Spell To Shapeshift Into A Wolf

How to determine the quality of forecasting models by multiple

One thing that is difficult is whether the forecast of an "expert" is strong. In other words, is what this expert is really good in its ability to predict the future?

Well, it looks like that says that I speak of fortune teller. But I actually spoke to these economists and the market analysts tell us that if the economy will go well or not and whether the markets will be positive or not.

Consider the situation of an analyst that we call ABC. So ABC does not believe in equity markets in their current composition and leans on a very conservative assessment of the value of a stock. ABC works for a large company and one of his colleagues, DEF, has a totally opposite and still believes that markets are undervalued as ever. XYS is more tempered in his vision of markets and through its analytical capability, meaning it will always find the market. Finally, OPQ is the lead manager of the firm which ABC, DEF and XYZ make their recommendations. OPQ does at its head by suggesting to analysts that are important and chose his title by chance (throwing dart, dice, etc.).. It is just 1 out of 2 in its final selections no matter the market direction.

Evaluating the quality of a forecast
Can we say that 3 analysts are as good? No. But how to analyze their performance? While markets have been rising for 8 years and down for 2 years, can we say that ABC is right 20% of the time and DEF instead because 80% of the time??

Such an analysis would leave aside the material weakness ABC and DEF: there are always biased (from the top or bottom). Then we must evaluate the analyst twice. Once the bull market, and once in a bear market.

The CFA proposes to calculate the probability that the analyst is correct in bull market (PA) and it is true in bear markets (PB). Calculated once you add the two values and subtract 1. Thus, our three analysts would obtain scores:

• ABC: 0 + 1 - 1 = 0
• DEF: 1 + 0 - 1 = 0
• XYZ: 1 + 1 - 1 = 1
• OPQ: 0.5 + 0.5 - 1 = 0

While we note that the value maximum for this ratio is 1. This means that the analyst is correct every time. A quick scan of the formula allows us to understand that the minimum value is -1. However, a result should not be below 0, because that would mean that the analyst is really bad! Indeed, if an analyst is below 0, it means that the orientation by chance would be preferable in terms of accuracy.

You see that ABC and DEF are also 0? This is because they are excellent in their respective markets (up or down), they are exceedingly bad in another. Their predictive ability is nil. It is as if Cole was telling us every day that will fine tomorrow. It will be right just when it will be fine. Other days it will twist. Does she give good weather forecast? Well no!