Should I Have Done the MIT Challenge Over 4 Years?

One critique I got of the MIT Challenge was that I should have taken the classes over the normal four years, instead of twelve months.

The naïve version of this critique was to argue that I should do the classes the way MIT students do—full time over a period of four years.

But this criticism doesn’t hold up to scrutiny. Of course, spending more time learning something will improve how much you learn. I could have spent fifty years on those same courses and mastered them to perfection. But then I’d have spent an extra forty nine years of my life. Opportunity costs matter, whether its an extra three years or thirty.

However, some readers pointed out a more sophisticated version of the critique. This one is based on research of the spacing effect, which more or less states that the same number of hours of studying spread out over a longer interval of time results in better retention of the material.

Under this critique, I should have done the MIT Challenge over 4 years, but only spent a quarter of the time each day working on it. Quarter-time for four years instead of full-time for one.

The Spacing Effect

The spacing effect is a fairly robust observation in learning. Studies show that spreading out the same amount of practice time over multiple sessions will improve retention compared to massed practice. I recommend reading cognitive scientist Daniel Willingham’s excellent article here for a complete summary.

Interestingly, this effect disappears when studying more complex skills. The research here is more sparse, but it seems to show that the spacing effect works better for things that need to be memorized.

To me, this seems to suggest that spacing would be a much bigger deal when studying something like languages or anatomy, which have a lot of facts to memorize, than computer science, which requires you to grok hard concepts and perform more complicated skills. Even if this is the case, however, no research I know of suggests spacing hurts retention, so it probably would still be advisable to make use of it.

If this is all true, and learning two hours a day for four years will help you retain more than eight hours per day for one, why did I run the MIT Challenge the way I did? And why run this language experiment similarly? (Especially when remembering vocabulary words seems even more susceptible to spacing effects.)

Project Constraints and a Writer’s Compromise

The most obvious reason for my choice of a compressed one-year challenge rather than a more leisurely four, was the nature of the project itself. I wanted to see if it would be possible to learn MIT’s program well enough to pass their exams in under twelve months.

Once given those constraints, I tried to use as much spacing as possible—taking 3-4 classes in parallel, rather than sequentially. Indeed, most of the research on spacing effect is over these observed time frames, so I believe that was probably a wise move.

For my current language learning experiment, the reason is even more obvious. My current method requires immersion which isn’t something I could space out without drastic alterations.

But for my MIT Challenge, a four-year spacing could have plausibly been accomplished.

The main reason I didn’t opt for a four-year spacing is that it wouldn’t have been nearly as interesting. As a writer, that’s a compromise I’m forced to make. I’m glad the project got as many people interested in self-education as it did, so my hope is that I can offer design improvements for those who want to do something similar, but aren’t necessarily under the constraint of writing and speaking about it.

Most the things I spend my time learning don’t form public projects. For those, I usually opt for greater spacing and a more leisurely pace.

Thoughts on Spaced Self-Education

For students in a classroom, the basic lesson of spacing is simple: don’t cram. Spread your studying time out evenly over the semester, and you’ll not only remember more for your exam, but you’ll remember more in your future classes which build off this one.

For autodidacts, I think there are some more interesting ways you can incorporate spacing into your life.

One of my favorite methods is to read blogs on the topic you’re interested in. Although blogging is less prestigious than authoring books, which may preclude some great authors from blogging, I find the format better for learning from someone’s expertise. A slow drip of ideas over a year helps digest a topic or philosophy better than reading about it in one big burst.

For skills, I think finding a way to inject regular practice into your everyday life is pivotal. When I returned from France, I made sure I always had a French book I could read to help maintain my French. Now that I’m taking on four more languages, I want to make sure I’m using all four at least once per week when I return.

My feeling is that the maintenance level of most skills is far less than the intensity required to reach a higher level. However, many people often don’t apply even that minimal amount and thus watch their skills and knowledge atrophy.

Because of this, I think the primary goal should be on finding enjoyable, interesting ways, to maintain your knowledge, not whatever system might be technically superior. Put the focus on making routines that involve the knowledge and skills you want to keep sharp.

Designing Your Learning Projects

Sometimes learning in a burst is the best way to go. Immersion with a language via travel, for example.

Generally, however, you’ll learn and retain more if you spread the same amount of time over a longer interval. This is certainly true for short intervals, such as spreading your studying over a term instead of right before an exam. But it’s likely also true for longer intervals, such as learning about a topic for years instead of all at once.

If you do have to learn in a burst, then I think the best compromise is to find a way to introduce the knowledge into your routine as a long-term habit. Small projects or weekly exposures to the topic can help ensure you won’t forget what you’ve learned.

