Stock Selection - Breakout Analysis

I tested two entry criteria for a breakout scenario:

• Price increases greater than a certain percent compared with yesterday's close
• Price increases greater than a factor of the 20 day average true range (ATR) plus yesterday's close

These are simple criteria to evaluate if there is any initial upside follow through over the short term.  All results were calculated using StockFinder 4.0.  Backtest criteria will be evaluated on a watchlist of stocks that composes IYY - Ishares Dow Jones US Total Market Index.  This ETF is composed of about 1300 stocks that will act as a proxy for the performance of the entire market.

Below is a chart summarizing the the results of stocks increasing by a specified percentage of the previous day's close:

The entry was the open of the day following the breakout day.  The exit was the close of the day two days after the breakout day, therefore holding times were about one day.  This chart summarizes the follow through behavior for the short term after the breakout.  It indicates that for breakouts with a magnitude between 1-8%, there is continued upside momentum on average.  For breakouts greater than 9%, the stocks tends to mean revert.  Breakout levels greater than 3-5% seem to show the best performance based on annual return.

Below is a chart summarizing the results of stocks increasing by a factor of their 20 day ATR over the previous day's close:

Although the stock selection criterion was different, entry and exit rules were the same.  The chart indicates that stocks increasing by more than 1 ATR tend to mean revert.  Breakout percentage seems to be a better selection criterion compared to ATR in terms of determining short term follow through.

It is important to note that the results for the lower breakout magnitudes include the results for the higher breakout magnitudes.  For example, the results for breakouts >10% are included in the results for breakouts >1%.  It would be a good exercise to separate out the breakout magnitudes independently by looking at the performance of breakouts with magnitudes only between 3-4%, 4-5%, 5-6%, etc.  This will indicate if there is a more robust edge by determining which magnitudes produce the best returns.

Currently, the edge for the breakout percentage criterion is quite small, as indicated by the average trade.  Future studies will also evaluate breakout behavior based on various entry/exit conditions.

ANALYSIS SUMMARY:

• Mean reversion is dominant for ATR breakout magnitudes >=1.0 over the 2 days following the breakout.
• Breakout magnitudes between >=3-6% appear to have the best performance based on annual return.  More testing on the results for each threshold independently is required to confirm this observation.
• Breakout magnitudes >=9% tend to mean revert over 2 days.