We invest primarily based on our S&P 500 model. This model predicts:
- Rallies in bull markets.
- Significant corrections in bull markets.
- Bear markets and bull markets.
We go 100% long UPRO (3x ETF for the S&P 500) when the model states that the S&P is in a rally in a bull market right now. We shift to 100% cash when the model predicts a significant correction or bear market. We shift back to 100% long when the model says that the significant correction or bear market has bottomed.
The model has a 100% accuracy rate for predicting bull and bear markets. It has only failed to predict 3 significant corrections in the past 67 years. The market made a significant correction every time our model predicted one. Sometimes the significant correction began several months after our model predicted one, but the market always fell below our SELL price.
As you can see, our model focuses on the medium-long term. In the U.S. stock market, the big money is made by riding the rally waves. Our model has yielded an average annual return of just above 30% since 1950. Keep in mind that our model does not “fit” the data. Instead, we implement a wide margin of error and back up indicators in case our model is wrong.
Our discretionary overrides
From time to time we will override our quantitative model. For example, sometimes we will sell everything and switch to 100% cash even though our model says that this is still a rally within a bull market.
On balance, investing/trading completely according to our model actually results in the best performance. So why do we override our model every once in a while and accept a lower return?
We override our model when:
- The risk of a small correction is high in the short term. (Our model is unable to predict small corrections).
- The risk of our model failing to predict a significant correction is high.
Our model will only provides the best performance IF our model continues to work well in the future. But what if our model is wrong? No model is perfect. Our model is based on the past. If the S&P falls 20%, UPRO will be down 50-60%! Sitting through a 50% loss and then waiting months for the market to reclaim its old high is very difficult.
By missing out on some returns, we are also shrinking our risk and performance volatility. As Reminisces of a Stock Operator says, “the final 1/8th is the most expensive 1/8th”. Whereas purely following our model would yield on average 30%+ per year, using discretionary overrides will yield an average of 20%+ per year. Although our profits are cut by 1/3, our portfolio’s volatility goes down by 1/2!
And as our discretionary skills improve, our performance will improve as well.
In short, discretionary overrides let’s us avoid the market when the risk:reward profile for being long isn’t good.
*We still completely ignore pullbacks. Pullbacks are 100% random and unpredictable.
Our model states that the S&P 500 is in a rally within a bull market right now.