| I work every day as a programmer on a quantitative | | | | calculation done it's possible to compare very |
| stock analysis team that manages billions of dollars in | | | | different indicators. |
| assets. That position has given me a lot of insight | | | | Getting close to the end of the process is another |
| into how such a professional quantitative stock | | | | very key part of the quantitative stock analysis |
| analysis process works. | | | | calculation. To get one final number for each stock |
| There are a number of distinct steps involved in | | | | you take some weighted combination of the |
| getting a quantitative stock analysis model running. | | | | standard deviations of the indicators. It doesn't |
| Start with a list of stocks. This is the entire domain | | | | necessarily have to be a linear calculation. This |
| of stocks that you'll be considering so you might | | | | function decides the importance of the different |
| want to include the entire NASDAQ. There are a | | | | indicators in the overall analysis. At this point it is also |
| number of considerations when choosing the stock | | | | possible to split out a few different final analyses |
| list. First, the more stocks the longer it'll take for the | | | | using the same inputs. For example a growth |
| computer to calculate everything. Second, there are | | | | strategy would be heavily weighted towards the |
| differences in behavior between for example large | | | | growth indicators, and a value strategy would be |
| cap and small cap or between different sectors that | | | | weighted towards value based indicators. This |
| make it hard to treat them all the same. | | | | formula requires lots of back testing to get just right |
| For each stock in the stock list selected the second | | | | and should be revised from time to time to deal with |
| part of the quantitative stock analysis process is to | | | | changes in the markets. |
| load all the data you can about them into a database. | | | | The final step is to rank each stock. Simply sort the |
| Some basic information might include price, earnings, | | | | stock list by the number that came out of the |
| forecasted earnings, cash flow, assets, and debt | | | | previous step. The result should be a list of stocks in |
| leverage. The more information you can get the | | | | which the ones at the top are probably the best |
| potentially more thorough the quantitative stock | | | | buys, and the ones at the bottoms are probably the |
| analysis will be. | | | | best to short. It's important to remember that the |
| Third, with all this basic information about every | | | | analysis is only as thorough as you make it so if you |
| stock the next thing to do is to calculate indicators. | | | | haven't accounted for things like merger speculation |
| These are usually basic ratios, so from price and | | | | or growth through acquisitions then you may want to |
| earnings you can get the Price-Earnings Ratio, or EPS | | | | do some further checking before placing any orders. |
| Yield. If you're maintaining this information over time | | | | The final list is very valuable and can be used for |
| then you can calculate things like price momentum. | | | | different strategies. There's still a question about how |
| This is a vital step to the effectiveness of any | | | | much of each stock should be bought or sold or |
| quantitative stock analysis model. Tweaking these | | | | whether or not to short the worst ranked stocks. |
| indicators can have a huge impact on the outcome of | | | | Every investor should decide how to do this on their |
| the process. | | | | own based on a lot of back testing. |
| Next there is a mathematical process. Since we can't | | | | That's a very high level over view of how a |
| directly compare price momentum to EPS yield there | | | | quantitative stock analysis could work. Of course |
| is an intermediate step that puts everything into | | | | there are many potential variations. The work |
| standard deviation space. For each indicator we have | | | | involved in getting something like this set up for |
| to calculate it's standard deviation from the mean | | | | personal investing is probably prohibitive. Check out |
| over all the stocks. This should result in numbers that | | | | my website for some ideas about how to start using |
| are roughly within the range -3 to +3. With this | | | | a quantitative approach today. |