Notes on 'Peak'

https://smile.amazon.com/gp/product/B011H56MKS

Based on the idea of building mental representations, I decided here to switch from taking notes sequentially as I read to creating my own summary of the arguments after reading, and then skimming the book again to correct details and fill in anything I missed. I don’t have any way to evaluate the effectiveness of different note-taking strategies, but this certainly produced a more readable reference. It does mean that I miss out on the detailed criticism I usually record, so maybe I should add an extra pass to check references and look for counter evidence.

(Author is psych professor. Studies expertise, and introduced the term deliberate practice.)

Expert performance is not as much a product of innate talent as is widely believed. First, the existence of prodigies usually turns out to be a myth - on further investigation every world-class expert has thousands of hours of practice/training behind them. Second, there are cases where effective training techniques have been able to raise arbitrary individuals to world-class levels eg:

• Perfect pitch. Very rare. Initially believed to be genetic, but also requiring training in early childhood. Sakakibara trained an entire cohort of children to have perfect pitch.
• Chess. Laszlo and Klara Polgar decided that their children would be chess prodigies. Of 3 daughters, 1st was first woman to become grandmaster through competing in mixed competitions, 2nd reached threshold for grandmaster but was not declared, 3rd was broke record for youngest ever grandmaster, male or female.

Physical adaptability has only been explored in last 50 or so years - consider the huge difference in Olympic performance in that time, a result of better understanding of how to train athletes.

Mental adaptability even less well explored. Vast improvements in chess and music performance in the same time period though.

Adaptability is highest as a young child, but possibly much higher in adults than previously believed eg:

• Presbyopia. Farsightedness in middle age as a result of physical changes in the eye. After three months of training subjects could read letters 60% smaller. Eyes didn’t improve, but brain got better at interpreting image.
• The knowledge. London cabbies memorize entire city down to the level of individual buildings.
• Reciting digits from memory. Author trained several unremarkable volunteers to world-record level performance.
• Perfect pitch. Experiments with adults have shown some limited ability to train perfect pitch.
• Golf. Dan McLaughlin, in correspondence from the author, set out to become a professional golfer despite no prior experience. So far has reached a handicap of 3-4.

General adaptability is lower in adults, but still high enough to reach expert level performance in many cases. What absolute limitations does an adult have? Not much known yet, except that:

• Height and build are partially determined by genetics. Being eg a short basketball player is a major disadvantage.
• Skeleton is fixed by adulthood. Eg ballet turnout has to be trained as young child.
• IQ is correlated with early success in chess and go, but not with expert performance (ie IQ gives contributes but is dominated in the long run by training)
• Average IQ for scientists is higher than general population, but beyond that does not correlate with career success eg Nobel prize winners have average IQs for their field
• From another source - beginner vs expert-but-not-world-class climbers have similar anthropometrics, but there is reasonable speculation that the pulling strength of world-class climbers depends on lucky tendon placement.

Brain physically changes. London cabbies, mathematicians, glider pilots, musicians, divers display measurable increases in sizes of certain parts of the brain.

Not free eg London cabbies get worse at abstract spatial memory, expert string musicians have worse sensation in the palm of their left hand.

Requires upkeep eg retired London cabbies show partial shrinkage.

Mental representations. Chunking. Build mental structures specific to domain of expertise and use ‘pointers’ to these to reduce load on working memory. Eg:

• Most people can memorize a sentence, but would struggle to memorize a much smaller number of random letters.
• Chess grandmasters are dramatically better at memorizing chess boards from real games, but not chess boards with random layouts.
• Digit memory subjects divided sequence into trees, and mapped 3-4 long sequences into pre-existing mental structures eg 314 -> pi.

Hallmark of expertise is the ability to see patterns where beginners see chaos. Eg

• Expert football players better at predicting the next action of a player in a paused video
• Rock climbers hands are primed to form correct grip as soon as they look at a hold

Effective practice is the process of building and honing mental representations. Good mental representations also make practice more effective - by making one better aware of what to strive for and better able to notice and debug mistakes.

