PresentationPDF Available

Predictability: An Empirical Approach (Seminar 4 of 4)

Authors:

Abstract

Seminar 4 of 4, on Day 6 of the 2008 of the Alpbach Forum
“Predictability:
An Empirical Approach”
Kesten C. Green
Business and Economic Forecasting Unit
Monash University
Australia
Alpbach Forum
15-20 August, 2008
kesten@kestencgreen.com
ForPrin.com
2
Day 6: Outline of Today’s Seminar
“Forecasting elections, climate,
and polar bears”
Forecasting election outcomes
Do forecasting principles apply to natural science (my) problems?
The problem of forecasting long-term climate
Carter DVD
Tea Break
Forecasting audit of IPCC’s forecasts
Polar bear population forecasts
3
Alan Lichtman’s Keys to the Whitehouse model
4
Randy’ Jones’s paper on equal weights models for election forecasting
5
Armstrong & Graefe paper on forecasting using politicians policy positions
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Combining forecasts
from valid methods
improves accuracy
See pollyvote.com
21
Your questions
I _ "
22
‘Forecasting principles
don’t apply to my discipline’
Hogarth tells of having met “[Ray] Fair in France in the 1970s and telling
him about equal weights results. Our conversation clearly had no impact”
(Hogarth 2006, 19). Similarly, Dana and Dawes ruefully point out that
“Numerous pleas to investigators to consider alternatives to regression
coefficients have fallen on deaf ears” (2004, 329).
‘Your principles don’t apply to us’ is a common opinion
My response: Prove it! Show me the evidence!
So far, no evidence
23
Refresher on principles
Principles are based on
§ evidence from research in many disciplines
§ standard practice where not contradicted by evidence
True that not all principles are relevant to all problems
§ many principles are conditional
Lenny drew a distinction between ‘scientific forecasting’ and
‘economic forecasting’. I think he meant ‘forecasting for the natural
sciences’ and ‘forecasting for the social sciences’.
When I refer to ‘scientific forecasting’, I mean forecasting
using methods that are based on empirical research comparing
the performance of reasonable alternative methods.
Little evidence comes from natural sciences
§ Why?
24
Possible solutions to domain jealousy’?*
1. Special interest groups (SIGs)
2. Collaborations between forecasters and domain experts
3. Willingness of domain experts (e.g. physical scientists) to
audit their forecasting procedures using principles
a. With open minds
b. Provide evidence for
i. New conditions
ii. New principles
*Making a start:
a. Collaborations (2, 3) with Harvard astrophysicist Willie Soon.
b. Lenny’s students?
25
Wave Pilkey & Pilkey-Jarvis
26
Show Carter DVD
27
TEA BREAK
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
Audit of the polar bear forecasts
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
ResearchGate has not been able to resolve any citations for this publication.
ResearchGate has not been able to resolve any references for this publication.