Game Theory Coffee Filter

May 22, 2014 8:42 PM

On weekdays, I get my morning beverage from one of the many coffee carts dotted around my office neighborhood. Most frequently, I go to either Cart A or Cart B. Cart A is across the street from the bus stop where I get off for work. A is open weekdays from 7am to 3pm. Cart B is conveniently located directly outside of my office, open 7am to 4pm, but only on the 3 or 4 days a week when the owner feels like coming to work, like when it's not raining or too hot or Thursday or whatever. The menus are identical, and the service is unremarkable. Given that my goal is to get my morning jolt with as little hassle as possible, which one should I give my business to?

I can't see from the bus stop whether B is there on any given day. So every morning I get off the bus, and I have to make a decision: do I cross the street and go out of my way to get my drink now at Cart A, then cross back and go to my office? Or do I walk past that cart toward my office until I turn the corner and can see whether or not Cart B is there, and then go to B on the way into the office if it's there, but have to walk back around the corner and across the street to A if B is not there?

Basically, to the extent that my tiny purchases have any effect on the proprietors' behavior, I'd like to use my buying to reward Cart A for being there every day, rain or shine, and incentivize Cart B to come out more often and/or punish them for being lazy. And I'd like to have things be convenient in the morning when I'm grumpy.

Is the best way to do that to use Cart B every single time it's open, so that the proprietors know it's profitable and will maybe want to come out more? Or should I use A most of the time, both so that I don't have to double back on the days when B turns out not to be there and to punish B's flakiness, and only use B when I'm coming out of my office in the afternoon for a pick-me-up?

This is a ridiculous question. But I think about it a lot. And I want to know what the best strategy is, assuming I have some tiny, marginal effect on the businesses I patronize. So please no "it doesn't matter" answers. Thanks!

posted by decathecting to Grab Bag (18 answers total)

In my mind, Cart A is the path of least resistance. You may have to cross the street every day, but you also know that it's always there.
posted by Ruki at 8:46 PM on May 22, 2014 [1 favorite]

Coffee *always* matters.

I would use A every time, but that's because I'd be angry having to double back.

Optimize for the shortest distance between myself and the next coffee.
posted by parki at 8:51 PM on May 22, 2014

An unintended consequence of rewarding B with your business is that you may actually encourage the coffee cart operator to show up less often. Presumably he's happy with the income earned now, or he would work more days. If everyone who rides the bus and works in your building makes the same decision to frequent B, he will earn the same amount of money faster and thus spend fewer days on site.
posted by carmicha at 8:56 PM on May 22, 2014 [2 favorites]

I think that you are an epsilon term and should act accordingly.
posted by zscore at 8:58 PM on May 22, 2014 [3 favorites]

Go tell the owner of Coffee Cart B, what you said above. But shorter. Basically, that you would use them consistently, but you never know when they aren't going to be there, which makes you want to use the other cart.
Tada!


If they may be able to tell you that actually, they will be there on MTW, or everything but the first week of the month, so, then use it those days. If they don't appear on the day they promised they would be there?
Switch that day permanently to Coffee cart A.
If they warn you they will be away, you can stay with Cart B.

Basically, if they keep flaking out on you, you'll end up with Cart A. If they make an effort to be predictable, you can keep using B.

If Cart A is ever closed, you'll check in with B, but you're basically rewarding professionalism.
posted by Elysum at 9:01 PM on May 22, 2014 [7 favorites]

Under my scenario above, A becomes deprived of sufficient business to make a living and departs for another site... and now your supply of sweet delicious morning goodness is really threatened. Unintended consequences are a bitch.
posted by carmicha at 9:04 PM on May 22, 2014

Riffing off my previous post, assuming that you can't affect the behavior of Coffee Shop B, what you actually need to do is to create a model that predicts the likelihood that B will be open on a given day.
posted by zscore at 9:08 PM on May 22, 2014

Did I misunderstand the location of Carts A and B? It seemed like you said you take the bus to work, and when you get off the bus, Cart A is right there (across the street) while Cart B is at your office (around a corner, some distance away). Why would you walk to Cart B only to find out it's not there/open/whatever and then have to go BACK to Cart A, like 30+% of the time? Go with Cart A every time, unless this street crossing is some herculean task.
posted by axiom at 9:14 PM on May 22, 2014 [1 favorite]

I think you're disregarding the most important factor here, which is that cart A gets a cup of coffee in your hand 30 seconds earlier (I'm guessing) than cart B.
posted by hydrophonic at 9:27 PM on May 22, 2014 [7 favorites]

I am no good at game theory, but if this were my scenario, I would go with Cart A every time to reduce the overall goddamn aggro of a pre-coffee morning commute. The idea of doubling back for coffee annoys me just thinking about it. In fact, I'd probably relish the fact that dumb old unreliable Cart B wasn't ever gonna make a cent off of me, while drinking delicious dependable Cart A on my way to work.

