A straight line is the shortest distance between two points i.e. but only on paper. Once we leave paper and two dimensions, however, things get more complicated. Theoretically a straight line projects in both directions to infinity—which is much more than we are ever prepared to prove. Where straight lines lead is “o-f-f t-h-e c-h-a-r-t-s” and nothing is ever so pure as a vector that continues to infinity—no matter what (else). What we think of as “straight lines” such as signature lines, etc. are only provisional straight line because they are contained, bounded, and limited.
A bicycle may surpass walking for efficiency, but the bike path is hardly a straight line. A rocket is among the fastest of man made vehicles, but it takes hundreds of millions of dollars and years of planning to launch, even if once it moves it beats all comers. Perhaps we can see "straight” to a particular object, but that hardly takes us there. So much for the usefulness of straight lines.
With two points, we can project a straight line through them. With three, we can run the circumference of a circle through them and a straight line if the are so aligned as well as any number of complicated curves. But more often than not, scientists, economists and data technicians often notice that scatter patterns strongly resembling linear formation so that they invent the tale that the "Straight line” says it all (onward to infinity; which can only be supposed since infinity is only a theoretical concept). Economists and prognosticators project straight lines and thus they project a line through the midst of a conflicting data as if the projection could and should speak past all dissension and non conforming data. All discrepancies are “thrown to the winds” in favor of the projected straight line (which inevitably turns out to be wrong, errant and a false prediction). But if one assumes for an instant that that is the way reality works and that the straight line will “hold true” for all values of X and Y, we are hopelessly naive. All sorts of trends that seem to “hold” to a particular straight line, are amazingly ephemeral subject to a host of factors (variables), but often "skew” off one way or another as we employ very low and very high values for X and Y variables. (A bullet fired from a rifle tends to travel more or less straight until we notice gravitation “hugging” and “coaxing” it earthwardly and bending the trajectory downward into a more or less parabolic curve. A breeze can similarly deflect the path sideways by some vector. The spin can buffer other forces, but if the bullet tumbles in mid air, that too will impact on trajectory.)
The point here is that economists and data prognosticators are so naive that they render themselves silly with their projections. These people speak of trends such as crime increasing three times as fast as the population. But while that might have held true for certain enumerated crimes of particular sorts for perhaps a year or two, looking over a decade, two decades of century, such trends totally lose meaning and become charades and more memorials to our naivete and ignorance than anything else.
Indeed, in most cases, straight lines are silly, foolish, naive, even dangerous and often trouble makers.