Recently, I had a chance to share with my son some ideas and comments about how testing is less about whether something passes or fails, but is about analysis of information and helping to make a qualified decision based on experimentation and feedback.
One of the games I love playing with my kids is “Armchair Economist”. I’m a bit of a geek when it comes to experiments, continuous learning, and interesting discussions, as well as ways of looking at problems that might have slightly ambiguous answers. This is one of those times that playing “armchair economist” led to an interesting discussion about testing.
For years, my commute was an easy thing to deal with and calculate. I worked in San Francisco, which meant, basically, two things. Either I did a rail commute, or I drove. Both have their positives and negatives, but for the most part, driving was just not a reasonable solution because traffic and parking prices made it genuinely unpalatable. Since I had a meeting I had to be at each morning that started at the same time, commute decisions were easy. Go to San Bruno Caltrain station, get on train at the appropriate time, get to meeting on time. No fuss.
When I changed jobs, and made the shift to working in Palo Alto, plus the fact that where I was working was very close to the CalTrain station, again, that was a no-brainer most days (because of the way that our company works with the City of Palo Alto, I do have a parking permit that gives me “free” parking in the Civic Center garage). Sometimes I have things going on that require me to drive in, but most of the time, train still makes the most sense.
Pat of this change-up involved me changing my schedule so I could drive my kids to school. I’d drop them off, and I’d go on to the next station (in Millbrae) rather than back track and go to the one in San Bruno.
Now that my son has a driver’s license, and a car of his own to drive, he’s now handling the “driving of the kids to school” routine. He’s also become very “price conscious” about how much fuel costs, and more specifically, just how much traveling and mobility that fuel actually gets him. Part of this was prompted by the fact that, since he’s received his licence, he would drive me to the train station and drop me off on various days, and as part of this (including being my ride home) he was wondering how much it was costing him to “ferry me to the train station and back”.
This prompted a discussion… “Hey, Dad, how much does your monthly commute cost you?”
I showed him the price tables, including some of the additional costs of parking, and at first, we both agreed, it looked like going to Millbrae over San Bruno is the better deal… but is it, really?
With that, I said “hey, let’s be ‘testerly‘ about this. Let’s build a model.”
We sat down and we started with the obvious comparisons.
I live in San Bruno.
I work in Palo Alto.
Caltrain divides its fare rate into zones, instead of an origin/destination price model.
Travel from San Bruno to Palo Alto covers three transit zones ($179/mo. if you use the Clipper card)
Travel from Millbrae to Palo Alto cover two transit zones ($126/mo. if you use the Clipper card)
Three zones is more expensive than two zones, $53 per month more.
OK, so it’s cheaper to travel from Millbrae than San Bruno. We’re done, right? Well, not so fast…
Let’s consider the parking situation. In San Bruno, although it can be hit and miss, and you may need to walk a bit farther on some days compared to others, it’s entirely possible to park on Huntington Avenue for free. It’s so much more possible that the San Bruno CalTrain lot usually only has a handful of cars in it on any given day.
Cost for paid Caltrain parking? $5/day, or $50/month with parking permit, doesn’t matter the station.
OK, so what’s the option for free parking in Millbrae? Actually, there’s very little in the way of nearby/available free parking. The side streets have strict time limits or “no parking” policies. The closest “Free” parking is about a half-mile away. This is reflected in the parking levels in the Millbrae lot. By 9:00 a.m on any given weekday morning, the lot is completely full, out to the furthest parking spots.
Translation: a $50 premium added to the price of the commute. So the difference now is just $3/month in Millbrae’s favor.
Hmmm… here’s a thought. What’s the difference between driving to San Bruno Caltrain vs. driving to Millbrae Caltrain?
