In this article I’ll set out my personal methodology, paradigm, and mental models for learning new skills efficiently. That isn’t to say that this is the only way to learn new skills (obviously) but it’s a framework that prioritises outcomes and efficient time usage, without any pop-psychology learning “hacks”.

Core Insight

The core mechanic we’ll be focusing on here is the idea of “neural adaptation”. Your brain currently receives inputs X and produces outputs Y as a result:

• In chess, the inputs are the board the outputs are the move
• In programming, the inputs are the current code and the outputs are the changes made

To improve at any cognitive skill, your brain needs to produce a different Y for the same X. Which means that your brain needs to transition from it’s current state, to a new state with a different structure that produces different outputs. This process of change is what I’m going to refer to as “neural adaptation”, and I’m going to hand-wavily model it as a fixed amount of individual changes that need to occur on the individual neuron level. That number might be millions, billions, or trillions, it hardly matters - what matters is that to move your brain from state N (now) to state N + 1, there are an amount of connections that need to be rewired, and the most efficient skill acquisition methodology is the one that achieves these rewirings the fastest.

Our methodology will focus on using exercises that maximise neural adaptation density strategised towards the elements of the chosen skill that give the best yield in outcomes.

The Three Phase Cycle

Our model will focus on a cycle of three phases:

1. Testing (or practice, discovery, play, research, data gathering)
2. Review (or strategy, leverage, study, research)
3. Neural adaptation

Testing

The testing phase is the period of time you expose yourself to the actual skill you’re trying to find out where you are with it. If you’re learning chess, this will be playing real games vs human players. If you’re learning martial arts, this will be your real sparring time.

Many see this is the entire process in and of itself, citing common idioms like “practice makes perfect” and “gotta put in the hours”. It’s true that learning can occur here, significant learning even, but I would like to challenge the role of this sort of practice in the learning cycle.

This sort of “end to end” practice is usually neither particularly well strategised to focus on the highest leverage elements of the skill, nor is it particularly dense with neural adaptation. In an average 20 minute game of rapid chess, you may only experience a couple of board positions that require the deep thought capable of stimulating neural adaptation, and often those positions arrive when the clock is ticking down and you don’t even have the time to mine all the neural adaptation you could.

In an average 3-5 minute MMA sparring bout, you may only spend one minute, if not only a few brief moments, in the situation that you most want to work for the best outcomes. For example, a particular kick could be, once practiced, a dominating, fight-ending technique. But practicing it during normal sparring bouts may only see you get 1-2 reps in, about 5 seconds worth of neural adaptation relevant to the highest leverage technique, and the other 4 minutes and 55 seconds are spent randomly responding to your opponents moves and not practicing the thing you would receive the most value from.

I propose that the best use of “live practice” isn’t to drive the development of motor skills and decision making itself, it’s to stay present and atuned to the direction your practice needs to move in order to see best results.

Review

The next phase, review, is where we identify the highest-leverage vector of improvement. It’s common advice to review your games afterwards when learning chess, but I think many learners lack direction during this review process because they’re trying to get the wrong thing from it - they’re trying to mine neural adaptation from a couple of individual games, not much better than trying to mine neural adaptation from them while you’re playing them - which is exactly the method I’ve just challenged above. Instead, I suggest that learners use the post practice review period to focus on strategy and learning direction. How are you most often failing? What sub-skills would make the most difference to your outcomes?

Neural Adaptation

The final, and most crucial element of the cycle: neural adaptation. This is the phase where we actually drive dense growth and adaptation to move our brain from its current state to a state that makes better momentary decisions and yields better outcomes.

To discuss this stage, we need to understand two significant neurological mechanisms for driving neural adaptation:

1. Dopamine-based operant conditioning
2. “Greasing the groove”

Operant Conditioning

Operant conditioning is the process through which the dopamine system rewards behaviours and encourages them in the future. Most obvious in dog training, a reward reinforces the behaviour and decision making that preceded it.

Two crucial caveats:

The first is that operant conditioning almost exclusively applies to intentions rather than specific motor skills. Just because something “feels good” doesn’t automatically make you mechanically better at it, it just makes you more likely to decide to do it in the future. You are not going to use reward based conditioning to improve your chess puzzle skills by itself, only to encourage you to engage with the process more.

The second caveat is that your brain is really quite intelligent, and it’s very hard to fool the dopamine system with faux rewards. If you reward yourself with a little piece of chocolate after doing some hard work, your brain will not backpropagate that reward to the actual hard work, it will instead reward the immediately preceding decision to eat some chocolate, and over time you will find yourself eating more chocolate while working just as hard as before - or worse, you will end up craving chocolate every time you engage in hard work. This is the crux of the reason you cannot simply use addictive substances to make yourself into an unstoppable work machine, you will just end up a drug addict.

