This book is perhaps the most useful of all the ones I have read [other than the ones on playing
positional chess] and should be the starting point for anyone who wishes to understand the parts of a decision making process.
I once got into an interesting discussion with someone
on the chessgames forum who insisted that computers do not 'play' chess. I could not at the time explain why I thought that
his statement was incorrect. Perhaps he was thinking about how different machines and humans are when they focus on move
sequences and evaluate the winning chances of positions. James March would explain that the machines follow a rational
procedure to make a choice as described below:
p.2"A rational procedure is one that pursues a logic of consequence.
It makes a choice conditional on the answers to four basic questions:
1. The question of alternatives: What actions are possible?
2. The question of expectations: What future consequences might follow from each alternative?
How likely is each possible consequence, assuming that alternative is chosen?
3. The question of preferences: How valuable (to the decision maker) are the consequences
associated with each of the alternatives?
4. The question of the decision rule: How is a choice to be made among the alternatives
in terms of the values of their consequences?"
In a rational process, we are concerned with the future states of the world that are possible
and likely, as well as how we would feel once we reached one of the possible states:
p.3"Within rational processes, choice depends on what alternatives are considered and on two guesses
about the future: the first guess is a guess about future states of the world, conditional on the choice. The second guess
is a guess about how the decision maker will feel about that future world when it is experienced."
Decision making involves dealing with uncertainty, and we have (perhaps) probability, previous experience,
advice of others, rules of thumb, sample polling modeling and analysis to help us imagine or predict what lies within that
uncertain area. If we cannot resolve the uncertainty, we might instead choose to take a path that avoids the question entirely.
p.5"Decision makers are assumed to choose among the alternatives on the basis of their expected
consequences, but those consequences are not known with certainty. Rather, decision makers know the likelihoods of various
possible outcomes, conditional on the actions taken."
How we allocate our attention and comprehend our world has a big impact on the decisions we make.
Sometimes the complexity is so large or the time is so limited that we need to begin making simplifying assumptions and using
them to form the basis of our decision.
p.10"Time and capabilities for attention are limited. Not everything can be attended to at once.
Too many signals are received. Too many things are relevant to a decision. Because of those limitations, theories of decision
making are often better described as theories of attention or search than as theories of choice. They are concerned with the
way in which scarce attention is allocated... Decision makers have limited capacities for comprehension. They have difficulty
organizing, summarizing, and using information to form inferences about the causal connections of events and about relevant
features of the world. They often have relevant information but fail to see its relevance. They make unwarranted inferences
from information available to them to form a coherent interpretation...As decision makers struggle with these limitations,
they develop procedures that maintain the basic framework of rational choice but modify it to accommodate the difficulties.
Those procedures form the core of theories of limited rationality."
Sometimes we need to simplify our complex world by using numbers which represent complex quantities.
I watched a post-game TV show yesterday where the announcers discussed the 'quarterback rating' of a certain American football
player. The art of passing, handing off the ball, and general leadership on the field has been reduced to a number. In fact,
postgame TV shows are often groaning with statistics as the announcers attempt to explain to the audience why a team won or
lost, and what the primary causes were. Our computer chess program must find a way to reduce the complexity of the situation
on the board to a single number so we can estimate which positions are more promising than others, and use that
information to focus our search.
p.15,16-17"Faced with a world more complicated than they can hope to understand, decision makers
develop ways of monitoring and comprehending that complexity. One standard approach is to deal with summary numerical representations
of reality, for example income statements and cost-of-living indexes. The numbers are intended to represent phenomena in an
organization or its environment: accounting profits, aptitude scores, occupancy rates, costs of production. The phenomena
themselves are elusive - real but difficult to characterize and measure... Decision makers and professionals try to find the
right answer, often in the face of substantial conceptual and technical difficulties. Numbers presuppose a concept of what
should be measured and a way of translating that concept into things that can be measured."
Seeking, investigating, exploring, evoking - these concepts of attention and how our computer program
addresses them will determine to a large extent how strong our computer program will be.
p.23"Not all alternatives are known, they must be sought; not all
consequences are known, they must be investigated; not all preferences are known, they must be explored and evoked. The allocation
of attention affects the information available and thus the decision."
The study of search and attention might improve the performance of a computer
chess program, as well as the study of decision making.
p.23"The study of decision making is, in many ways, the study of
search and attention."
When we decide how to allocate the attention of our computer chess program, we are making critical
decisions that will affect the overall performance.
p.24"Decisions will be affected by the way decision makers attend (or fail to attend) to particular
preferences, alternatives, and consequences... Decisions happen the way they do, in large part, because of the way attention
is allocated, and 'timing' and 'mobilization' are important issues.
Decision makers appear to simplify the attention problem considerably."
Rational action seems to involve imagining the future and taking steps to make the imagined world
real. Perhaps predicting the future is a better term. Prediction begins by determining what is possible,
and from that what is likely. The difference is the consideration of constraints in addition to goals
and strategic factors which might help us to achieve the goals or to put pressure on the opponent to
prevent us from achieving our goals.
p.78"Much modern thinking about decision making presumes that the expectations and willful actions
of human beings enact the future in the present. The presumption is reflected in theories of rational action and power, including
theories of strategic action... In these perspectives, change stems from imagining the future and imposing it on the present.
Visions of the future, or destinies, are confirmed by following courses of action necessary for their fulfillment."