San Bruno: 4.2 miles (round trip)
Millbrae: 8.4 miles (round trip)
When I was driving my kids to school, that didn’t make as much of a difference, since their high school was close to the Millbrae station, and I’d be going out that way anyway. It didn’t make an sense to double back. Now? It’s completely my choice, so I’m choosing to drive double the distance. I drive a traditional car, a 2001 Ford Escape 4WD w/ a V6 engine. Nice vehicle for off road play, camping and snowboarding, but it’s strictly average when it comes to driving around town, to the tune of about 15-20 miles per gallon. To be conservative, let’s use a 15 MPG value. If I were to take an average of 23 days for commuting in a given month, it’s 97 miles total per month to the San Bruno station, 193 miles total per month to the Millbrae station. Gas right now is $3.99/gallon, and at 6.44 and 12.88 gallons respectively, the cost for gas (specific to my commute and nothing else) is $25.69 for San Bruno, $51.39 to Millbrae.
So where does that put us?
– San Bruno to Palo Alto Total Cost per Month: $204.69
– Millbrae to Palo Alto Total Cost per Month: $227.39
Net difference? $22.70 per month extra to use Millbrae instead of San Bruno.
This is where I paused, and did my classic testers smirk and said “so, are we done?”
My son said “yeah, it actually costs less to commute from San Bruno, so you should commute from San Bruno.”
Testers, I see you smirking (well, i don’t see you smirking, but I know you are 😉 ).
“Yes, on a pure dollar basis, it would make sense to say “Commute from San Bruno, you’ll save money”. Is it really that simple, though?”
My son, to his credit, knows that, whenever I throw out the “Is it really that simple” comment, or something to that effect, that I’m about to lead him on a bit of a Socratic adventure. To
his credit, he hasn’t gotten to the point of rolling his eyes and walking away when I do this. Today, he didn’t disappoint.
“Well, it would make sense if everything were the same, right? If every train stopped at each station, and if the schedule were the same, then the cost savings would make sense… but Caltrain doesn’t work like that”:
– None of the “baby bullets” stop at San Bruno.
– In fact, there’s a lot fewer trains that stop at San Bruno.
– Also, trains that stop at San Bruno are either full stop trains or on limited stops.
– An average commute out of San Bruno to Palo Alto will take about 35-40 minutes (33 minutes being the fastest option on one specific train).
– The fastest commute out of Millbrae will take less than 25 minutes
– Even if we don’t catch the baby bullet trains, there’s more trains that stop at Millbrae than San Bruno in any given hour, about three times as many.
Here’s where I smile a bit, since I see he’s considering something. I leaned in and said “ok, so now that you’ve said that… which is the better value for commuting?”
As expected, he thought about this for a while, and said “well, it depends on what matters more to you. If you care about the total cost, then commuting from San Bruno (and parking on the street for free) will be the better choice. If you care about flexibility, then commuting from Millbrae would make more sense. The difference is $23, or one dollar a commute day. If, in my schedule, I had to change up my routine for some reason, or needed to get there at a specific time, then sure, it would cost me $5 to park in Millbrae on that day, but I could still do that 4 times and come out ahead. $2.70, in fact. If I had to vary it more than 5 times, then commuting that month out of Millbrae and buying a monthly parking pass would be a better deal.”
I stopped the discussion at this point. I told him that, if we wanted to, we could go into serious minutiae and get measurements under various conditions to see total time performance (door to door, how long it took to walk to the platform from where we parked, total time difference on an average vs optimized schedule, etc.) but that this exploration was pretty good in and of itself. It gave us a lot of good information.
What was interesting was how he saw that the dollar amount, which would have been easy to quantify in isolation, masked a lot of nuance of the situation. I explained to him that the nuance is where a real thinking mind needs to make an analysis, a total calculus of the whole situation, and then make a judgment call based on all the information. As he learned, it’s not all cut and dry. Additionally, there’s a new construction project happening in San Bruno, and the station is going to be changing location in a few months. At that point, the “Free Parking” option may no longer be a viable option, and we’ll have to reconsider this whole exercise again at that point.
So, what do you and your kids talk about ;)?