different types of judgement(ucti slide)
TRANSCRIPT
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
1/27
FORECASTINGJUDGEMENTAL
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
2/27
Module Code and Module Title Title of Slides
Methods Based on Judgment
Unaided judgment It is common practice to ask experts what will happen. This is a
good procedure to use when
experts are unbiased
large changes are unlikely
relationships are well understood by experts (e.g., demandgoes up when prices go down)
experts possess privileged information
experts receive accurate and well-summarized feedback abouttheir forecasts.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
3/27
Module Code and Module Title Title of Slides
Prediction markets
Prediction markets, also known as bettingmarkets, information markets, and futuresmarkets have a long history.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
4/27
Module Code and Module Title Title of Slides
Between the end of the US Civil War and WorldWar II, well-organized markets for betting onpresidential elections correctly picked the winner
in every case but 1916; also, they were highlysuccessful in identifying those elections thatwould be very close. More recently, in the fourelections prior to 2004, the Iowa Electronic
Markets (IEM) has performed better than polls inpredicting the margin of victory for thepresidential election winner.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
5/27
Module Code and Module Title Title of Slides
Contd
In the week leading up to the election,these markets predicted vote shares forthe Democratic and Republicancandidates with an average absolute errorof around 1.5 percentage points. The finalGallup poll, by comparison, yielded
forecasts that erred by 2.1 percentagepoints.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
6/27
Module Code and Module Title Title of Slides
Delphi
The Delphi technique wasdeveloped at RAND Corporation
in the 1950s to help capture theknowledge of diverse expertswhile avoiding the disadvantages
of traditional group meetings. Thelatter include bullying and time-wasting.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
7/27Module Code and Module Title Title of Slides
To forecast with Delphi theadministrator should recruit betweenfive and twenty suitable experts and
poll them for their forecasts andreasons. The administrator thenprovides the experts with anonymoussummary statistics on the forecasts,and experts reasons for theirforecasts.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
8/27Module Code and Module Title Title of Slides
The process is repeated until there islittle change in forecasts betweenrounds two or three rounds are
usually sufficient. The Delphi forecastis the median or mode of the expertsfinal forecasts. Software to guide youthrough the procedure is available atforecasting principles.com.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
9/27Module Code and Module Title Title of Slides
Structured analogies
The outcomes of similar situations from the past(analogies) may help a marketer to forecast the outcomeof a new (target) situation. For example, the introductionof new products in US markets can provide analogies for
the outcomes of the subsequent release of similarproducts in other countries.
People often use analogies to make forecasts, but theydo not do so in a structured manner. For example, theymight search for an analogy that suits their prior beliefsor they might stop searching when they identify oneanalogy. The structured-analogies method uses a formalprocess to overcome biased and inefficient use ofinformation from analogous situations.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
10/27Module Code and Module Title Title of Slides
The outcomes of similar situations fromthe past (analogies) may help a marketerto forecast the outcome of a new (target)situation. For example, the introduction ofnew products in US markets can provideanalogies for the outcomes of the
subsequent release of similar products inother countries.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
11/27Module Code and Module Title Title of Slides
People often use analogies to makeforecasts, but they do not do so in astructured manner. For example, they
might search for an analogy that suits theirprior beliefs or they might stop searchingwhen they identify one analogy. Thestructured-analogies method uses a formalprocess to overcome biased and inefficientuse of information from analogoussituations.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
12/27Module Code and Module Title Title of Slides
To use the structured analogies method,an administrator prepares a description ofthe target situation and selects expertswho have knowledge of analogoussituations; preferably direct experience.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
13/27Module Code and Module Title Title of Slides
The experts identify and describeanalogous situations, rate their similarity tothe target situation, and match theoutcomes of their analogies with potentialoutcomes in the target situation. Theadministrator then derives forecasts from
the information the experts provided ontheir most similar analogies.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
14/27Module Code and Module Title Title of Slides
Game theory
Game theory has been touted in textbooksand research papers as a way to obtainbetter forecasts in situations involvingnegotiations or other conflicts. A Googlesearch for game theory and forecasting
or prediction identified 147,300 sites.