  • Radhika

    For students, then, would it be better to study one subject a day or multiple subjects a day?

  • Lorena

    Very interesting. You mentioned anatomy, many people think that the med school in the US should be a longer program combining undergrad and grad as it is in many countries, instead of 2 years of hell memorizing anatomy and other difficult subjects in a short period of time.

    I think your 1 year MIT challenge was way more interesting than it would have been a 4 year do it yourself and it has been very inspiring to me. Are you as good as an MIT graduate? I do not know, and this was probably not the objective. I attend MIT events frequently and as you know MIT has wonderful labs, competitions and group activities, that maybe difficult to replicate at home and contribute to better learning. But I think you are probably better than some graduates from less rigorous schools and you prove that someone can study very difficult material at home in a short period of time.

  • Pete


    I suggest two subjects a day. One in the morning for a max of 3 1/2 hours, the second in the afternoon for a max of 3 1/2 hours.
    More than two subjects means you won’t have enough time to go deep in any one. It also wastes a lot of time continually switching mental gears, setting up & tearing down.
    Evenings are reserved for light work: making flash cards (hand-written better than computer generated), re-writing lecture notes, getting books from the library, reading first or second drafts, and physical fitness.
    All this presumes a six day “work” week. On the 7th day play.

  • Pete

    Gee, Scott,

    If you had taken four years, you could have double majored, plus picked up a couple of minors besides. Ha-ha!

  • Pau

    I think the spacing effect applies mostly to the regular studying techniques. For more advanced learning methods it doesn’t, plus they require you to study many less hours overall. I am thinking techniques like taught in Phenomenal Memory. I am not sure about Photoreading, I have heard of some good results from it, and increasing IQ with techniques like Image Streaming can be highly beneficial in cutting study and retention time.

  • Ellie

    Would it have made a difference if you did it over 4 years? Maybe. I know the courses I take have so much additional material to read that isn’t required but is helpful to learning the subject.

    Did you read all the extra materials?

    Do you think it is possible for someone to hold down a job and do this at the same time as I would really like to try it.

    Do you have any advice for whiners and complainers and people who quit too easily?

    That isn’t me but I live with someone who has just drifted through life and spends most days just wasting time online or watching tv. They say they want a career but really they try one thing like a programming tutorial and they say it is too hard and they give up after an hour.

    Would your habit changing book help as I really do think complaining and time wasting are learned bad habits?

    Oh and you languages challenge is awesome and I will follow it. I have pimsleurs Mandarin and will get started on this in addition to my other studies. One thing I found from taking years of Spanish in school is that when I went to Mexico and decided I would only speak Spanish, it was much easier to do because everyone else was.

  • MJ Bush

    “Interestingly, this effect disappears when studying more complex skills. The research here is more sparse, but it seems to show that the spacing effect works better for things that need to be memorized.”

    Could you provide a link or two on this? As a fiction writing coach, this is a point that could have a big impact on how I teach.

  • Vince Lu

    The spacing thing is something I’ve gotten familiar with. But I have some thoughts about spaced repetition for technical, complex subjects.

    Yes, it works well for languages, or anatomy, or anything similar that should require memorization. This seems to differ – on at least one level – from the rote memorization that students use when they may need to cram. I have found that trying to create Anki flashcards on math or computer science is a little tricky, because if you want to most the make of your learning by understanding concepts, and getting a good hang of how to apply them to problems, it could be a bit slow and inefficient at spacing them over a period. Learning a STEM subject, in general, is a bit like sports.

    So with that said, I still think spacing is pretty good. It could help when trying to remember how a certain theorem works, or the basics of knowing a concept. But I wouldn’t rely on that as the primary way of learning a difficult subject that requires more on web connections.

  • Williams

    I guess doing it in four years wouldn’t even have made it a “challenge” in the first place.

  • George

    Hi Scott – interesting take on spacing…. will you be testing your retention of the MIT challenge at some set time in the future?

    I would be really interested to see how much you have retained.

    As always, this was a great read, thanks.

  • ben S

    Hm. I always thought the point was that you could do the MIT level curriculum in one year, then more advanced or graduate level classes that built off of that the next, etc.

    Ed Witten is a particularly famous physicist as he’s the father of string theory. He got his undergraduate degree in history and worked as a journalist. There is not much about his background online, but from what I’ve heard, he learned extraordinarily quickly and was soon on his way to proving amazing things about quantum field theory/string theory.

    Considering that just getting accepted to Princeton for a PhD is an accomplishment, and that he did so without much of a non-math/physics background, I’ve always assumed that he was able to work with similar focus as you did in your MIT challenge.

    Note—he did have a significant background in math/physics before college.