Saturday, May 21, 2011

Lateness Factor

So today as I was laying triple bogey in the Commerce 40 (that's stuck 3 racks) trying to hole out a 7 iron to make a 6 (which I did eventually accomplish) Danielle responded to my text saying that she would "home by 5 for sure". I like to have this information, especially these days now that we have no kitchen, couch, TV, or really internet poker to speak of to entertain me. It motivates me to play until a certain time, and gives me a sense of accomplishment when I make it. It got me thinking, though, that implicit in my digesting of that text message is a bit of translation work that I have to do whenever Danielle gives me anything resembling a completion or arrival time. This translation is basically based on the sum total of my experience over the last eight years, and is pretty hand wavy at best. There are lots of things that need to be considered, obviously, like whether or not she actually cares if she's going to be late, if she enjoys her current activity, etc etc (this is the woman who once looked at the clock and said "Oh good, I'm late I can leave now"). This got me thing if maybe we can do better, and I believe the answer is a resounding yes. I give to you now a new subjective and therefore non-sabermetric-approved statistic; Lateness Factor (LF).

Lateness factor for a given person is defined as followed. Simply ask the person's significant other (or if single, best friend) to translate various arrival time predictions and apply the following rules:

If they translate "definitely"or "for sure" or "will be" to "probably", add a point
If they translate "probably" to "maybe" or "might be", add a point
For every half hour of lateness they assume, add a point

The sum of this scoring system is the Lateness Factor. For example, I've decided Danielle's current lateness factor is about 2 (she used to be a zero, for the record...much like athletic performance, I predict lateness factor will change for the worse with age), which means her text of "I'll definitely be home by 5" could translate into any of the following:

1. I'll definitely be home by 6
2. I'll probably be home by 5:30
3. I might be home by 5

Obviously the higher the lateness factor, the more possible translations exist, but practically speaking the statistic really only has meaning from values of about negative one to three (the big potato almost certainly has a negative lateness factor; if I schedule a lunch for 12pm, he will definitely be there by 11:30). Values outside of that range are pretty much useless, as they indicate a basic disdain for any sort of scheduling whatsoever.

Mildly Related Tangent

Most people use the commonly accepted scoring system of 1 to 10 to gauge a person's attractiveness (the poster formerly known as cgrohman prefers a binary 1 or 0 scoring system, but many feel this is short sighted). What does this mean? People aren't really sure, but here is my interpretation, specifically that it is a log scale:


These people are merely mildly pleasing or displeasing to the eyes. They are in short average.


These people are "1 in 10s", meaning that saying someone is a seven means that roughly speaking if you saw 10 other people on the street, you'd expect one of them to be about as attractive as said person (just the opposite for a 4).


Same deal, except 1 in a 100. So someone who rates an 8 is reasonably likely to be the most attractive person of a group of 100. A 3 would likely be the homeliest.


Same deal, but 1 in a 1,000.


Same deal, but 1 in 10,000. In short, at a large university like, say, Penn State, with 30K undergrads, there should only be 1 or 2 women you'd rate as a 10.

1 comment:

ExMember said...

A couple of years ago I suggested abandoning the 1-10 scale for the simpler and more useful number of standard deviations away from average.

Let the population be 18-40 year-olds of your preferred gender from first-world countries.

Zero is average. +1 is about the threshold for me to be interested. +3 is what a supermodel should be.

Note that despite the fact that it is impossible to quantify attractiveness, it is possible to measure where someone fits on this scale via a well ordering of the population (or sample thereof).

Also note one-point or half-point levels of precision are wide enough to be easily agreed upon, but narrow enough to tell you everything you need to know about how attractive the person is.

You can also do interesting things like redefine the population. For example, if you limit it to female poker players, you'll find some people who would score in the -0.5 to 1 range, suddenly jump to the 1-2 range which explains the attention they subsequently receive.