Here is the idea:
We are all familiar with the notion of Support and Resistance. Most of us also know that these price levels “work” because they are widely recognized to work. The more acknowledged a price level is as either support / resistance the more power that level has, which is why the 58 SMA is fairly useless, while the 20, 50, 100 and 200 are significant.
Decisions on when to enter and exit a position are often based on these levels, whether they are from the moving averages (simple or exponential), horizontal lines of well-tested price points, intersections, down/upward sloping trendlines, etc. retail traders and more importantly, institutional traders, use these areas with remarkable consistency.
However, as we have seen, some areas of support/resistance are more easily breached than others. Some stocks are more likely (i.e. more volatile) to break through support/resistance regularly, whereas others tend to respect those boundaries unless something extraordinary happens to either the stock, sector or market.
It is clear that certain variables are required for a stock to have the strength (or weakness) to transgress these borders – heavy volume, a strong/weak market, news, earnings, etc. Some combination of atypical elements pushes the price of an underlying past that of which is expected.
Thus, wouldn’t it be theoretically possible to do the following:
1) Identify which types of support/resistance have more strength than others (i.e. a S/R price point that is from a major SMA is stronger than horizontal S/R over ‘X’ number of days).
2) Identify which forms of catalysts are stronger than others (i.e. earnings is typically more potent than sector rotation).
and then finally
3) Identify which types of stocks are more susceptible to breaking out of those ranges.
With these three identified and put into some form of taxonomy (i.e. A Level 4 Support met with a Level 3 catalyst on a Level 1 Stock) one can begin to devise models that would indicate the required conditions for a stock to violate their Support and Resistance lines.
Thus, as, for example, if AAPL is approaching resistance, and you know that given the category of stock AAPL falls into, the type of resistance it is approaching (e.g. well tested horizontal support over a two month period), it would require the following catalyst to breakthrough (e.g. Positive product news or stronger). Absent that type of catalyst you can have reasonable certainty that the resistance will hold.
Anyway, just a thought. This type of model would most likely require an intense amount of data and iteration to work. I imagine one could try to crack this for many years and still get nowhere, but I do believe there is something here, so I wanted to throw it out there.