Nice one. A book for non-mathematicians. The approach of the book is more on logical and rational thinking rather than number-crunching.
Although, I think this book will either increase your confidence in your decision-making abilities, or make you more depressed or worried because you keep getting the quizzes wrong!
Or you could argue the reasoning over the preferred solution is wrong.
But making a decision, when options are presented to you, is one thing. The book does just that – present you options to choose. useful exercise to develop a discipline in rationality. Though in real life, it’s also about being able to develop creative solutions rather than just make/ choose the most rationale decisions.
I was also made even more aware that decision-making is easier when emotions are left out of the equation. For instance, there’s a question on whether the groom should go with the wife’s interest, or the parents, or just choose the groom’s when it came to planning a wedding. The rationale decision was to go with the wife, since she’s the person the groom will have to live with. But again, reality would depend on what you, the groom, will ultimately accept as a long term gain. It could be also giving time and explaining to the parents. Or the wife.
As I understand it:
- recognise the need to make a decision
- determine the options and alternatives and probably outcomes (P39. there are basically two types of decisions: where you can be certain of the outcome when you select an alternative, and where you can only estimate the probability of outcomes).
- there are 4 major decision criteria in decision theory: admissibility criterion (eliminate the improbable ones or those with lowest payoff), Minimax (the least-worse of the outcomes), Bayes criterion (the outcome with the greatest long-term gain/ expected value), Maximax criterion (choosing the best from positive outcomes).
The right decisions also depend on the right information.
I thought this book is really about “being conscious about one’s decisions”. For what is a correct decision depends on perspectives. But the book deals with the need to make decisions in situations that involves a lesser degree of subjectivity (i.e. Personal values and emotions) and more of achieving logical and rational outcomes.
Of course, no guide can allow us to predict everything with certainty. What this book does is allow us to approach decision making, and hence articulate it if necessary, in a rational manner.
A third of the book, the last part, is on quizzes, to reinforce learning.
p11/ 14. Decision Theory only deals with alternatives that can be selected (not make bets on the impossible). It evaluates decisions/ choices based on a number of “rational criteria”.
I think we apply aspects of decision theory, but may not be conscious about it.
Payoffs vs compromises
Long term vs short term goals
Immediate returns vs future (can work both ways; exercise on what would you do if you were Edison: build a stock price printer that has immediate demand but payoff is less than potential returns of a yet to be invented light blub?)
Probability of success/ outcomes?
Chapters contain “thought exercises” based on actual or hypothetical examples like:
- being Leslie Groves, the US general in charge of building the atomic bomb, on whether to hire a popular but left-winged scientist or a personally brilliant one?
Choosing a PhD thesis supervisor: a very tough one and likely unpleasant experience but has proven track record of producing PhD students; an upcoming researcher whose findings may be a nobel prize winner; or a supervisor you can connect with and likely to have good experience with PhD program but track record is less certain
P74. Bayes’ Criterion: select alternative that has biggest payoff in the long term. Long term gained involves math concept of expected value. P77. Computing the long term average gain, “expected value”, which involves knowing the probability of outcomes and the associated payoffs.
P87. Maximax criterion. For situations where one should throw caution to the wind (provided you also can identify correctly this is the situation) and choosing the alternative that gets the most out of the decision.
P98. Inadmissible situations should be avoided.
P99. Bayes’ Criterion – for recurring situations. Though author qualifies situations like whether to use Atom Bomb is, while one-time, considered a unique situation but with recurring probability.
- Minimax decisions: characterised by a disaster we wish to protect against (i.e. Minimise worse case scenarios)
- Maximax decisions: for optimistic (e.g. payoffs is so huge) or pessimistic (only option is to win).
P103. Suggested approach for gray areas/ hard to tell situations:
- Layout all alternatives and try to assess the associated rewards and risks.
- Eliminate inadmissible alternatives
- Then consider if it is a Minimax situations (i.e. Is this a disaster avoidance situation? “when a disaster looms it’s not a faint cloud on the horizon but a thunderhead”)
- if not, then choose the alternative and ask what if the outcome don’t work out? If answer is acceptable, then take it.
P51. On game theory; learning from WWI and WWII; Treaty of Versailles and later actions by victorious Allies: “Imposed punishment in just measure for a deserved crime stands a good chance of achieving the goal; a vindicative solution involving excessive punishment leaves a residue that the passage of time fails to erase”.
P100. Winston Churchill reportedly said that if you’re not a liberal at age 20, you haven’t got a heart. But if you’re not a conservative at age 40 you haven’t got a brain.