In competitive situations such as games, actors struggle for positions of power which offer the
promise of obtaining objectives. Perhaps we should be looking at the potential of each position on the chessboard to achieve
a position of power, and the abilities of pieces on the board to make a contribution in the struggle for power.
p.141"The basic idea of decision making in the face of inconsistency is that different people want
to have different things or to fulfill different identities, and not everyone can have everything desired. As a result, individuals
(and groups) struggle, competing and cooperating with each other, trying to satisfy their individual preferences and identities.
Power is the capability to get what you want or fulfill your identity... The standard presumption of decision making is the
struggle for power and, through power, for desired outcomes."
In a game such as chess, Botvinnik [Computers, Chess and Long-Range Planning, p.9] suggests
that the players exchange intangible positional factors with each move. The strategic plan is to properly evaluate the
worth of these difficult to estimate positional factors, and over time to accumulate small positional advantages that lead
to a position of power or strength. From this powerful position, an attack on a weak spot in the enemy position can be launched
that cannot be countered without undergoing a further weakening of the position.
p.148-149"Exchange models of power address that problem [power weights cannot be independently observed
but must usually be estimated from their consequences] by focusing on a small number of factors that provide a trading advantage
in a system of voluntary exchange... The fundamental idea in an exchange model is that participants... enter into voluntary
exchange relationships regulated by some system of rules. Each participant brings resources into the arena... The process
of choice is one of arranging mutually acceptable trades within the rules. Each individual seeks to improve his or her own
position by trading with other individuals... Power in an exchange model comes from control over resources desired by
others... When decision makers have something others want, they can exchange it for something that they want... the possession
of desired resources gives power. In order to be powerful, decision makers seek control over resources."
If we desire our computer chess program to behave in a rational manner, James March suggests that
it make two specific guesses, or estimates, about the 'world', which in this case is the chess game currently underway. Our
program must attempt to estimate the future consequences of the possible moves, and it must attempt to estimate the
future preferences which will then come into existence several moves into the future as we proceed down the
predicted lines of play.
p.180"Rational action stems from two guesses about the world. One is a guess about the uncertain
future consequences of possible current action. The other is a guess about the uncertain future preferences by which the outcomes
of a current action will be evaluated in the future."
The proposed heuristic attempts to interpret the current situation on the chessboard by making predictions
about future consequences. We use this information to form a rough but realistic estimate of the winning chances that
is useful and efficient in steering our search efforts. This attempt is part of a rational procedure to
make a decision on which move is the best move to play.
p.207"This book is built around two alternative visions of decision making: The first is a vision
of rationality in which actions stem from expectations of their consequences and preferences for those consequences. The second
is a vision of rule following in which actions stem from a matching of the demands of identities with a definition of the
situation. Each vision assumes that decision makers interpret their situations and their experiences, that they make sense
of them in order to make decisions. Rational actors - whether acting alone or in negotiation with other rational actors -
interpret their situations and experiences to predict future consequences of current actions and their future feelings about
such consequences. Rule-following actors - whether acting alone or in concert with other rule-following actors - interpret
their situations and experiences to identify appropriate identities and rules. They interpret history to develop the rules
they follow."
We gather information, through the process of search and attention, in order to resolve uncertainty
and to evaluate decision candidates.
p.207"Uncertainties are reduced through the accumulation and retrieval of information. Information
systems are designed, and information is used to facilitate judgments about consequences or appropriateness. Meaning is established
in order to make decisions.
From such a perspective, decisions are important because they allocate resources
and produce measurable consequences for the decision maker. Information is meaningful if it resolves uncertainties about preferences,
consequences, situations, and identities."
Decision intelligence is defined by either a process or a measurement of results. Perhaps it is
really bouncing back and forth between these two - we use results to drive process and process to create results.
p.222"Decision engineering is dedicated to producing decisions that are intelligent, but the definition
of intelligence is often left unclear. Students of decision making oscillate between process and outcome definitions of intelligence
and have never been able to resolve satisfactorily some difficult issues associated with key tradeoffs underlying the definition
of good outcomes."
Should we take the short-term or long-term view when evaluating the worth of outcomes?
p.227"Decision outcomes unfold over time. The short run is nested in the long run. Many actions
that contribute to short-run well-being are deleterious in the long run, and vice versa. Moreover, preferences and identities
change over time, partly as a result of taking actions. Are outcomes to be evaluated in terms of preferences and identities
that existed at the time of the decision or in terms of those that exist at the time at which the effects of the decision
are realized?
The complications of weighting consequences that are distributed across time
constitute a prime topic in both the psychology and the economics of choice."
We should look at the distant consequences of a decision if we truly desire that our decision exhibit
intelligence.
p.232"Making an assessment of decision intelligence involves a judgment with respect to distant
consequences of actions taken here and now."
Making a decision should involve the search and exploitation of knowledge - especially knowledge which
allows us to anticipate the consequences of our decision.
p.240"Decision processes presume the exploitation of knowledge. For example, rationality involves
anticipating the future consequences of present actions, as well as future preferences for those consequences when they occur.
The ability to use knowledge to anticipate consequences and establish preferences for them is essential."
The proposed heuristic uses adaptive decision making, which James March defines below:
p.252"Adaptive decision making is predicated upon skill at understanding the environment and responding
to it. Many adaptive strategies involve monitoring the environment, understanding its causal structure, storing inferences
drawn from that understanding, and retrieving the implications of those inferences at the appropriate times and places."
