Notes on the lectures from the course (2023) Intelligence Failures. INCLUDES notes from lectures 1-6
(Total: 12 pages).
Intelligence Failures Lecture Notes (Lectures 1-6)
Table of Contents
Lecture 1: Introduction 1
Lecture 2: Warning Problems - Diagnosing Intelligence Failures 3
Lecture 3: Step 2 - Warning Scenario 5
Lecture 4: Step 3 - Critical Indicator 7
Lecture 5: Step 4 - The Intelligence Collection Plan 9
Lecture 6: Alternative System 10
, 1
Lecture 1: Introduction
Introduction on Warning
Indicator & Warning: Basic model of indicator-based intelligence for a warning analysis.
Intelligence’s purpose:
● Warning is a fundamental reason for intelligence activity.
● Aim = prevent a threat from coming to fruition.
● Why do we warn?
○ Provide early warning of potential threats by an in-depth study of a developing
situation.
○ Inform consumers/decision-makers of future developments in time to take action.
○ Reduce the effects of adverse developments.
What causes an intelligence failure?
1. Primary cause = lack of intelligence.
➔ Tasking too limited (e.g., UK/Iraq).
➔ Special means of collection are inadequately used (e.g., NL/Srebrenica).
➔ Events NOT planned.
➔ Encryption, one-time pads (e.g., Pearl Harbour attack).
2. Failure to correctly interpret the upcoming events:
a. Failure to connect the dots.
b. Unavailability of information.
c. Overestimation.
d. Mirror-Imaging: Analysts may perceive/process information through the filter of
personal experience, undermining objectivity.
e. Lack of communication.
f. Overconfidence.
g. Complacency (e.g., Falklands War).
h. Conventional wisdom.
i. Underestimation: Through national/cultural biases (e.g., ignorance, arrogance,
prejudice, ethnocentricity).
➔ Worst case scenario NOT developed, NO accurate critical indicators or a
misinterpretation of critical indicators.
j. Subordination to policy.
3. Failure to evaluate the existing intelligence in time (e.g., overload problem).
➔ Causes:
◆ Encryption = takes too long.
◆ Overload = too much information.
◆ Translation (e.g., fast evolution of the Tuareg language).
◆ Sources in enemy territory CANNOT be reached.
➔ Timeliness = easier to get insight into the long-term aims, BUT harder to gain good
insights into short-term aims.
➔ Level = easier to get insight on the strategic level, BUT harder to gain insight on the
operational & tactical level.
, 2
➔ When could it happen?
◆ Cry wolf mechanism.
◆ Intelligence is filtered by ‘noise barriers’ (e.g., enemy, international
environment).
◆ Self-Generated Noise: When policymakers are NOT able to adjust their
expectations about a party’s intentions & capabilities on the basis of reality.
◆ ‘Sheer Nerve Scenario’: Opponent will NOT have the ‘sheer nerve’ to
conduct an attack.
Intelligence cycle = an intelligence failure is rarely due to 1 single cause.
Warning: Can be associated with 5 possible aspects:
1. Hit: A warning is given + event happens.
2. Miss: NO warning given + event happens (e.g., poor weather on Normandy Beach landing in
WWII).
3. False Alarm: A warning is given, BUT the event does NOT happen.
4. Correct Rejection: A negative warning [or NO warning] is given & the event does NOT
happen (e.g., Nazi Germany was NOT capable anymore of producing nuclear weapons).
➔ Saves resources from being spent.
5. Success: A warning is given + event does NOT happen by intelligence-based intervention.
How to deal with warning issues?
● Analytical methods.
● Based on methodology.
● Methodological insight in the method used.
Explanatory research:
➔ Characteristics & demands:
◆ Alpha (α): Chance that you incorrectly conclude that there is a significant
relationship between phenomena.
● Social sciences 5%.
● Intelligence 10-90% depending on the issue at hand.
◆ Beta (β): Chance that you do NOT discover a relationship between phenomena.
● Social sciences 20%.
● Intelligence 0.01-5% depending on the issue at hand.
◆ Analytical accuracy: Degree of closeness/agreement of the assessed course of action
to its actual (true) course of action.
● E.g., a broken watch with the right time twice a day.
◆ Analytical Confidence: Probability/certainty that the actual (true) course of action
will be covered by a described variation of the assessed course of action.
● E.g., a broken watch with the actual time within a 6-hour margin of error.
➔ Warning methods always have issues of α & β reduction. Emphasis depends on the RQ &
release process’ phase:
● Warning problem = to NOT miss a threat → reduction β.
