Heuristics to steer by in study exploration

When I’m getting oriented in a domain, most of what I’m doing is figuring out where and how to invest my attention and effort. 

After a while, I’ll get the hang of it, such that I feel like I can reliably sit down and turn time and attention into progress towards my learning goal. But before I reach that point, I’m exploring (read: flailing around), trying to get a foothold. (This is very related to getting “hooked in.”)

The following are the heuristics that I’m currently using to steer that exploration process, in order of application. That is, the first one takes precedence over the second and so on.

Follow the hope

Often, when I’m trying to learn something new, it feels daunting. In fact, “daunting” doesn’t really cover it. I feel hopeless despair: the topic is huge, and there’s so much of it that I have to learn, and its going so slowly, and I don’t trust my futureself to do enough to get it to every pay off. 

When I feel like this, I want to follow the hope. That is, I’ll consider, and maybe try, several approaches, paying attention to if any arouse a slight glimmer of hope, a subverbal sense of “oh, I that might lead to progress”, that you can sort of sim following that path to your goal.

When you notice that flicker of hope, try the action that inspired it. Use your sense of hope as the heuristic function guiding your exploration.

Steer toward difficulty / intensity

But once you have a little bit of a foothold, you might still end up engaging in fluff, content that is easy to digest, but not the core hard part of what you are trying to learn. As an example, most versions of passively watching video, as opposed to actually trying to do the thing, are fluff.

Remember that learning = time * intensity. Once you have a foothold that has some hope about it, you want to dig into the hardest part of it. Look for something that would strain your effort some.


First look for hope, then move toward the hard part. (Point yourself in the right direction, then increase your magnitude.)

Some musings on deliberate practice

[Epistemic status: unverified postulates. Probably not the right ontology, or even first order factors.]

A few years ago, I taught myself to touch-type in Colemak mostly using the online software keyber.com. I came back to this recently, to see if I could increase my typing speed further. In particular, it seems like one of the main things slowing me down is punctuation, which I didn’t train at the time. So I’m focusing on that this week.

This seems to be going much better / faster than when I was originally learning to touch type, and it’s inspired me to write some notes about deliberate practice.


I used to practice Parkour with Duncan. He had a saying, about learning new parkour motions:

First do it right. Then do it smooth. Then do it quiet. Then do it fast.

Where, actually, by the time you’re doing it right smooth, and quiet, you’re automatically doing it at speed.

I think this generalizes. Or at least the “do it right, before you do it fast” part does. When you’re learning something you first and foremost want to focus on doing it right, no matter how slowly.

Practicing touch typing, at least when I’m starting out with some new keys, is extremely deliberate: I might pause for a half second before I hit each key, verifying that I am about the hit the right one.

I tend to speed up automatically as I start to get a handle on it, and have more of a sense of the rhythm. If I make more than a few (3?) errors, I go back and slow down.

Obviously, in some domains, its pretty hard to adjust the speed: juggling comes to mind. There’s probably something that can be done about that.

Note and counter-rehearse errors

Going slow also lets me pay attention to specific errors. The nature of the touch typing task also helps a lot: keystrokes are discrete, and I have clear feedback about if I hit the right key. (This is probably harder in other domains.)

But every time I make an error, I notice specifically what it was (“I hit the “y” key with my ring finger instead of the “semi colon”), and mentally pseudo rehearse the reflex that I want to execute instead (reaching up to hit “semi colon”).

I’ll typically go back to the beginning of the session and play it through again. Having just noted the error and the correction, I usually do it correctly on the second run-through.

Vary psychological sliders

In addition to speed, it feels like there are other high-level scaler variables that I can adjust up and down. I’m not sure that I’m naming these right, but some of them include

  • “intensity”
  • “commitment to accuracy”
  • “speed”
  • “sloppiness”
  • “calm” or “settledness”

This is something like the intention that I’m holding when I’m practicing. It seems like good deliberate practice is mostly a matter of identifying which slider(s) are relevant and holding those in the right place (which might be an extreme or a sweet spot in the middle of the scale, depending), while doing the activity.

Perception of Progress

If it feels unmotivating to sit down and do deliberate practice, that’s probably because you don’t viscerally perceive yourself to be making progress. Feeling yourself get better is almost always engaging (?).