Today is the day I make changes to the Quant, which is a monthly occurrence, but also change the Bubble Basket and the GARP portfolio. The Bubble Basket is a managed portfolio of the most overvalued stocks in America — broken down cleanly by my methods and diversified amongst all sectors. They’re supposed to be dangerous stocks — but they tend to do best in bull markets.
YTD the Bubble Basket is +41%
The GAPR is the Quant, but only the value aspect of it. No hedges or protection — just alpha.
YTD the GAPR is +23%.
My Quant was switched around today and plenty of changes were made. I sold my bonds and doubled up on gold and we’re 100% in value stocks now. Bear in mind, this is totally a mechanical process, so none of my real time opinions here are relevant.
Last month the Quant rose by another 3.5%.
Here are the returns YTD — by month.
Jan: +6.6%
Feb: +7.2%
March: +2.6%
April: -0.2%
May: +0.6%
June: +3.5%
Granted, it has been a very easy year to make money in the market. But the Quant keeps me honest and regimented. I feel a certain sense of normalcy by keeping it going, knowing the methods have been tried and tested and they work. The stated goal here, naturally, is to beat the SPY. But the bigger plan here is to find a repeatable process from which I can package and deliver to my kids and say “you want to know how to invest your money? Here you go.”
A best ideas approach is too esoteric and entails an artistic approach that is not repeatable. Indexing is fucking boring. In my opinion, this is the only way to structure a long term portfolio and feel confident that you’re actually working towards something meaningful.
I have a meeting now. If you have any questions, Exodus member of not, feel free to fire away in the comments.
If you enjoy the content at iBankCoin, please follow us on Twitter
Are you going to be incorporating machine learning or neural networks?
Any way to use your tools and go back in time to simulate your results?
No backtesting abilities yet. We are rewriting whole platform, so maybe in future.
I know you are on it. I know a guy if you want to do machine learning routes
Machine learning is often like giving a 16-yo the keys to a Corvette: without the understanding and respect for the power and the accident potenetial, disaster is a high-probability.
The issue is that machine learning is just a fancy way of saying pattern-matching, except that often the logic of the pattern is lost (depending on the algorithm). Neural networks are a complelte black box, so you have to remember GIGO. The key assumption is that the future will behave like the past: which is the one thing that Investment advisors can’t explicitly satte to their clients – even while creating portfolios based on the opposite.
I wouldn’t use ML for anythign othere than long-term trend-spotting, eg, are we in bull market or just a bear-market rally? You can read “Components of bull and bear markets:
bull corrections and bear rallies” (March 2010) by Maheu, McCurdy, and Song to see what I mean. Don’t worry, Ferd, that paper is spell-checked so you should have no trouble understadning it.