The Partial Control Fallacy

Around the time I started grad school, I ap­plied for a few pres­ti­gious fel­low­ships. Win­ning is de­ter­mined by sev­eral fac­tors. Some are just an ap­pli­ca­tion, while some have a fol­low-up in­ter­view, but the ap­pli­ca­tions all get scored on a rubric that looks roughly like this:

  • 50%: Past research

  • 30%: Let­ters of recommendation

  • 10%: Transcript

  • 10%: Per­sonal Essays

Nat­u­rally, I pro­ceeded to pour mas­sive amounts of time into the es­says, let­ting it con­sume much of my free time for the month of Oc­to­ber.

Get­ting that Fel­low­ship will re­ally help me have a suc­cess­ful grad­u­ate ca­reer. Writ­ing bet­ter es­says will help me get the Fel­low­ship. There­fore, to the ex­tent that I care about hav­ing a suc­cess­ful grad­u­ate ca­reer, I should be will­ing to work hard on those es­says.

But if the real goal is a suc­cess­ful grad­u­ate ca­reer, then at some point shouldn’t I put those es­says down and do some­thing else, like read­ing a few pa­pers or prac­tic­ing pub­lic speak­ing?

This, I dub the Par­tial Con­trol Fal­lacy. It’s where, if there’s some out­come you want, and you only con­trol a cou­ple fac­tors that af­fect that out­come, you de­cide how much to try to im­prove those fac­tors as if you were ac­tu­ally im­prov­ing the en­tire out­come. It’s closely con­nected to the 8020 prin­ci­ple: it’s when you only have con­trol over that last 20%, but you pre­tend it’s the whole thing and work on it ac­cord­ingly. It’s when the 8020 prin­ci­ple would sug­gest do­ing noth­ing at all.

Here are some more ex­am­ples:

  • Try­ing to get any com­pet­i­tive award that’s judged mostly by your past. The best col­lege ap­pli­ca­tion is stel­lar grades and some good awards, the best re­sume is a great net­work and lots of suc­cess sto­ries, and the best pitch to VCs is a rock-solid busi­ness.

  • Think­ing re­ally hard about what to say to that cute guy or girl across the room. Most of what hap­pens is de­ter­mined be­fore you open your mouth by what they’re look­ing for and whether they’re at­tracted to you.

  • Wor­ry­ing about small op­ti­miza­tions when writ­ing code, like avoid­ing copy­ing small ob­jects. Most of good perfor­mance comes from the high-level de­sign of the sys­tem.

I think I’ve been guilty of all three of these at one point or an­other. I don’t want to think about how much time I spent on my Thiel Fel­low­ship ap­pli­ca­tion and prepar­ing for my YCom­bi­na­tor in­ter­view. Mean­while, most peo­ple who get into ei­ther don’t spend much time at all.

In par­allel com­put­ing, there’s a con­cept called Am­dahl’s law. If your pro­gram takes tsteps to run, and you can make s steps faster by a fac­tor of f (say, by split­ting them across mul­ti­ple pro­ces­sors), then the new speed is t-s+s/​f, for a speedup of t/​(t-s+s/​f). There­fore, if you op­ti­mize those s steps ex­tremely hard and split them across in­finite cores, the best speedup you’ll get is t/​(t-s).

Ap­ply­ing that to the above, and you can see that, if I worked in­finitely hard on my es­says, I could only make my apps 11% bet­ter ver­sus not do­ing any­thing at all. (At least to the ex­tent that it re­ally does fol­low that rubric, even if I sub­mit a blank es­say.)

Some­times, that out­come is all you care about it, in which case you’re perfectly jus­tified in try­ing to eke out ev­ery ad­van­tage you can get. If you’re in a mas­sively com­pet­i­tive field, like sports or fi­nance, where there’s a re­ally big differ­ence be­tween be­ing #1 and be­ing #2 at some nar­row thing, then, by all means, go get that last 1%. Wake up early, get that 7th com­puter mon­i­tor, rinse your cot­tage cheese. But if you’re putting this kind of effort into some­thing be­cause it’s your ter­mi­nal goal — well, you’re not do­ing this for any­thing else, are you?

I think the solu­tion to this fal­lacy is always to think past the im­me­di­ate goal. In­stead of ask­ing “How can I get this Fel­low­ship,” ask “How can I im­prove my re­search ca­reer.” When you see the road ahead of you as just a path to your larger mis­sion, some­thing that once seemed like your only hope now be­comes one op­tion among many.