● Critical indicators = if a scenario is unfolding → reduction α Overall.
(Total: 12 pages).
Intelligence Failures Lecture Notes (Lectures 1-6)
Table of Contents
Lecture 1: Introduction 1
Lecture 2: Warning Problems - Diagnosing Intelligence Failures 3
Lecture 3: Step 2 - Warning Scenario 5
Lecture 4: Step 3 - Critical Indicator 7
Lecture 5: Step 4 - The Intelligence Collection Plan 9
Lecture 6: Alternative System 10
, 1
Lecture 1: Introduction
Introduction on Warning
Indicator & Warning: Basic model of indicator-based intelligence for a warning analysis.
Intelligence’s purpose:
● Warning is a fundamental reason for intelligence activity.
● Aim = prevent a threat from coming to fruition.
● Why do we warn?
○ Provide early warning of potential threats by an in-depth study of a developing
situation.
○ Inform consumers/decision-makers of future developments in time to take action.
○ Reduce the effects of adverse developments.
What causes an intelligence failure?
1. Primary cause = lack of intelligence.
➔ Tasking too limited (e.g., UK/Iraq).
➔ Special means of collection are inadequately used (e.g., NL/Srebrenica).
➔ Events NOT planned.
➔ Encryption, one-time pads (e.g., Pearl Harbour attack).
2. Failure to correctly interpret the upcoming events:
a. Failure to connect the dots.
b. Unavailability of information.
c. Overestimation.
d. Mirror-Imaging: Analysts may perceive/process information through the filter of
personal experience, undermining objectivity.
e. Lack of communication.
f. Overconfidence.
g. Complacency (e.g., Falklands War).
h. Conventional wisdom.
i. Underestimation: Through national/cultural biases (e.g., ignorance, arrogance,
prejudice, ethnocentricity).
➔ Worst case scenario NOT developed, NO accurate critical indicators or a
misinterpretation of critical indicators.
j. Subordination to policy.
3. Failure to evaluate the existing intelligence in time (e.g., overload problem).
➔ Causes:
◆ Encryption = takes too long.
◆ Overload = too much information.
◆ Translation (e.g., fast evolution of the Tuareg language).
◆ Sources in enemy territory CANNOT be reached.
➔ Timeliness = easier to get insight into the long-term aims, BUT harder to gain good
insights into short-term aims.
➔ Level = easier to get insight on the strategic level, BUT harder to gain insight on the
operational & tactical level.
, 2
➔ When could it happen?
◆ Cry wolf mechanism.
◆ Intelligence is filtered by ‘noise barriers’ (e.g., enemy, international
environment).
◆ Self-Generated Noise: When policymakers are NOT able to adjust their
expectations about a party’s intentions & capabilities on the basis of reality.
◆ ‘Sheer Nerve Scenario’: Opponent will NOT have the ‘sheer nerve’ to
conduct an attack.
Intelligence cycle = an intelligence failure is rarely due to 1 single cause.
Warning: Can be associated with 5 possible aspects:
1. Hit: A warning is given + event happens.
2. Miss: NO warning given + event happens (e.g., poor weather on Normandy Beach landing in
WWII).
3. False Alarm: A warning is given, BUT the event does NOT happen.
4. Correct Rejection: A negative warning [or NO warning] is given & the event does NOT
happen (e.g., Nazi Germany was NOT capable anymore of producing nuclear weapons).
➔ Saves resources from being spent.
5. Success: A warning is given + event does NOT happen by intelligence-based intervention.
How to deal with warning issues?
● Analytical methods.
● Based on methodology.
● Methodological insight in the method used.
Explanatory research:
➔ Characteristics & demands:
◆ Alpha (α): Chance that you incorrectly conclude that there is a significant
relationship between phenomena.
● Social sciences 5%.
● Intelligence 10-90% depending on the issue at hand.
◆ Beta (β): Chance that you do NOT discover a relationship between phenomena.
● Social sciences 20%.
● Intelligence 0.01-5% depending on the issue at hand.
◆ Analytical accuracy: Degree of closeness/agreement of the assessed course of action
to its actual (true) course of action.
● E.g., a broken watch with the right time twice a day.
◆ Analytical Confidence: Probability/certainty that the actual (true) course of action
will be covered by a described variation of the assessed course of action.
● E.g., a broken watch with the actual time within a 6-hour margin of error.
➔ Warning methods always have issues of α & β reduction. Emphasis depends on the RQ &
release process’ phase:
● Warning problem = to NOT miss a threat → reduction β.
● Critical indicators = if a scenario is unfolding → reduction α Overall.
