liveBook · Manning

Maximax method

As explained in Statistics Every Programmer Needs

The Maximax method is a decision rule that selects the alternative with the highest possible payoff across all states of nature.

The maximax method is an optimistic decision rule for uncertain outcomes. Without probabilities, it selects the alternative whose best possible payoff is highest, assuming the most favorable state of nature occurs. section 10.0 section 10.1.1

Apply the method systematically:

  • List the alternatives, states of nature, and payoff for every alternative–state combination. section 10.1.1
  • Compute each alternative’s maximum payoff across all states: maximax(i) = max(payoffᵢⱼ). section 10.1.1
  • Choose the alternative with the largest row maximum. section 10.1.1

Investment alternatives

For strong and weak economic conditions, the maximum payoffs are $100,000 for Stock Market, $60,000 for Professional Soccer Club, $80,000 for a 50/50 combination, and $35,000 for a Savings Account. Maximax therefore selects Stock Market, despite its -$120,000 weak-economy payoff. section 10.1.1

>>> maximax['Maximax'] = \
>>>    payoff_table[['Strong Economy', 'Weak Economy']].max(axis = 1)
>>> print(maximax)
             Investment Alternatives  Strong Economy  Weak Economy  Maximax
0                       Stock Market          100000       -120000   100000
1           Professional Soccer Club           60000        -20000    60000
2 Stock Market / Soccer Club (50/50)           80000        -70000    80000
3                    Savings Account           35000         35000    35000

A pandas payoff table adds a Maximax column by taking the row-wise maximum across economic states. section 10.1.1

How Statistics Every Programmer Needs covers it

How deep it goes
Defines the rule, gives procedural steps, and demonstrates pandas implementation, but does not estimate probabilities or evaluate sequential decisions.
Examples
Uses an investment payoff table covering strong and weak economies, including stock, soccer-club, mixed, and savings alternatives.
What it assumes
Assumes a defined set of alternatives, states, and payoffs while teaching the selection procedure without requiring probabilities.
What you can do after
Apply via pandas by calculating row-wise maximum payoffs and selecting the alternative with the highest result.
How it connects
Relates to payoff tables, states of nature, Maximin method, Minimax Regret method, Expected Value method, and decision trees.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Statistics Every Programmer Needs. The section links point to the parts of the book it draws on.