Capital CasebookEducation of a Speculator

All analyses

Count behavior in its context

The same player can behave differently under different incentives.

Open the interactive lesson & exercises ↗

Illustration for: Count behavior in its context
Conceptual illustration · not a historical photograph or market data

The father’s sporting records distinguish scores, shots and errors. The discussion also contrasts formal and informal matches and describes how a player’s behavior changes with the score.

Avoid an average that hides the question

If performance changes with context, one overall average may conceal the relevant comparison. Define the groups before viewing the result and keep group sizes visible. Context helps refine a question; it must not become an unlimited excuse for excluding inconvenient cases.

Worked example

A fictional player attacks more when behind than when ahead. Combining both states into one shot-success rate hides the change in choices.

Case connection

A match score alone cannot reveal decisions in different positions. This case is a prompt to record context alongside outcomes.

Deep Blue meets Kasparov

Source-grounded facts

IBM’s Deep Blue defeated world chess champion Garry Kasparov in a six-game match.

Context

IBM set out to build a chess computer capable of beating the reigning world champion. The contest gave both sides a clearly defined board, legal moves and tournament time controls.

Outcome

Deep Blue won the match 3.5–2.5. This demonstrated a specific competitive capability under chess rules, rather than proving superiority at every activity requiring judgment.

  1. In 1996, Deep Blue won the first game, but Kasparov recovered to win the six-game match 4–2.
  2. The team improved the computer’s evaluation function and endgame databases before the rematch.
  3. In May 1997, each side won one early game and the next three were drawn. Deep Blue then won the sixth.

IBM history

Case analysis

The final match score compresses many different positions into one number. To study behaviour, keep the state of play beside each move: what was known, what alternatives existed and whether the player was ahead or behind. Those labels should be defined before comparing results. Context can refine a question without becoming an excuse to discard inconvenient observations.

Try it

For an everyday observation, record one context variable such as time, setting or task difficulty along with the outcome.