Investment Strategies
What They Are
How to Use
Backtesting Details
What They Are
Searching for Patterns
Like quantitative hedge funds - large investment firms that usually manage the money of very rich individuals - Portfolio Strategies find stocks by searching for patterns in the financial performance of companies.
For example, imagine that you looked at historical data and found that, on average, stocks that pay high dividends as a percentage of their current price tend to outperform those that pay low dividends as a percentage of price (this actually is true based on data from the past 30 years). You could write a computer program that looks at the ratio of Dividend Per Share divided by Price and gives you the top companies based on that ratio.
How YCharts Strategies Uncover Patterns
We spend lots of time researching what worked in the past1 and can look at much more complex relationships than simple ratios. Once we find a good strategy, we program it into our system. Every day, our system checks the world of stocks for ones that meet the strategy's requirements, and we create a list of the top companies for every strategy.
The result is that you always have lists of stocks that fit strategies that worked in the past.
There are currently five Portfolio Strategies:
- Large Cap Value
- Dividend Power
- Peter Lynch Universe
- Ben Graham Formula
- Growth at a Reasonable Price
(1) Note that what worked in the past is never guaranteed to work in the future. We have done our best to only share strategies that outperform over long periods of time, but you should never expect results identical to those that occurred historically.
How to Use
The Details Matter
Pro strategies were tested in a very specific way. Some people who look at the historical performance believe that every stock that our strategies select will perform well. That is not true. This information will tell you reasonable ways to use strategies.
Reasonable Uses
There are two reasonable ways to use them, but either way, read the “watch out for” section below carefully.
- Use the portfolios to generate ideas for good stocks to research more deeply
- Follow the entire portfolio, not trying to distinguish between individual good and bad picks
When using, remember that strategies were tested over long periods of time (20 to 30 year periods), but there are strings of months where every strategy lost money. We judge returns over years, not days. Do not expect instant results, and focus on earning large amounts slowly.
Watch out for: Four very important issues.
- Every strategy was tested as a portfolio - Individual stocks within the strategy were not checked for performance. Therefore, never assume that every stock will behave as the portfolio did - in fact, they are usually much more risky individually. Some will win, and some will lose. If you want to see approximate probabilities, check the winning percentage statistic for each portfolio.
- Performance statistics are historical and no guarantee of future returns - We aim to give you an edge based on history and feel that there is much to be learned from how our strategies performed in the past, but you must decide if you believe such strategies will continue to work over the long haul.
- Volume - Throughout certain periods, stocks may enter portfolio strategies that have low daily trading volume. This means that trading the whole portfolio may prove complicated, especially with large amounts of money. Proceed with caution.
- Transactions Costs - People with small amounts of money to manage will find that transactions costs from trading many positions may outweigh the benefits of investing in backtested portfolios of this nature. Please consider the costs carefully.
Backtesting Details
This page is for those of you who want the nitty gritty details of how the portfolios were tested. We have deep respect for you if you go this deep.
How We "Trade" the Portfolios
Each portfolio has a specific quantitative algorithm that is used to generate the portfolio. First, our system selects "eligible" stocks based on characteristics such as Market Cap, current price, debt to equity ratio, etc. For example no stocks under $50 million in market value are included in any strategy.
After eligible stocks are picked, we sort the stocks based on a metric. Dividend yield is the most common sorting metric.
Finally, we select the top 35 stocks based on the sorting procedure. That gives us the portfolio for the month. In testing, we re-ranked the stocks on the last day of each month and bought at that day's closing price. We buy the same dollar value of each stock as opposed to some other systems which buy different amounts based on the market capitalization of the stock (for those familiar with the terms, we use equal-weighted instead of value-weighted purchases).
Further, each month we re-balance. This means that if any stocks value has grown to more than 1/35th of the portfolio, we will sell some of the stock and use the proceeds to buy new stocks. Then, at the start of each new month, every stock composes only 1/35th of the portfolio.
Summary of Trading Rules
For those who want a quick summary to know what settings we used, here they are:
- Portfolios were equally weighted (not value weighted) meaning that each month, we started with the same dollar value of each company, regardless of the size of the company
- We recreated, traded and rebalanced portfolios every month, selling any stock that was no longer in the top 35 by rank
- Whenever the stock was no longer in the universe of eligible stocks, we sold it
- 3% of assets were held as cash which earned 1% annually
- In the Large Cap Value and Dividend Power portfolios, dividends were not reinvested. For other strategies, they were.
- We assumed a short term capital gains tax rate of 28% and a long term capital gains tax of 15%, which was deducted from returns (1 year was the cut-off between short term and long term holdings)
- Transaction costs were 2.5 cents per share traded
- We assumed "slippage" of 0.5% when we executed the trades