Why humans lose to an eight-state automaton
In 1953, mathematician Claude Shannon built an electromechanical console using telephone stepping relays at Bell Labs. The machine played a binary guessing game against a human operator. Across hundreds of sessions, the machine won between 60% and 70% of all rounds.
The machine did not read minds. It exploited three universal biases in human cognition:
When asked to generate random sequences, humans actively avoid repeating the same symbol more than twice in a row. In a genuine 20-bit coin toss sequence, the mathematical probability of encountering a run of four or more identical symbols is 82.5%. Humans produce runs of four in fewer than 20% of trials because they intuitively assume short sequences must remain evenly balanced.
Extensive psychophysical trials (Wagenaar 1972) demonstrate that humans alternate choices at an average rate of 65%, well above the mathematical 50% expectation. After choosing 0, people feel an unconscious pressure to choose 1 next. The automaton recognizes this bias and turns it into points.
When the machine correctly anticipates a choice, players immediately switch strategies. The automaton's 8-state transition memory tracks how you respond after wins versus losses, predicting the exact countermove.
How the machine operates
The automaton maintains two binary variables from previous rounds:
- Whether the machine won or lost the preceding round;
- Whether the human repeated or alternated their choice compared to the step before.
These two variables produce four operational states. For each state, the machine stores the player's subsequent decision. When that state reoccurs, it wagers that you will repeat that past behavior. The guess is locked in before you touch the button.
Methodological bounds
This test measures rapid binary motor decisions across 20 rounds. The score reflects inductive motor patterns in working memory under interface constraints; it carries no clinical or diagnostic weight.