Opinion: Reed Should Not Gamble

On Wednesday, September 16, a decent chunk of the senior class packed into the library lobby at noon to participate in the great and holy ceremony of the Selection of Thesis Desks. Apart from wondering who the hell all these people I’d never seen once in the last three years were, the only question I was left asking was “Why the hell are we doing this?”

Simply put, the problem of assigning seniors to their desks is an example of what is (predictably) known as an “Assignment Problem,” which is thoroughly studied in mathematics and computer science. Given that this problem (or something very akin to it) crops up frequently in the real world, it has been studied to death, and, in most situations, effectively solved. 

Reed’s current solution to this problem is to assign every eligible senior (not counting those who get fancy-schmancy offices or studios) a random lottery number, and then have students come up one at a time and select from whatever is available. The (fairly obvious) problem with Reed’s current way of doing things is that if you get unlucky and get a number way down the list, then… damn, sucks to suck. Anecdotally, even being in the top 10 is not always enough to guarantee your first choice of desk. Also, if you’re unable to hang out in the library for an hour on Wednesdays, then you better hope your proxy can accurately reflect your choices.  

Still, there is the distinct possibility that you’re lucky enough to have a friend who has a better number than you and who you want to share desks with, in which case you’re both allowed to choose desks at the same time. However, this (a) does not escape the random nature of assignment, (b) does not account for people with no friends… sucks to suck, and (c) does not consider trios or larger friend groups who would like to be near each other. 

Ultimately, with any system, there’s no way to make everyone happy. Reed’s system seems unnecessarily unfair and, frankly, a thoughtless solution to the problem considering we’re (allegedly) a pretty smart group of people. In these critiques, I mean no offense to the people who are tasked with assigning thesis desks, merely giving food for thought. 

I do not see it as unreasonably difficult to design and implement a system where students submit their preferences and an algorithm takes care of the rest. Of course, asking every student to rank every thesis desk would be overbearing, but dividing the library in ~10 zones and asking students to rank these areas. A question could also be posed as to whether students mind being near bathrooms, hallways, etc. 

In terms of getting desks near your friends, it also seems quite possible that if students would like to sit together, they would apply not as individuals but as groups (as is already the case with pairs). They would submit preferences and be treated as a single unit. All of this data could be treated via a random round robin, the Hungarian algorithm (with some modifications), or any other off the shelf method.?

The point of my argument is not to fully detail the inputs and algorithm necessary to more holistically fulfill the desires of seniors, but rather to suggest that we at least think about this as a possibility. 

I would like to acknowledge two defenses of the lottery system in its current form, starting with that of ceremony. It’s pretty fun to have everyone pile in together in the library lobby (although immunocompromised people might rightfully disagree) and get hyped about theses to come. This is a good part of our current system, but I wonder if it’s really worth keeping the old system around just for that. Could we not replicate this celebration some other way?

Secondly, one might argue that it’s best to just keep the process simple and avoid convoluted forms and black-box algorithms. I find this take unconvincing in that filling out a form with preferences is already something we do frequently (such as in getting assigned thesis advisors, in some departments, where assignments are made algorithmically if my sources are correct), and if people have no opinions, then they need not fill out the form. The “black box bad” take is an additional and reasonable complaint, but code can be made open source, and if not, it seems to me no worse than the black box of uncaring fate. 

I conclude this lengthy and self-indulgent opinion by taking a moment to rail against the other lottery-based assignment system at Reed: dorms. I mean no offense to the employees of Res-Life, but I must  ask why, in the name of all that is holy, do we have to fill out a housing preference survey if we’re just going to be given a random lottery number and have all of our opinions disregarded? Is there some way in which our preferences determine how numbers are given out? Then why not just stick straight with peoples’ preferences, and ignore the whole randomness aspect?

While room assignment has many more factors to take into account than desk assignment, in principle it's still a fairly similar solution. Our current lottery implementation of dorm assignments falls victim to all the same problems that desk assignment does, with considerably higher stakes. 

I am baffled by the current implementation of dorm assignments at Reed, and am additionally baffled by whatever algorithm was implemented in coming up with housing assignments for 2024. Anecdotally, I heard from multiple sets of people whose given preferences would’ve been better fulfilled by an easy swap of two people. Considering how impactful getting one’s choice of room can be, it seems like this problem should have some significant thought put into it. 

I assume that whoever is in charge of thesis desk and dorm assignments have already considered most of, if not all of, these points, and I additionally choose to assume that they have the student’s best interests in mind. I humbly request renewed reevaluation of our current systems in favor of a more fair, methodical approach.