“These Models are too good to be true!”  While we are flattered by this statement, we are convinced we can do better.  Why?

Quite often, statements such as “It’s too good to be true” are the result of our own point of reference.  As an example, the media and financial services industry have force-fed the S&P 500 down our throats, so we compare and measure every investment to the S&P 500. However, if you expand your comparative point of reference to include indices that are broader in scope and equally-weighted, you would realize the hypothetical returns of our models are quite reasonable.

It's All in the Weighting models are equal-weighted.

The following chart compares the S&P 500 Index (a common benchmark for stocks) against a more comprehensive Wilshire 5000 Equal Weight Index.  As you will see, the Wilshire 5000 Equal Weight Index has averaged nearly 16% per year since 1975 compared to the S&P 500 Index's 11% (per year) over the same time-frame.

If a strategy is meant to choose “the best of an index,” the strategy should provide hypothetical average annual returns that are a few percentage points higher than its appropriate benchmark – which should be an equal weight index.  As an example, if an equal-weight index has averaged nearly 16% per year, it is reasonable to believe a Model that selects the top 10% of the same index could achieve an average annual return of 19% per year (only 3% annual increase).

In an equal-weighted index, each component stock is balanced so that even the smallest company and the largest company have an equal say in the index.  Cap-weighted indices (i.e. S&P 500 Index) weight stocks by their market capitalization, or the share price multiplied by the shares outstanding. Therefore, the market segment's biggest and most expensive stocks can dominate an index and make it "top heavy."  Because of this, cap-weighted indices may have their biggest positions in stocks that are priced near their peaks.

Winning by Losing Less

Much of the Alpha generated by models is derived from positive cumulative returns during the 2000-2002 Bear Market.  While this (again) may seem "too good to be true" it makes sense when considering the equal-weight broad market Wilshire 5000 gained 6.4% (cumulative) during this time-frame while the S&P cap-weighted index lost nearly 45% cumulatively.


No matter how good the model, at some point in time it will lose money and under-perform it's peers.  That's when having a number of low-correlated models running at once can really help.


The Wilshire 5000 Equal Weight Index was originally comprised of 5,000 stocks, but today it is made up of more than 6,700. This is one of the broadest indexes and is designed to track the overall performance of the American stock markets. Stocks of the index must meet the following criteria: (1) The companies are headquartered in the United States, (2) The stocks are actively traded on an American stock exchange, and (3) the stocks have pricing information that is widely available to the public.