, R-Mesa, and Daybreaker as a representative slice of day trading activity. These three also have the longest actual trading track records. So in a theoretical month which saw those systems record +10%, -10%, +15%; the average gain would be 15%.

We began searching for answers in the usual places. The oft quoted relationship between volatility and day trading profits was first up. The argument for this relationship is that day trading systems need markets to be "moving" in order for there to be ample opportunity for profits, and that volatility is as good a measure as any for the "movement" of the stock index markets.

To test which factors contribute to a good trading environment, we assigned a value of 1 for every month in which our aforementioned day trading index (the average monthly gain/loss across three systems) was positive. We then assigned a value of 1 to each factor for every month which that factor was greater than its 2 year average.

Volatility?

So, if increased volatility really does help day trading systems, we would expect the day trading index to be positive in those cases where the average daily volatility of the S&P futures was above its two year average. If the day trading index was positive every time the average daily volatility was greater than the two year average, this factor would have a correlation of 1 and the factor would be present in 100% of the profitable months.

The relationship between volatility and day trading profits failed to materialize in our tests, with higher volatility present in just 6 out of the day trading index's 13 profitable months for a correlation of just 0.1712 . The correlation was positive, as expected, but not high enough to be statistically relevant. This should not come as a surprise, however, given that the trading system index has been able to post profits despite market volatility hovering at historic, five to ten year lows for much of the past two years.

Another measure of volatility is the average true range, and we tested against both the average daily range each month and the percenatge of the futures price contained within the range to lessen any effects of lower prices and therefore lower ranges. Both range factors showed no corrleaiton, with correlations at 0.0989 and 0.3026 respectively.

Volume?

We next tested volume as a factor, looking at the average daily volume of both the S&P futures and NYSE. Both tests were again inconclusive, showing correlations of 0.2423 and 0.1209 respectively. But higher volume was present in 54% (S&P) and 69% (NYSE) of profitable months.

Rate of Change?

Maybe day trading systems have a directional bias, and prefer a bullish or bearish environment. We tested this by looking at the average daily rate of change, again assigning a variable of 1 when the average daily rate of change was positive, and 0 when negative. Day trading systems should be able to perform whether the market goes up or down, and true to form, there was no correlation at all to day trading profits, with a 0.0605 calculation.

Following the conclusion that the direction of the market doesn't matter, our next effort took into effect the size of the average daily change, with no regard to the direction of the move. We again found low correlation, at 0.3187, as higher daily moves were only present in 46% of profitable months.

Our Approach

So where do we go from here? Is there nothing which can explain why day trading profits came roaring back in August. We fell back on two trusty allies, common sense and experience.

What we know from experience is that the best environment for day trading success is a market which sees "follow through" into the close, meaning a trading day in which the market closes near its high or low for the day. A perfect example of such "follow through" was this August 5th, when the S&P futures market opened at a price of 1097.80, set a high for the day shortly after at 1098.00, then proceeded to sell off and move lower throughout the day before making a low of 1077.20 shortly before the close and closing at 1078.50. Trading systems enjoying profits that day included Compass, R-Mesa, Cipher, Magnitude, Impetus eRL, and Spectrum.

The opposite of this follow through is a day such as May 12th of this year, which saw a 25 point trading range, yet prices finish well away from either the low or the high. As any investor who has traded systems for long can tell you, the worst days for day trading systems are those days which see morning or early afternoon highs/lows followed by a reversals to the lows/highs. These days cause day trading systems to initiate positions based on strong early moves, only to see them stopped out as prices make even stronger moves in the opposite direction.

We attempted to quantify this "follow through" characteristic by realizing the ultimate follow through day (on the short side) would be a day when the open and high are the same as the close and the low. In this scenario, the daily change (difference between open and close) would be 100% of the daily range (difference between high and low). We took the differences between the open-close range and high-low range to compute daily "range captured" percentages, with a higher number representing better "follow through"

With this data in hand, we did our final test and found a high correlation between the "range capture" indicator and day trading profits. The correlation was 0.7338, with "range capture" values higher than the two year average present in 10 out of 13 profitable months. We also found a possible explanation for the August profits, as the "range capture" value in August was 56.20%, its highest value since March of this year.

Conclusion

Before picking up the phone and telling us to trade 10 contracts on Compass the next time the "range capture" reading is higher than 55%, think over the following. One, we have fit this statistical model to our assumptions. We knew that day trading system do well when most of the range is captured, and set out to create an indicator which measured just that. Two, this is a lagging indicator, and can only tell us when such an environment has been present - not when its going to surface again.

The trick, therefore, is finding out what drives higher "range capture" percentages. We ran a correlation test between the range capture indicator and each of the other factors - hoping to see that higher volume or increased volatility created such an environment, but found nothing significant there.

The conclusion may be that there are no good factors for explaining day trading profits. Such systems are designed to profit in nearly any environment, after all, and trying to see what environment they perform best in may be a fruitless cause. The possibility of predicting when a higher "range capture" environment is about to occur will remain out there, however, and with it every investor's dream of enjoying the positive months while sitting out the negative ones.