Analytical Methods (Decision Analysis)

Front 1

Value of decision analysis (3)

Back 1

(1) understand different options;
(2) make decisions in objective, rational ways;
(3) forces express consideration of options

Front 2

Decision Trees (function)

Back 2

Organizes all decisions and their consequences, with their corresponding probabilities and importance
- Used by Gov’t to help design policies by predicting people’s behavior (e.g., EPA w/ hazardous waste to determine whether people will want to test for wastes given the costs)

Front 3

Decision Trees (Components) (4)

Back 3

1) Decision Node
2) Decision Branches

3) Chance Node:
4)Chance Branches

Front 4

Decision Node

Back 4

An action (i.e., a decision) that must be taken

Front 5

Decision Branches

Back 5

Represent all of the possible decisions stemming from a decision node; indicates payoffs/consequences

Front 6

Chance Node

Back 6

Different possibilities (i.e., outcomes); NOT a decision

Front 7

Chance Branches

Back 7

every outcome possible from a chance node (i.e., from a given possibility); indicates likelihoods as well as payoffs/consequences
- Demonstrate Uncertainty

Front 8

Expected Values (Definition)

Back 8

- Derived at chance nodes
- The amount you would obtain on average, from repeated exercises under the same terms I.e., the probability of a payoff multiplied by its value
- Best way to evaluate a series of different possibilities

Front 9

Expected Values (Equation)

Back 9

(probability1)(value1) + (probability2)(value2) +…+ (probabilityX)(valueX)

Front 10

Ways of Gathering data for decision trees (3)

Back 10

1) Decisions and Outcomes:
2) Probabilities:
3) Payoffs

Front 11

Ways of Gathering data for decision trees (Decisions and Outcomes)

Back 11

- Discussions w/ clients, other lawyers, experts, etc.
- Consider proceeding chronologically
- Construction of the Tree itself can reveal new options

Front 12

Ways of Gathering data for decision trees (Probabilities)

Back 12

- Hard data (review of case outcomes, audit rate for tax issues, etc.)
- Research: E,g, mock trials, experts, etc.
- Try the Roulette Wheel Method (pp. 28-29)
- A lot of times, attorney and/or client must supply the %s

Front 13

Ways of Gathering data for decision trees (Payoffs)

Back 13

- Ask client, hire experts, attorney self-knowledge, etc.
- Don’t forget to include non-monetary costs: e.g., time, emotional cost of trial or embarrassment, etc.

Front 14

Sensitivity Analysis

Back 14

Def: How wrong can you be and still reach the right decision?
- Answers: Whether the best decision is sensitive to particular changes in the data
- One Method: Ask how much the estimate of probability or payoff can vary w/o changing the result reached; sometimes small changes in estimated values can lead to different preferred decisions, and sometimes even big changes won’t

Front 15

Crossover points

Back 15

- points where your decision changes
- Focal points when reflecting when considering reliability

Front 16

How a Decision Tree Question is Asked (3)

Back 16

1) WHAT SHOULD YOU/CLIENT DO?
2) SHOULD YOU ADVISE CLIENT TO DO X?
3) WHAT METHOD SHOULD BE CHOSEN?

Front 17

Game Theory (types) (2)

Back 17

1) Sequential Decisions, and
2) Simultaneous Decisions

Front 18

Game Theory (Functions) (4)

Back 18

1) Designing Ks,
2) formulating litigation strategies,
3) conduction negotiations,
4) explaining corporate takeovers and anticompetitive behavior,
5) etc.

Front 19

Game Theory (goal)

Back 19

Each party want's to maximize its potential payoff

Front 20

Payoff

Back 20

All things that matter to a player at any time during the game.

Front 21

Decision Payoff Matrix (use)

Back 21

Used for simultaneous decisions ONLY

Front 22

Decision Payoff Matrix (Procedure)

Back 22

Utilize Game Tables
1)Write the LEFT players payoffs first, followed by the TOP players
2) Circle each player's best option when working out a table

Front 23

Collusion

Back 23

- Can solve several problems, but if it can not occur, or there is a threat of breaking collusion, problems such as the prisoner’s dilemma can occur
- Collusion can occur through direct communication b/w parties, reputation of a party for a certain action, etc.

Front 24

Dominant Strategies

Back 24

- When one decision will produce the best outcome for a party regardless of what the other party does; therefore, that decision will always be made, irrespective of the action of the other party
1. If each party has a dominant strategy, the game is solved
2. BUT, lack of a dominant strategy is not damning; we can still say something about what will happen by predicting the status quo will continue

Front 25

Nash Equilibrium

Back 25

- Each player chooses the best action for itself given what it thinks the other parties will do

1. There is NO deviation from a Nash equilibrium

2. NEs often do NOT exist
a.Frequently do NOT exist when one party does not have a NE

3.IF several NEs exist, the one occurring more often OR the one involving subgame perfect actions will be the focal point

Front 26

Subgame Rational/Perfect Decisions

Back 26

- rational decisions made when the choice about that action arises
1. Players should only believe that other players will make subgame rational/perfect decisions
2. E.g., empty threats are NOT subgame rational and should be ignored

Front 27

Decision Trees (use)

Back 27

For sequential decisions ONLY

Front 28

Decision Tree (procedure)

Back 28

Utilizes game trees
i. Decision Nodes are marked, on the inside of the box, with the name of the player making the decision

ii. Payoffs are listed at the ends of decision brances in (X,Y) format
1. The action is on the top of the branch, with the cost on the bottom
2. “+” or “-“ signs: used for extra benefits/detriments that we know exist, but are unsure how to quantify

iii. Chance nodes are also used; the top of a chance branch contains the resulting action, the bottom contains the probability of occurrence

iv. Write the result in (X,Y) format under each chance and decision node
b. Nash Equilibriums can be determined through decision trees

Front 29

Moral Hazard (Definition)

Back 29

Once a contract is made, a party to it may have incentives to act in a way detrimental to the other party

Front 30

Solving Moral Hazard

Back 30

- Obtaining information about the problematic behavior/situation of the other party to induce different behavior
- BUT, acquiring info is costly, can be difficult, the info can be imperfect, etc.

Front 31

Output-based incentive

Back 31

- Moral hazard solution
- impose risk on people which make them unpopular; imposing risk on risk averse individuals can have negative consequences (e.g., higher salary demands for CEOs)
- Note: Output also hard to measure