Loading ...As you may have seen, I write a weekly technical analysis summary for the stock market. At the end, I calculate an average rating for that week between 1 (most bearish) to 3 (most bullish). A while back, I started tracking the performance of the S&P 500 index over the 13 weeks after each of my issued average ratings. This allowed me to see if there was a relationship between my rating and the 13-week performance of the index. More interestingly, it helped me make an experimental forecast for where the S&P 500 index would be 13-weeks after each of my weekly articles. But the results were very disappointing. The average rating alone was not useful. So I decided to go for something a little more robust.
I am now testing two models at the same time. One is a 13-week forecast using the individual components of my technical analysis summary weekly articles (BarChart.com, CXOAdvisory.com, StockTA.com, and 200-day moving average from Finviz.com). Those four components are specific to about a 13-week time frame. I will calculate a linear regression model using the actual values from each of those four sources, not just a 1, 2, or 3 as a rating. The other model is a 4-week forecast using four sources with a shorter time frame (BarChart.com, StockTA.com, PowerRatings from TradingMarkets.com, 20-day moving average from Finviz.com). Again, I will calculate a linear regression model using the actual values. The BarChart.com and StockTA.com are different for the two models. I use the “long term” analysis for the 13-week model and the “short term” analysis for the 4-week model.
Right now, this is still early in the data gathering process. But I thought there might be some people out there who, like myself, enjoy watching forecasting models as they evolve from their very early forms into more accurate ones. So, I’ll go ahead and begin publishing the forecasts that are derived from my analysis on this page.
S&P 500 Index Forecast
Using a 13-week model and a 4-week model, here is the forecast for the S&P 500 index at the close of each week for both models. The two lines without points represent the overlap of the confidence 95% confidence interval or the two models. In other words, the area between those two lines represents the area that both models are reasonably certain the S&P 500 will close on that date:
Stock Market Forecast Archive and Buy-Sell Simulation
Once a forecast has been issued, that forecast might change a little each week as new data comes in. That’s because every new week of data will adjust the models slightly. Therefore, I will keep an archive of every forecast made, including when that forecast was made.
Also, if you click the “Buy-Sell Simulation” tab below, you will see a chart that shows what would happen to $10,000 invested each week into the S&P 500 index based on whether my weighted combo forecast is predicting a rise or fall in the S&P 500 the following week. For the sake of practicality, the simulation only buys or short sells if the index actually hits the signal level the following week. In other words, if it takes off without ever giving you a chance to buy at the level that the signal said you should buy, the simulation won’t buy – just as you couldn’t. Also note that the system will not buy or sell if the price is even better than it should be without hitting the actual price. Example: If the signal says to buy at 1,300 but the S&P 500 starts the week by opening down around 1,290 and never recovers, this simulation won’t buy because it never hit 1,300. There’s two reasons I do this: 1) Practicality, so I don’t have to find a custom price point at which to enter each week and 2) This could be a way of insuring against an unusual external event that occurs between Fri and Mon, like a big crisis development overseas that couldn’t have been picked up by the technical analysis data that went into Friday’s signal. The column on the far right is how the simulation is doing versus a buy-and-hold strategy.
For better viewing of the archives of forecasts, here are the column headings to help you if you scroll down (I can’t get it to freeze rows when I publish the spreadsheet):
Disclaimer: All of the information provided at Easynomics is for informational and educational purposes only and should not be construed as financial, legal, or any other kind of advice. Per FTC guidelines, this website may be compensated by companies mentioned on the site. My reviews are my opinion and positive reviews are not for sale. References to 3rd party sites are not updated daily so please visit the official site for accurate information.
