Easy Science in Real Life: Quickest Route

[livingdeb]

There are two big north-south routes in my town. They are too wiggly to be perfectly parallel to each other. I live just east of the easternmost one, near the center of town, and my sister lives just west of the westernmost one, pretty far south.

There are also two big east-west routes, one of which is perfect for going to my sister's house but terrible for coming back because of all the construction on the northbound side of the easternmost north-south route. The other is north of my place and actually runs northwest and southeast.

The question: Is it quicker to take that northern road, going out of the way but staying out of stop-and-go traffic, or is it quicker to take 15th Street, one of the shorter routes between the two north-south roads, but with lights at most of the intersections?

I've always wondered this so I decided to--can you believe it?--pay attention. I tried both routes and timed myself and now have one good data point for each route. Traffic conditions were perfect for both routes, and I actually remembered to look at the clock both at the very beginning and at the very end.

I feel these two data points are enough to answer my question when traffic conditions are good. (Extrapolating my results to rush hour conditions would be risky.)

Oh, enough with the suspense already?

The answer: The longer northern route took 30 minutes; the other one took 22 minutes. The latter route required me to stop at two lights on 15th, but the previous time I tried, the lights were perfectly timed for me, so I'm going to assume that having only two red lights is not biasing my conclusion for that route to be too short. (I forgot to check the time when I pulled in that time, though.)

It is possible that the construction reduced the traffic density on the part of the easternmost north-south route I was using (north of the construction) enough to make it significantly faster than it would normally be. But since the difference between the two routes was so huge, I feel sure that 15th Street will remain the quicker route when construction ends.

(And then I'll test the normally perfect east-west route.)

Comments

More samples

[indigo-rose99.livejournal.com]

Two data points are generally considered insufficient. For continuous data like this, 30 data points for each route would probably do it quite well. For your purposes, you can probably get away with fewer because the consequences of being wrong are pretty low-key. Perhaps 13 each?

I would suggest that with only the two data points you have no idea of the natural variation (time-wise) for either route. So is an 8 minute difference a big deal or if you took the same route over and over again does it vary as much as 10 or 15 minutes? At this point, you don't know.

Re: More samples

[livingdeb.livejournal.com]

This is where you and I disagree. I think that sometimes two data points can be enough. I feel that I have a fairly good idea of the natural time variations. I'll just leave it at that.

Re: More samples

[(Anonymous)]

I agree, of course, with IndigoRose that your sample is inadequate for answering your question of which route is actually the fastest, but if you were merely attempting to pick a route to use on a regular basis that you find satisfactory (and part of the satisfaction comes from feeling that you've at least done a cursory comparison of noted drive times and not relied entirely on a general sense of the drive), that's a different story.

I find it difficult, a priori, to have a good sense of the variation in driving times. I have traveled enough routes where hitting the lights a certain way made a difference.

sally