Out with the old:

I pulled the WhaleIndex from the site last month to address a major flaw in the way WhaleScore filers were chosen.  I also took the opportunity to address some limitations in the WhaleScore and improve the performance of the WhaleIndex.

The old WhaleScore was limited to the 200 most popular funds currently on the site.  The problem with that is that part of the reason these funds were so popular is that their historical performance was better than other funds.  So using this method to choose the historical WhaleScore population is really a form of look-ahead bias.  Backtesting by choosing the best of the 200 most popular funds on the site naturally looked very good for past quarters, but isn’t accurate in forecasting how it would perform in the future.  And attempts to choose 200 filers by simply using past performance known at the time to be backtested resulted in poor performance. So clearly a different approach was needed.

In with the new:

The new WhaleScore doesn’t limit itself to 200 funds.  All 13F filers for a given quarter are evaluated providing they meet some basic qualifications.  The fund must have between 5 and 750 holdings, have at least 3 years history, at least $100 million in 13F assets and have at least 20% of their portfolio concentrated in their top 10 holdings.  Banks, insurance companies, trusts, and pensions are also excluded from consideration.  For Q1 2016, 2,016 filers were included in the new WhaleScore calculation.

I also re-evaluated the metrics used in scoring funds.  The old algorithm used 1 year, 3 year, 5 year, and 10 year performance along with 5 year Alpha, Sortino, turnover percentage, and top 10 holdings concentration to generate the WhaleScore.  Through extensive regression testing, I found that of those metrics, only Sortino had any significant correlation with predicting future performance.  So the old metrics were thrown out.  Through further regression analysis I identified metrics that do seem to have a strong correlation with future Alpha.  In addition to Sortino, eight additional measures (Omega, Calmar, etc.) showing strong relationships to Alpha are used.

The WhaleIndex has been revised somewhat as well.  In particular the Whale 30 and 20 are now the Whale 30 and 20 Omega and seek to maximize Alpha by taking more risk.  The Whale 30 and 20 Low Turnover have become the Whale 30 and 20 Conservative seeking to minimize turnover and volatility.

Are you done making changes?

Not likely.  I reserve the right to make changes to the way WhaleScores are calculated whenever necessary in order to improve the accuracy and performance of the model.

WhaleWisdom’s popular Excel Add-in has been updated to support Excel 2016.  The add-in is available for Windows users only.

Use the add-in to easily pull 13F data directly into Excel.

WhaleWisdom has a new hedging strategy for backtests, the Monthly Moving Average Hedge.

Now you can set a monthly moving average for your hedge to watch (the default is 10 months).  If your long portfolio’s performance drops below the moving average then your portfolio will be hedged by shorting the S&P 500.  So if you set a 25% hedge, then 25% of your long portfolio will be sold and the funds reallocated to the short whenever the hedge is in effect. Once your long portfolio’s performance goes back over the moving average, the short hedge will be removed and the funds from the short will be reallocated to your long positions.

The new hedging strategies require a pro level subscription to use

Please visit the contact form if you have any comments, questions, or suggestions for WhaleWisdom.com.

We’ve been busy building out some new enhancements to the 13F backtesting engine. Some of the improvements include:

Hedging Strategies

Now you can optionally add hedging to your backtes to protect against downside market risk.  Hedging is done by shorting the S&P 500.  Internally this is accomplished by buying shares in SH, the S&P 500 short ETF.

Hedging will assume that 100% of the backtest’s invested principal is always in the backtest’s long positions.

For example, if you choose a 25% hedging rate, then for a $10,000 portfolio, you will be long $10,000 and have a $2,500 short on the S&P 500. The short position will adjust each rebalancing period to match 25% of the total portfolio. Any gains/losses from the short are added to the overall available capital for the long positions in the portfolio.

Moving Averages

Now you can setup moving averages in the chart that is displayed after your backtest runs.  Moving averages are calculated using monthly data.  Just click the “configure Moving Averages” link to add one or more moving averages.

Sell Buffer

Now you can optionally include a stock sell buffer in your backtesting model. It works by keeping a stock in your portfolio even when it would otherwise drop out during rebalancing as long as the stock remains within your sell buffer threshold. For example, say you are backtesting with the top 20 stocks each quarter and set a sell buffer of 50. In the first quarter, AAPL is one of the top 20 stocks and is included in the portfolio. In quarter 2, AAPL drops down to number 25. Without the sell buffer, AAPL would drop out for that quarter, but since it is still within the sell buffer threshold of 50, it will remain in the portfolio. Your portfolio size remains the same. The stocks that would have fallen out stay in and the stocks that would otherwise have been added stay out. Using the sell buffer may help produce a lower turnover rate.

Rebalance Frequency

By default backtests rebalance quarterly using new 13F filing data. Now you can adjust the frequency for how often the rebalances occur.  Options include quarterly (the default), twice a year, and once a year.  Rebalancing less frequently may help lower turnover in the portfolio and possibly give you tax advantages by keeping stocks at least one year before selling.

Now you can get an email alert whenever a security you’re tracking releases a specific Form 8-K event.  Form 8-K’s are sent out by publicly traded companies within 4 days of a material events.  Events such as

• signing a material transaction agreement
• bankruptcies or receiverships
• mine shutdowns or violations of mine health and safety laws
• delisting from a stock exchange at closing
• issuing unregistered equity shares
• election or departure of directors and officers
• holding a stockholder meeting to approve the transactions
• and many more.  See the complete list at https://www.sec.gov/answers/form8k.htm

Using WhaleWisdom’s email alert system, you can choose which events (or all events) within the 8-K you wish to receive an email on.  Just go to step 3 of the email alert form and choose 8-K as one of the form types.  Once selected you will be presented with a list of events to choose from.