Sunday, April 13, 2008

Closing Price and Daily Range

The results of this week’s study really surprised me. Few things are more painful when trading an end-of-day system than to have a stock price drop substantially the morning of your entry. There is obviously no way to minimize this risk completely but anything to reduce the frequency of this would represent a tangible statistical edge, not to mention a major psychological advantage. Conventional wisdom suggest that a closing price in the upper percentages of a stock’s daily range is a sign of strength and narrowing trade selection to these stocks that closed strong could potentially offer a terrific technique to help accomplishing this.

These test will use our “standard” entry criteria discussed last week with the following criteria:
Average Volume = 10-day average not including today - Average(Vol,10,1).
Average Volume >= 100,000 shares
Today’s Volume >= AverageVolume * 2 (100% increase)
Today’s Close > Yesterday’s Close

In addition to this, I am calculating the daily range of each stock on the day of the breakout and will then run a series of tests where we take only closes that fall in the top X% of that daily range. The top 100% represents taking every trade while the top 10% represent only taking trades where the close is in the top 10% of the daily range.

For exits, we will use a simple 1, 5 and 20-day timed exit. As usual, we will run the test on three separate portfolios to increase our confidence in the robustness of the results.

Our primary evaluation ratio will again be our EDR ratio but we will also look at the two other evaluation ratios used previously on this blog. I acknowledge that it may be difficult for a new reader to comprehend my approach by starting here but rather than continuous overly redundant posts, I encourage anyone interested to read through the previous backtesting studies to flush out exactly what we are doing.

EDR=Expectancy per dollar risked. We’ve covered this plenty already but it is calculated by Average Trade/Average Loss.

EER=Efficient Expectancy Ratio. This is our EDR/Avg. Days in Winning trades. This number is meant to reveal the most efficient use if capital as it reveals expectancy per dollar risk per day in trade.

DDR= Drawdown ratio. This number is calculated by Net Profit/Maximum Drawdown. A higher number is better.

I will begin this week by posting the summary graphs followed by the raw data of the tests.



Summary Graphs:





Raw Data:




Conclusion:
If this does not highlight the advantages of backtesting an idea or strategy prior to implementation, I don’t know what does. Using such a strategy, only limiting EOD trade selection to those stocks that closed in the top 25% of their daily range, could have been disastrous. I say “could” because backtesting in no way predicts the future and this could start working terrific tomorrow. Personally however, I am not willing to bet my money on it.

The reduction in expectancy was substantial and consistent in all portfolios. Looking at the Nasdaq portfolio for example, the top 10% of trades when compared to the top 100% (all trades), had an expectancy decrease of 35% 1-day out, 27% 5-days out, and 17% 4-weeks later. I suspect this was perhaps once a profitable strategy but any edge has long been traded away. Skepticism towards conventional wisdom remains a valid strategy for capital preservation.

As we would expect with any entry criteria, the entry criteria itself has less and less of an impact as the time held increases. The underlying fundamentals of price movement ultimately have the largest impact on longer-term expectancy but even after 20-days, there are lingering affects from using this type of entry filter. I would not suggest one reject those stocks that close in the top 25% of their daily range. Often, these are the most powerful movers but limiting ones selection to these is clearly not an advantageous decision.

3 comments:

Jeff said...

B-
The results remind me of what Bill R's research with RSI(2). In the 50s and 60s, RSI(2) was a great trend following indicator. Starting in the 70s through the present, it changed to become an overbought/sold signal.

bhh said...

haha, I haven't tested to see if it works in reverse but it could pair up nicely with the RSI(2) for a mean-reverting system.

Anonymous said...

First time @ the blog, but always appreciate anyone willing to do some research.

Recalling that Larry Williams said markets close to a swing high tend to close on their highs, markets near a swing low tend to close on their lows, it becomes no wonder that such a system has such a low success rate.

perhaps if we add some context. if a market closes in the top 25% of the range, and an oscillator (RSI, %R, stoch) is not overbought, does it work then? Or, previous day closed in the **bottom** 25% of its range, and current day closed in the top 25% of its range.

that sounds like a reversal day to me, and a much better time to get in.