As we wind down Financial Literacy Month, I want to point out Joel Greenblatt's book The Little Book That Beats the Market. In the book, Greenblatt describes a trading model that he developed. While I have not tested his model and therefore can't recommend an investment strategy that's based on it, the book is worth reading because it outlines how he developed a profitable and consistent model for trading.

I have my Seton Hall students read Greenblatt's book in advance of a class I teach entitled Simple Guide to Building an Investment Model.

Models give us an objective manner in which to trade. They help take some of the guesswork and subjectivity out of deciding when to buy and when to sell. Just like chart formations provide technicians with trading models, a well-tested statistical model can provide a framework for successful trading as well.

Thus, as you develop your investing knowledge, here is an eight-step overview of my investment modeling class.

Step 1. Develop a Hypothesis

Begin by outlining a theory which you may believe can lead to a profitable model. Pose the hypothesis on some sort of logical statement ("If X then Y.") For example, the hypothesis might be something such as, "If the S&P 500 drops by more that 2% on a Friday then the index will rise on the following trading day."

Make sure that the hypothesis is a bona fide relationship. In other words, comparing tech stock prices to consumer spending might work, while a comparison to pharmaceutical prescriptions makes no sense.

Step 2. Build a Database

In order to prove or disprove your hypothesis, you need to accumulate data over a long period of time. The data should be in some sort of time series so that if you need to compare it to other data points, you can line up the many variables which you will be utilizing.

Remember, the more variables and the more data points which you use, the more significant your output will be from a statistical perspective.

You must make sure that the data is from a reliable source, such as a premium service like Bloomberg or Reuters, or a solid free services like Yahoo! Finance (see " Investment Research: Ignore the Ratings, Read the Reports"). Also, I suggest that the data is retrievable in some electronic and downloadable form.

Once you download the data, you will want to check to make sure that there are no missing, incorrect or "corrupted" data points. For example, sometimes a holiday will show up as a data point which must be deleted.

Step 3. Make Observations

With the raw data that you have now aggregated and organized in a time series database, you will now scroll through the data and make some observations on what you see.

Wondering what you're looking for?

Your goal here is to notice any significant changes in the "dependent variable" (the "Y" in the "if X then Y" example above) based on changes for one or more of the "independent variables" (the "X" in the example above).

This exercise is one of visual observation that seeks to achieve one of two objectives. First, you want to confirm -- in a subjective and non-quantifiable manner -- that your original hypothesis (step one) is directionally correct and worthy of additional analysis. Second, you might detect a pattern or anomaly that was not part of your original hypothesis and could form the basis of a new or modified theory.

As an example, you might recognize that "down" Fridays may be followed by "up" Mondays.

For more on how to analyze data, read "What Investors Need to Know About Historical Data" and "15 Ways to Check Data" on TheStreet.com.

Step 4. Develop Calculations

This is by far my favorite step in the process. I call it "torturing the data."

In my class, the students take their hypotheses and observations and we begin to test them out using formulae and calculations. This may take one of two forms: logical queries or "regression analysis."

Logical queries will say if condition "A" exists, then perform calculation "B" and provide me with the desired output "C." Regression analysis is a mathematical or statistical operation in which you attempt replicate or predict the dependent variable, by using the independent variable.

If you perform a regression, then it is important to determine that your output meets or exceeds many of the statistical tests or requirements that confirm the statistical soundness and significance of the output.

The calculations that you perform in step four will now dictate a set of trading rules that you can now "codify" (or "systematize").

For example, I started out with the hypothesis "If the S&P 500 drops by more that 2% on a Friday, then the index will rise on the following trading day."

Now, through observations and calculations, let's say I determine that the trading rule here is "If the S&P 500 drops by more than 2% on a Friday, then buy the S&P 500 on Friday's close and hold it until the it gains 1%."

A great example of a trading rule was developed by TheStreet.com's James Altucher, which he calls the "QQQQ Crash System." Here is how the rule is stated:

• Buy when the price of the Nasdaq-100 Trust (QQQQ) closes at a 1.5 standard deviation below the 10-day moving average of the low price of each day.

To calculate this, you can use the lower Bollinger Band using a 10-day moving average, 1.5 standard deviation and the price series of the lows of each day.
• Sell at either a loss after 20 days or on any day that the QQQQ closes higher than the entry point.
• Step 6. Back-Test

With your trading rule, which was developed through observation and hypothesis, now comes the true test, which is to determine whether the rule -- if traded on a consistent basis -- can actually make money.

You do that by simulating the trading rule "back in time," applying it to the historical data in your data base and calculating the gain or loss, which would have resulted from the rule - as you have stated it, without modification.

Step 7. Does the Model Make Money?

If the answer is yes, then move onto the next and final step. If the answer is no, then you can take one of two paths. First, you can go back to step three (Make Observations) and develop a modified trading rule. Or, you can alternatively, simply abandon the model and begin it anew when a new hypothesis is developed.

Step 8. Can the Model Be Traded?

When you have created a model, which on a back-tested basis can make money, there is one final step you need to take before "going live" it. Step eight is where the practical aspects of trading take over from the academic aspects of the model development. Here, you need to insure that you can trade this model in the real world.

For example, if your model dictates the short sale of a stock or index, you must make sure that such or a stock or index can be borrowed (see " How Short Selling Works").

Another example may be one where you buy Chinese stocks and short-sell Hong Kong stocks. This rule may be impractical if you do not have access to the Chinese "A shares" because of regulatory or legal reasons.

I will elaborate on each step in more detail in follow-up installments of The Finance Professor. In the meantime, formulate your own hypothesis and start to aggregate the necessary data required to develop your own trading model.

At the time of publication, Rothbort was long Ultra S&P500 ProShares (ticker: SSO, the S&P 500-leveraged ETF), although positions can change at any time.

Scott Rothbort has over 20 years of experience in the financial services industry. In 2002, Rothbort founded LakeView Asset Management, LLC, a registered investment advisor based in Millburn, N.J., which offers customized individually managed separate accounts, including proprietary long/short strategies to its high net worth clientele.

Immediately prior to that, Rothbort worked at Merrill Lynch for 10 years, where he was instrumental in building the global equity derivative business and managed the global equity swap business from its inception. Rothbort previously held international assignments in Tokyo, Hong Kong and London while working for Morgan Stanley and County NatWest Securities.

Rothbort holds an MBA in finance and international business from the Stern School of Business of New York University and a BS in economics and accounting from the Wharton School of Business of the University of Pennsylvania. He is a Professor of Finance and the Chief Market Strategist for the Stillman School of Business of Seton Hall University.