…or when to stop counting rivets!
Tony Koester of MR/RMC fame has often spoken of a so-called “good enough” philosophy (citing V&O creator Allen McClelland as the source) for model railroading. In other words, for each one of us, there’s a point at which a model or scene is “good enough.”
I thought about this as I tried to codify my own threshold of “good enough,” and it began to remind me of differential equations from calculus. “Good enough” is nothing more than a unique solution to an initial value problem.
Let’s call the solution the “sweet spot.” This solution is the point at which the amount of work required to make a model more accurate exceeds the fun the modeler would have in doing so. So, let’s define two curves:
The red dashed line represents fun (scaled on a dimensionless, normalized range between zero and unity) and the solid black line represents work (scaled the same way). As a store-bought or scratch-built model becomes more and more accurate, it requires more and more exacting, tedious, and time-consuming work to accomplish. Theoretically, the amount of fun a modeler is having is simultaneously decreasing (i.e., the law of diminishing returns). The “sweet spot” is that level of accuracy whereby the modeler is still having fun but working hard to accomplish his goal; any more work and it stops being fun. Notice the curves are asymptotic; no model can ever achieve 100% prototype accuracy.
What makes this an initial value problem (i.e., the sweet spot is a unique solution to a very specific set of circumstances) is that the slope of these curves varies greatly from modeler to modeler, and from project to project. In other words, the skills, desires, and patience of the modeler affect the sweet spot location as does the choice of prototype, starting model (if applicable), availability of after-market details, paint, decals, photos, diagrams, etc.
So “good enough” is a constantly moving target; the reference frame is always changing. Perhaps quantum mechanics is a better context than math? You decide. But the sensitivity to initial conditions reminds me very much of partial differential equations, where each variable is dependent upon the others and a minor change in the choice of prototype or starting point yields a vastly different version of “good enough.”