Deliberate practice:

• Near-maximal effort, outside of comfort zone
• Well-defined specific goals
• At any time, aiming at a specific target, not just general improvement
• Deliberate - requires full conscious attention and self-monitoring
• Accurate, timely feedback and modification of efforts in response to feedback
• Depends on effective mental representations, and involves improving those representations
• Building new skills step-by-step on top of existing skills, in a structured fashion. Equivalently, breaking complex skills down into simpler components that can be individually improved.

Better to train at 100% effort and stop when exhausted, rather than less effort over more time.

This is hard! Motivation and engagement. Can come from:

• Meaningful positive feedback - knowing that you are improving.
• Desire to challenge self.
• Satisfaction in the process itself.
• Support from environment - family, coaches, team-mates.
• Identity - “I am an X, that’s what I do.”

Ensure you always have positive feedback by dividing training into manageable milestones, and measuring yourself against those regularly.

Environment is partially under your control. Sleep well, maintain health, remove potential distractions, surround yourself with motivated peers.

Is grit/willpower innate? Open question, but:

• Seems to be domain specific eg kids who spend thousands of hours practicing music but procrastinate on homework
• Circular reasoning - I don’t practice because I don’t have enough willpower, I know I don’t have much willpower because I don’t practice

Mental representations don’t have to be built form scratch. Well-developed fields build up an understanding of how to transmit representations, and what kind of practice methods are good at honing them. For such a body of practice to build up, there must be:

• Reliable measures of performance. Either totally objective measures, such as jump distance, or at least strong correlation between expert judgment, such as in music.
• Incentive to practice and improve. Up until memorizing digit spans became popular, there was no reason for anyone to figure out how to improve.
• Time. Knowledge builds up over decades or centuries.
• Teachers/coaches. Individuals whose time is spent transmitting the previous generations accumulated knowledge.

So ideally, deliberate practice also involves:

• Developing skills in a field where effective training techniques have been established
• Training by an experienced teacher/coach

Failing that:

• Find experts in the field. Be wary of subjective judgments of expertise. In many fields experience, seniority, prestige etc have been found to correlate very loosely with objective measures of ability. Come up with your own measures in advance and strive to make them as objective and reproducible as it possible.
• Figure out what makes the experts effective. This is difficult, because mental representations may not even be consciously accessible to the expert. But figuring out what makes their training different may be sufficient.
• Design training techniques to replicate their skills. Since the previous steps are unreliable, it’s important to keep applying the same objective measures to your own performance to separate out what works and what is just coincidence.

Examples of deliberate practice:

• Top Gun fighter pilot academy put pilots in simulated combat at the edge of their abilities, and followed each session with a feedback session, based on videos of the session and advice from the experienced trainers.
• Expert chess players practice mostly by studying puzzles - given a board position and asked to figure out the next few moves - with feedback from even better players.
• Expert musicians focus on specific challenges/weaknesses (eg play this 4 bar section perfectly, or as fast as possible) rather than just playing through their repertoire repeatedly.
• Library of radiology cases with known outcomes, to provide timely feedback for radiologists practicing cancer diagnosis. Similarly for pediatric ankle x-rays.
• Franklin practiced writing by taking notes on essays he enjoyed, writing his own version a week later and then comparing to the original. Found his vocabulary was lacking compared to good writers, so also tried writing his own version as poetry to force him out of his vocab comfort zone.

In response to plateaus:

• Practice different, not harder eg athletes pro-actively vary training schedule to constantly create novel challenges and stresses.
• Isolate problems. Figure out which micro-skill is the bottleneck in your current performance by pushing yourself and seeing what breaks down first eg type 20% faster than usual speed and record which combinations causes mistakes.

(Incidental rebuttal of the popular 10000-hour rule, which was popularized by Gladwell based on a misinterpretation of the authors work. ‘World-class’ is a measure that is relative to others in the same field, so the amount of time it takes to reach depends as much on the maturity of the field as anything. The first subject in the authors experiments digit span experiments reached what was at the time world-class performance after ~200 hours.)

Skills, not knowledge. Knowing something alone is not enough to build the mental representations to employ that knowledge effectively. Focus practice/training/education on what skills/abilities should be developed, not on what knowledge should be learned.

You pick up the necessary knowledge in order to develop the skills; knowledge should never be an end in itself

Given the drastic improvements in sports, chess and music in the last century, it’s natural to wonder what improvements could be achieved in other fields by applying the same principles.