Despite a vast research effort, there is noresearch that directly tests the forecastingability of game theory.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
15/27Module Code and Module Title Title of Slides
However, Green (2002, 2005) tested theability of game theorists, who were urgedto use game theory in predicting theoutcome of eight real (but disguised)situations. In that study, game theoristswere no more accurate than university
students.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
16/27Module Code and Module Title Title of Slides
Judgmental Decomposition
The basic idea behind judgementaldecomposition is to divide the forecastingproblem into parts that are easier toforecast than the whole. One thenforecasts the parts individually, usingmethods appropriate to each part. Finally,
the parts are combined to obtain aforecast.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
17/27Module Code and Module Title Title of Slides
One approach is to break the problemdown into multiplicative components. Forexample, to forecast sales for a brand, onecan forecast industry sales volume, marketshare, and selling price per unit. Thenreassemble the problem by multiplying the
components together.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
18/27Module Code and Module Title Title of Slides
Empirical results indicate that, in general,forecasts from decomposition are moreaccurate than those from a globalapproach (MacGregor 2001). In particular,decomposition is more accurate wherethere is much uncertainty about the
aggregate forecast and where largenumbers (over one million) are involved.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
19/27
Module Code and Module Title Title of Slides
Judgmental bootstrapping
Judgmental bootstrapping convertssubjective judgments into structuredprocedures. Experts are asked what
information they use to make predictionsabout a class of situations. They are thenasked to make predictions for diversecases, which can be real or hypothetical.For example, they might forecast nextyears sales for alternative designs for anew product.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
20/27
Module Code and Module Title Title of Slides
The resulting data are then converted to amodel by estimating a regression equationrelating the judgmental forecasts to theinformation used by the forecasters. Thegeneral proposition seems preposterous. Itis that the model of the man will be more
accurate than the man. The reason is thatthe model applies the mans rules more
consistently.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
21/27
Module Code and Module Title Title of Slides
Judgemental bootstrapping models aremost useful for repetitive complexforecasting problems where data on thedependent variable are not available (e.g.demand for a new telecommunicationsdevice) or data does not vary sufficiently
for the estimation of an econometricmodel.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
22/27
Module Code and Module Title Title of Slides
Judgmental bootstrapping also allows experts tosee how they are weighting various factors. Thisknowledge can help to improve judgmental
forecasting. For example, with respect topersonnel selection, bootstrapping might revealthat some factors, such as height, weight orlooks, are used, even though they are notrelevant for the job. Bootstrapping also allows forestimating effects of changing key variableswhen historical data are not sufficient to allow forestimates.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
23/27
Module Code and Module Title Title of Slides
Expert systems
As the name implies, expert systems arestructured representations of the rulesexperts use to make predictions or
diagnoses. For example, iflocalhousehold incomes are in the bottomquartile, thendo not supply premiumbrands. The forecast is implicit in theforegoing conditional action statement:i.e., premium brands are unlikely to makean acceptable return in the locale..
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
24/27
Module Code and Module Title Title of Slides
Rules are often created from protocols,whereby forecasters talk about what theyare doing while making forecasts. Where
empirical estimates of relationships fromstructured analysis such as econometricstudies are available, expert systemsshould use that information. Expertopinion, conjoint analysis, andbootstrapping can also aid in thedevelopment of expert systems
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
25/27
Module Code and Module Title Title of Slides
Expert systems forecasting involves identifyingforecasting rules used by experts and ruleslearned from empirical research. One shouldaim for simplicity and completeness in the
resulting system, and the system should explainforecasts to users.
Developing an expert system is expensive andso the method will only be of interest in
situations where many forecasts of a similar kindare required. Expert systems are feasible whereproblems are sufficiently well-structured for rulesto be identified.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
26/27
Module Code and Module Title Title of Slides
Simulated interaction
others. It is especially useful when thesituation involves conflict. For example,one might wish to forecast how best tosecure an exclusive distributionarrangement with a major supplier.
-
8/4/2019 Different Types of Judgement(UCTI SLIDE)
27/27
To use simulated interaction, an administratorprepares a description of the target situation,describes the main protagonists roles, and
provides a list of possible decisions. Roleplayers adopt a role and read about thesituation. They then improvise realisticinteractions with the other role players until theyreach a decision; for example to sign a trial one-year exclusive distribution agreement. The roleplayers decisions are used to make theforecast.