Does anyone (ping: @Markku_Kurtti) have a view on what would be the most optimal metric for measuring the accuracy of target prices?
Link: measuring accuracy (MAPE)
I was considering that accuracy could be measured with at least the MAPE (Mean Absolute Percentage Error) metric. In this case, the better the accuracy, the smaller the MAPE. If the target price and the actual share price are equal, then MAPE is 0%. On the other hand, the larger the deviation between the actual outcome and the forecast, the larger the MAPE.
However, the MAPE metric has at least one weakness. It gives greater weight to situations where the actual outcome is below the forecast than to situations where the actual outcome is above it. Example:
| t | Actual | Target Price | APE |
|---|---|---|---|
| 1 | 100 | 80 | 20% |
| 2 | 60 | 80 | 33% |
| MAPE | 27% |
Target price underperformance may be more negative for an equity investor than target price overperformance. In that sense, the MAPE metric might be quite good as it is. However, I wonder if there is a metric that equally weights both target price overperformance and underperformance. In other words, is there perhaps a better suited metric for measuring accuracy?
Edit 4.11 at 2 PM. Typos corrected.