1. is a bad guide to stock market investing because it does not take judgmental factors into account.
2. might well be good at identifying high-risk stocks.
3. will probably also select stocks that out-perform the market average in bear (declining) markets.
4. is a good guide to stock market investing because it selects stocks that out-perform the market average.
Choose the correct option?
The correct answer is option 2. High-risk portfolios tend to yield higher returns than the average of the market. They also tend to earn greater than average capital gains in rising markets and greater than average capital losses in falling markets.
Option 3 is clearly wrong. The model will probably select stocks that decline more than the market average in bear markets because stocks that rise more than the average in bull (rising) markets tend to fall more than the average in bear markets.
Option 4 is also clearly a bad choice. Risky stocks out-perform the market average in terms of rate of return, but that does not mean that one should choose risky stocks---it will depend on one's degree of risk aversion. Moreover, the particular portfolio selection model in question may not even be good at selecting stocks that will earn higher returns than average over the very long-run, averaging out both bull and bear markets. It has merely identified a set of stocks that rose more than the average in the a particular bull market.
With respect to option 1, one can use judgment along with the model to make selected departures from the model's recommendations if one wishes. The performance of the market, however, reflects the judgment of the average market participant---the purpose of the model is to attempt to do better than one can do on the basis of judgment alone by extracting evidence from available data.
It is interesting to note that any model that turns out to be a truly successful method of portfolio selection---that is, picks stocks that will earn a higher return than the average after adjusting for risk---and becomes widely used will destroy its own success. Any information it is able to produce will be incorporated in the prices of the stocks and one will not be able to use it to "do better" than the market. Indeed, one will have to use it to achieve the same portfolio return, adjusted for risk, as the rest of the market achieves.