Before I take your argument apart piece by piece let me start by saying you have chosen a terrible visualization to represent a polynomial regression. You should be using scatter plots with a line of best fit.
Now let me ask, where exactly did you get some of those numbers from? They don’t exactly relate to the single source link you posted.
Now going back to the point on your terrible visualization - I bet you chose this because you don’t have a proper scatter plot of data points. Once again you are working with average values whose distributions you do not know of. This puts your entire model under question. What makes this worse is that you essentially have 8 data points and are making a generalization on an entire population for that. Of course, these 8 data points are averages, but once again, you don’t have a single clue about their distribution or any other assumptions for that matter. You need SIGNIFICANTLY more information than what you are outputting right now to draw any remotely objective conclusions.
Once again, you are using what appears to be 3 DATA POINTS and then applying a line of fit on them. This is so laughable that the only people you could convince are those who have 0 background in statistics. As someone who KNOWS statistics, let me tell you: you are TERRIBLE at using statistics.
You know, let’s say you ACTUALLY had a valid model, large working data set with assumptions filled out, and proper statistical analysis. THEN I would be more on board with what you have to say. However, not only are you absolutely terrible at using statistics, once again you fail to address all the other surrounding factors that lead to the definition of skill.
So I’ll say it once again: CORRELATION DOES NOT IMPLY CAUSATION.
Your entire thread is thus invalid. Come back with a better paper.