From: veritasium
Expertise, at its core, is fundamentally about recognition [04:01:00]. Experts can recognize complex stimuli as single, meaningful entities, a process known as “chunking” [03:32:00]. This ability is developed through extensive experience and allows for fast, automatic thought (System One thinking) [00:21:00].
How Chunking Works
Chunking refers to the process where information stored in long-term memory enables an individual to recognize complex stimuli as just one thing [03:33:00]. For example, the digits “3.14159” are recognized as “pi” rather than a string of six unrelated numbers or meaningless squiggles [03:40:00]. This contrasts with System Two thought, which is conscious, slow, and effortful [00:16:00].
Examples of Chunking in Expertise
Chess Masters
Decades ago, scientists studied what makes experts, such as chess masters, special [01:45:00]. They found that chess masters are not exceptional in general measures like IQ, spatial reasoning, or short-term memory spans [01:57:00]. However, their performance is vastly superior when dealing with chess-specific situations [02:02:00].
A 1973 experiment by William Chase and Herbert Simon demonstrated this [02:08:00]:
- Three chess players (a master, an advanced amateur “A player,” and a beginner) were shown a chessboard with about 25 pieces arranged as they might be in a real game for five seconds [02:10:00].
- From the first look, the master recalled 16 pieces, the A player recalled eight, and the beginner recalled only four [02:36:00].
- The master needed only half the peeks to perfectly replicate the board compared to the A player [02:45:00].
- When the researchers arranged the pieces in random, unrealistic positions, the chess master performed no better than the beginner, recalling only three pieces after the first look [02:50:00].
This demonstrates that chess experts do not have better general memory, but specifically better memory for chess positions that could occur in a real game [03:08:00]. Their brains have learned patterns from seeing numerous chess games [03:22:00]. Instead of individual pieces, they see “recognizable configurations” or chunks [03:28:00]. This allows chess champions like Magnus Carlsen to recognize chess positions instantly, similar to how people recognize faces [04:05:00].
Memorizing Pi
Grant Gussman, who memorized 23,000 digits of pi, uses chunking to aid his recall [00:33:00]. He associates sequences of numbers with meaningful information [03:47:00]. For instance, “three zero one seven three” means “Stephen Curry number 30, won 73 games” from 2016 [03:49:00].
Recognition and Intuition
The ability to chunk and recognize patterns leads directly to intuition [04:10:00]. Just as seeing an angry face provides an instinctive idea of what might happen next [04:13:00], chess masters recognize board positions and instinctively know the best move [04:19:00]. This means they often “know what to do” without having to “figure it out” [04:24:00].
Developing Chunking and Expertise
Developing the long-term memory of an expert takes a significant amount of time, often cited as around 10,000 hours [03:32:00]. However, simply accumulating hours is not sufficient [04:40:00]. There are four critical criteria that must be met to develop true expertise:
- Many repeated attempts with feedback: Experts get clear feedback on each attempt [04:54:00].
- A valid environment: The environment must contain predictable regularities that can be learned [06:46:00].
- Timely Feedback: Feedback needs to be delivered quickly enough to allow for learning and adjustment [11:18:00].
- Deliberate practice: This involves practicing at the edge of one’s ability, pushing beyond the comfort zone, and methodically attempting things one is not yet good at [14:25:00].
When these criteria are met, the human ability to recognize patterns and perform as an expert is astonishing [16:40:00]. However, when these conditions are absent, individuals widely considered experts may not actually demonstrate expert performance [16:45:00]. For example, financial professionals fail to consistently beat the market due to the low validity of the stock market environment [09:56:00]. Humans also have cognitive biases, such as trying to find patterns in randomness, which hinders effective learning in low-validity environments [11:08:00].