forked from mike19106/model_thinking
-
Notifications
You must be signed in to change notification settings - Fork 0
/
model_thinking.1
130 lines (93 loc) · 3.1 KB
/
model_thinking.1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
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
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
Model Thinking
==============
https://www.coursera.org/modelthinking/class/
http://ccl.northwestern.edu/netlogo/
http://modelpractice.wordpress.com/
http://openstudy.com/study#/groups/Model%20Thinking%20%28Coursera%29
http://www.geazle.com/group/21
Section 1 - Introduction: Why Model?
====================================
Intelligent citizens of the world
---------------------------------
All models are wrong, but some are useful
models are the new lingua franca - business, politics, non-profit
models clarify logic
give us the conditions under which certain things hold true
two heads are better than one, versus
too many cooks spoil the broth
opposite proverbs website
models are better than we are
y - discrimination
x - calibration
hedgehogs use a single model
terrible at predicting
hedge-foxes
fox-hogs
foxes - use lots of loose models in their heads
better at predicting
mindless competition
formal models
models are fertile
they can be applied to multiple, distinct domains
models make us humble
we're much worse at predicting than we think we are
they make us see the full dimensionality of the problem
Many-model thinkers know best
foxes did better than hedgehogs
formal models did better than foxes
lots of formal models did better than individual formal models
Thinking more clearly
---------------------
developing a model
name the RELEVANT parts
identify the relationships (between the parts)
Choosing a restaurant
name the parts
restaurant
people going
how much money they have
how much time
preferences
cuisine
models allow us to
inductively explore
understanding class of outcome
equilibrium
cycle
random
complex
identify logical boundaries
the conditions under which one thing holds, and another doesn't
when is it the case that two heads are better than one?
when is it the case that too many cooks spoil the broth?
communicate
Using and Understanding Data
----------------------------
Why you model
1) understand patterns
flat line, sloped line, wave cycle, spikes, etc
2) predict points
if houses cost $100/sqft, a 2000 sqft house is $200,000
3) produce bounds
inflation 10 years from now will be between 0-3%
4) retrodiction
5) predict other stuff
derive unknown variables based on known variables
predict previously unknown things
6) inform data collection
name the parts
7) estimate hidden parameters in the model
8) calibrate
model refinement
using the model to tell us more about the data
using the data to tell us more about the model
Using Models to Decide, Strategize and Design
---------------------------------------------
1) real time decision aids - help figure out when to intervene (or not)
2) comparative statics - what's likely to happen if we make this choice?
moving from one equilibrium to another
3) counterfactuals - what would have happened if we hadn't chosen that course of action?
4) identify and rank levers - which entity/factor among many has outsized impact
5) experimental design -
6) institutional design - should we have a market? a democracy? a bureaucracy?
7) help choose among policies and institutions