We had a nice little snow last week. And by ‘nice little’ I mean the roads were absolute shit, traffic was at a standstill (aside from the cars sliding down inclines they were trying to stop at), and vehicles littered the sides of the road. Some amount of people had it even worse: instead of a ditch they found themselves in a major accident.
If you were an insurance adjuster you may get called in to investigate the scene of the accident due to there not being a clear at-fault driver. You’ll need to look into traffic patterns, stop lights and signs, look for tire marks, and, if you’re called early enough, sort through the wreckage before it’s even fully cleaned up.
Think Dexter, but glass instead of people, shatter instead of splatter.
Now, to figure things out you usually plug information into some calculator or spreadsheet and voila! you have a report generated and you can call it a day.
Those calculations are using good ol’ Newtonian physics. The classical theory of physics serves us quite well in situations like these. Dynamics, or Newton’s laws of motion, gives us a way to describe the moment-by-moment snapshots of what occurs in a physical system. So, given a final state (the wreckage) and enough variables (car weights, damage, roads, etc.) one can work backwards towards an initial state.
But you recently watched a dozen YouTube’s about Einstein’s theories of relativity and learned that—gasp!—Newton’s laws are in fact false!
You want to do the best job possible, right? RIGHT! So, instead of using the data on hand and plugging it into the calculators to get to the bottom of the incident, you decide to rebuild them from the ground up using Einstein’s equations. Needless to say, it’s no small task.
Days (read: weeks) go by and you’re still tuning things up. But by god you’re gonna get this right, you’ll stop at nothing—the precision of your data determines the quality of your choice of who was at fault!
But before you’re done with the work they pass the case on to an adjuster who can actually deliver something. The claim’s needs included being delivered in a reasonable time horizon wayyyyyyy before finding precision at the Nth decimal place.
You can see where this is going can’t you?
Sure, Einstein’s equations will give you a truthier result, but at what cost?
Put another way: are you optimizing for the output or the outcome?
Einsteins’s equations proved that Newton’s were in fact wrong. Yet, while living in the messy mesoscale, they do just fine as an approximation to the truth [1]. Simply: they get the job done.
The reminder then is to not be so dead-set on being slightly more right that you end up being entirely dead wrong.
(The thoughtful reader will rightly apply “…but some [models] are useful” here to flip the story and its moral when the character is designing a GPS)
[1] Checkout Popper on what he calls “verisimilitude”