However, Elysum's idea of asking Cart B if they have a schedule (in my mind: really suggesting that they most certainly should) also appeals.
posted by juliplease at 10:27 PM on May 22, 2014 [1 favorite]

How long does street crossing take? How long does doubling back take? Approximately how often is B not there? To me this is a math question.

If [street cross time] < [percentage of times B isn't there] × [double back time] then go with A every time.

If the two sides of that equal are equal, then decide if you like variability in your life and the exciting chance to "come out ahead" or predictability and the knowledge that you won't have to double back. Me, I'd take A every time.

If what you want is for B to be there every day, skip this guessing game and just ask the cart worker why they're there sometimes and "if I came every day you're here, would that make you more or less likely to come every day?"
posted by salvia at 12:31 AM on May 23, 2014 [1 favorite]

What you need to do is build a little electronic gadget that detects light, and place it somewhere so that the coffee cart B casts a shadow on it at whatever time you arrive in the morning. Set it to email you if it is in the dark at that time. You could build this with an arduino. (But I guess it would probably get stolen. But it would be SO cool.)
posted by lollusc at 1:07 AM on May 23, 2014 [3 favorites]

I don't suppose you can set up a Dropcam or similar so you can check on Cart B with a smartphone.

I think this problem will yield to a analysis using elementary game theory. You have two plays: check B or don't check B. Coffee World has two plays: B is there or B is not there. You have to put a numerical value on your satisfaction for each of the four possible combinations and do a little arithmetic.
posted by SemiSalt at 5:01 AM on May 23, 2014

Do you know why Cart B is intermittent or are you assuming? If you don't know then ask and suggest they attach a schedule of their open days to the cart. Then decide, game theory is critically dependent on the nature and quality of your assumptions and the accuracy of your modelling.
posted by epo at 5:57 AM on May 23, 2014

In my city a significant number of food trucks have twitter feeds that post their daily specials, etc. Granted, these are the businesses that have it so together that they can manage a varying menu and a twitter feed, and it sounds like your CartB owner is challenged on the "go to work every day" aspect of being a business owner. But, especially if CartB is a youngish type rather than a retiree, maybe they'd be interested.

Take a mid-morning coffee break sometime (i.e. not during breakfast or lunch rush) and go out to CartB to plead your case. Point at your office window and tell him how convenient it would be, tell him that frequently you already have your coffee by the time you get to his cart because you assume he'd be closed. "So I'm sure there's a system that I just don't understand, can you explain to me which mornings you're open?" Tell him how awesome a twitter feed would be, or ask if there's a number you can call. Basically gush about how much you want to buy your coffee from him every day, and see if he can suggest a way to make that happen.
posted by aimedwander at 6:49 AM on May 23, 2014

Basically, to the extent that my tiny purchases have any effect on the proprietors' behavior, I'd like to use my buying to reward Cart A for being there every day, rain or shine, and incentivize Cart B to come out more often and/or punish them for being lazy. And I'd like to have things be convenient in the morning when I'm grumpy.

Clearly you aren't French.

I have lots of non-monetary solutions "out-of-the-coffee-cup" solutions:
- I would talk to them and find out their story. So do you have a schedule? Wife & kids?
- Do you tip? You can tip B when they're there and say "I'm glad you were here when I needed you!"
- Ask them to put out a sign around the corner (when you first get off) so you can see if they're there.
- walk past B drinking A's cup "oh man I wish I knew you were here!"

My favorite:

- give B your ph# and tell them to text you in the morning because you're lazy and will buy it at A if they don't tell you otherwise
posted by St. Peepsburg at 6:52 AM on May 23, 2014 [5 favorites]

I think that the "correct" solution to the problem requires knowing the probability of Cart B being there on a particular day. If it's only there, say, 25% of the time, that's different than if it's there 90% of the time. And the precise number of extra steps you'd have to take if you went for Cart B, found it to be closed, and had to double back would also matter — if your goal is to minimize the number of steps (or time) that you take getting coffee.

Personally, I'd rather go for consistency and also support the cart that's always open every day. And hey, an extra few steps could be viewed as a good thing anyway.
posted by Kadin2048 at 8:55 AM on May 23, 2014

Cart A is the critical morning coffee cart. Assume cart B is not there for this purpose.

Cart B is the occasional mid-day spontaneous man, I should take a break and get a coffee to clear my head cart. If cart B is even there. Unreliable supply earns unreliable demand. Universe in balance.
posted by ctmf at 10:24 AM on May 24, 2014 [1 favorite]