I'm not an economist, but I suspect it's probably wrong and useless. The idea that the entire tax code and it's effect on the economy (and therefore the revenue taxes bring in) can be put into a neat little curve is preposterous.
You really don’t understand mathematics? At the end of the day, quantifiable data has correlation and causation. The curve is not comparative from year to year: it is relative to one budget year only, and as such is completely accurate due to the fact that it lacks specific numbers and is based upon a relatively similar economic situation regardless of taxes.