Business Strategy, Luck, and Poor Judgment

1
Business Strategy, Luck, and Poor Judgment

• Jerker Denrell, Stanford
• Christina Fang, NYU

2
(No Transcript)
3
(No Transcript)
4
(No Transcript)
5
Strategizing and forecasting

• Accurate forecasting of the value of resources
  is the only systematic way to earn above normal
  returns
• Barney, 1986, 1990

6
Forecast accuracy versus skill

• Skill

Accuracy
7
Accurate vs Good Forecasts

• I think there is a world market for about five
  computers.
• Thomas J. Watson,
• Chairman of IBM, 1943

8
Accurate vs Good Forecasts

• I think there is a world market for about five
  computers.
• Thomas J. Watson,
• Chairman of IBM, 1943

• I think there is a world market for about five
  billion computers.
• Thomas Crank,
• Chairman of Wild Ideas, 1943

9
Accurate vs Good Forecasts

• I think there is a world market for about five
  computers.
• Thomas J. Watson,
• Chairman of IBM, 1943
• Accuracy bad
• Judgment good

• I think there is a world market for about five
  billion computers.
• Thomas Crank,
• Chairman of Wild Ideas, 1943

10
Accurate vs Good Forecasts

• I think there is a world market for about five
  computers.
• Thomas J. Watson,
• Chairman of IBM, 1943
• Accuracy bad
• Judgment good

• I think there is a world market for about five
  billion computers.
• Thomas Crank,
• Chairman of Wild Ideas, 1943
• Accuracy good
• Judgment bad

11
Accurate vs Good Forecasts

• If someone predicts that an activity will be
  very successful

12
Accurate vs Good Forecasts

• If someone predicts that an activity will be
  very successful and the prediction turns out to
  be correct

13
Accurate vs Good Forecasts

• If someone predicts that an activity will be
  very successful and the prediction turns out to
  be correct the individual is probably a poor
  forecaster

14
Accurate vs Good Forecasts

• If someone predicts that an activity will be
  very successful and the prediction turns out to
  be correct the individual is probably a poor
  forecaster
• Accurate foresight about the next big thing
  signals bad judgment

15
Forecast accuracy versus skill

• Skill
• 
  

Accuracy
16
Forecast accuracy versus skill

• Skill
• 
  

Accuracy
17
Model

• You observe a noisy signal

18
Model

• You observe a noisy signal
• S m error

19
Model

• You observe a noisy signal
• S m error

20
Model

• You observe a noisy signal
• S m error

21
Model

• Task Make a Forecast

22
Model

• Task Make a Forecast
• Forecast bS

23
Model

• Task Make a Forecast
• Forecast bS

Bayesian b 0.5
24
Model

• Task Make a Forecast
• Forecast bS

Bayesian b 0.5
Overreactor b 1
25
If m 0, distribution of forecasts
Forecast
26
If m 0, distribution of forecasts
Bayesian
Forecast
27
If m 0, distribution of forecasts
Bayesian
Overreactor
Forecast
28
If m 3, distribution of forecasts
29
If m 3, distribution of forecasts
Bayesian
Forecast
30
If m 3, distribution of forecasts
Bayesian
Overreactor
Forecast
31
Accuracy vs Expected Accuracy

• Mean Square Error (MSE)
• E ( forecast actual)2

32
If m 3, what is the expected MSEgiven the
forecast?
33
If m 3, what is the expected MSEgiven the
forecast?
Forecast
34
If m 3, what is the expected MSEgiven the
forecast?
Forecast
35
Experiment

• You observe a noisy signal
• S m error
• Task Make a Forecast

36
Experiment

• You observe Test Sales
• Task Make a Forecast of the Actual Sales

37
Experiment

• You observe Test Sales
• Task Make a Forecast of the Actual Sales
• Can observe 50 previous test sales and
  actual sales

38
Experiment

Artist Test Actual 1001 58.62 75.33 1002 27.68 1
6.17 1003 63.81 70.79 1004 47.86 46.46 1005 43.94
64.67 1006 48.05 44.31 1007 71.04 52.03 1008 54.6
51.37 .
39
Experiment

Artist Test Actual 1001 58.62 75.33 1002 27.68 1
6.17 1003 63.81 70.79 1004 47.86 46.46 1005 43.94
64.67 1006 48.05 44.31 1007 71.04 52.03 1008 54.6
51.37 . Testi mi error (mean of mi
50) Actuali mi error
40
Suppose the actual outcome is between 40 and 60
41
Suppose the actual outcome is between 40 and 60
42
Suppose the actual outcome is above 60
43
Suppose the actual outcome is above 60
44
Conclusion

• Accurate Foresight about the Next Big Thing
  Signals Bad Judgment
• Are entrepreneurs who made money by betting on a
  vision that turned out to be correct worse
  forecasters?