Ok,
rant mode on
You'll have to excuse me here, I really hate when people bend statistical methods to their own sick ends. It contributes to the continuing erosion of the reputation of science because the interpretation of scientific data is inescapably entwined with a faith in statistics. When people abuse statistics they're abusing science. It is truly unfortunate that the accuracy of any statistical interpretation depends so much on the impartiality of the practitioner of statistics.
There are several things amiss here. Now I'm not saying he's being deceptive, he's probably just ignorant. But it is one of the two.
#1 is this graph the result of regression?
If it is it looks like wildhagen used a linear model for all but DPC for which he used some type of power model. This could belie some sort of deception. If it isn't the result of regression then theprojection into the future is mere artwork. So its likely that regression is involved. Wildhagen should have stated what models were used and why.
#2 Regression analysis is used as a predictor of future outcomes. It cannot be used to predict outcomes for events that were not a part of the random sample. Future dates cannot be sampled therefore use of the regression to predict the future is invalid. I understand the need to project the overtaking of a competitor and can overlook this problem. But we all know that an assumption is made when we use regression analysis to project something. In this case the assumption is that production rates will remain unchanged with time. We all know that to be false. There is no statistical method to validly do this.
#3 Wildhagen should have included a scatter plot of the data overlaid on the regression model as well as an indicator of fit. This is so that we can see how well the model fits. This graph could be an invalid fit. It is very difficult sometimes to find the best model for a given set of data. Excel will fit the data to a model irregardless of whether the model is appropriate. By showing some sort of indicator of fit the author is demonstrating that he has ascertained the proper model to use. There are no computer programs that can do this.
rant mode off
Just some items to think about when you look at that graph. Hope its been helpful. I would also like to state that this is in no way an attack on wildhagen himself. I just tell the truth, I can't help it, if it stings a little. Better yet consider it a challenge to his statistical ability.