Quick update on the performance of the forecast model:
- I’ve increased the spread to 1.25 std so that the close forecast is falling within the forecast high and low range 85% of the time.
- There’s still large dips (as you will see in the precision measure) because I haven’t plugged the recession forecasts yet
- My goal here is to reduce the std spread to 1 std and below while maintaining a 80%+ accuracy. As long the recession plug is working, this might get there
- Precision needs a lot of work, but I’ll need to think of a way to improve the accuracy without running the risk of overfitting
- Right now, I’m on average 50% off when it comes to forecast price. In the long run, it’s not too bad as long I’m buying the near lows and selling near the highs.
- I want to reduce the forecast error or improve precision to about +/- 25%
- Right now the model is strictly looking at fundamentals. I think adding more layers into the forecast model should continue to better the forecast
- Fundamentally, the north star should be whether a stock is more likely to beat the market next year. Will need to build this metric out and will provide an update when I do
- Also, more analysis needs to be done to see where the forecast errors is occurring. Need to narrow down the pattern to see if it’s coming from a specific sector, subsector, market cap etc…