The reason food rewards work with dogs is because they do not have the option to reward themselves. The decisions they make that actually result in reward are the skills that convince you to reward them. They cannot simply learn to choose to eat the reward without performing the skill like a human with agency can.

This understanding immediately debunks a significant amount of tiktok pop-psychology learning advice. If you’re going to use operant conditioning to learn to do something, such as learning to work harder, you will either need the reward to come from the act itself and not a secondary decision (e.g. the intrinsic feel-good factor that inherently follows a period of good hard work), or needs to come from an external source that will not provide the reward unless the action is performed (e.g. money received from a paying client, or perhaps even just a reward provided for you by a family member or partner by agreement - if they give you the chocolate then, like the dog, you will learn to do the work instead of learn to treat yourself).

This way, we can use operant conditioning as the main driver for neural adaptation regarding any skills or sub-skills that are primarily emotional, such as learning to work / focus harder, learning to tackle unpleasant habits.

”Greasing the groove”

This is the other primary driver of neural adaptation that we’ll look at, and it’s the simpler mechanism. If you’re ever heard the adage “neurons that fire together wire together”, that’s this.

In short, any motor skill / momentary decision / cognitive process you ever engage with becomes mechanically easier every time you engage with it. The neurons involved wire themselves more efficiently and become faster and more precise. You do not need a reward for this, the simple act of those neurons firing is the work factor.

For example:

• Doing mental maths makes you faster at mental maths
• Juggling makes you better at juggling
• Whistling makes you better at whistling

Basically anything you do that involves technical skill rather than emotional decision making is improved by the simple act of doing it.

Something to draw your attention to: this should be quite uncomfortable. It should feel hard. Neurons don’t like reconnecting, because it’s expensive and difficult - especially as an adult. Forcing them to reconnect requires a hard effort to engage in exercises that are just on the edge of your ability, which is one of the main components adults learning new skills lack.

One further note: there is often a practical limit to how much neural adaptation can be achieved in a single session before rest is required, spending 10 hours doing mental maths is not 10x more effective than spending one hour. In fact, 10 separate 1 hour sessions would be significantly more effective for the same amount of time.

Tying it Together

The flow we’re going to use, looked at in worked examples below, is testing -> identifying relevant exercises -> exercises -> repeat.

The outcomes from our testing / play / practice sessions go directly into our review & stregy sessions. Those review & strategy sessions parse those outcomes and understand: what sub-skills are most relevant here? What exercises drive the most dense neural adaptation in those sub-skills? Then the neural adaptation exercises can remain drivers of dense development that feel exhausting, but effective.

Worked Example: Chess

A new player learning chess can easily fall into the trap of endlessly, listlessly playing games to achieve utterly marginal elo gain over time, if any. Often burning out through lack of progress never achieving what they’d wanted.

I propose that a highly efficient learning flow for a new chess player according to our model could try something like this:

Firstly, play a couple of matches in a day. It’s not really possible to stay realistic about your strengths and weaknesses without it, and it’s also quite good fun - enjoy this process, it’s the main point of entertainment in the process.

Next, they might review those games using chess.com’s review tools - stepping through each move they made until the computer informs them of a mistake. They might make a note or a tally of the nature of the mistake they made, and whether or not it was relevant to the outcome of the game. For the second half of the review & strategy session, they might look over the tally of all their outcome-deciding mistakes and notice that one particular area is weaker than the others, endgames for example. They could then determine that the highest-density neural adaptation mechanism available here is endgame puzzles, an unlimited supply of which are available on chess.com with a rapid feedback loop and plenty of opportunity for the deep thought and groove-greasing that drives neural adaptation. They may spend 20-30 minutes practicing endgames, but they will have gotten more endgame practice than a hundred normal games would give them, which could take weeks to achieve normally.

Finally, they play through the end games until their brain practically hurts and they see endgame patters when they shut their eyes. They decide they need a rest, so they reset for the night and continue tomorrow.

Tomorrow comes, and they observe that the games they’ve played were not primarily lost through bad endgames, and that a different sub-skill has taken the lead.

Day-to-day the practice might be noisy, especially if you’re only playing a few games per day you’re not going to get a complete distribution of importance over all possible sub-skills, but a simple mechanism like this is almost always enough to keep on top of what’s actually costing you games, and ensure that every day you are getting a week’s worth of the neural adaptation you would’ve had if you just played more games, massively accelerating your timelines.