Overfitting is a model that explains historical data too closely will fail on new data. It sits in the Cognition dimension (COG) of the Human Behavior Taxonomy™ as element HBT-COG-0510, within the Learning family. The core principle: a model that explains historical data too closely will fail on new data. In incentive terms, it matters because it changes the payoff people perceive before they choose — which means it can be designed for, or exploited.
Scientific Definition
A model that explains historical data too closely will fail on new data.
Plain-English Definition
A model that explains historical data too closely will fail on new data.
Feynman Explanation
You memorized the test. You did not learn the subject.
Core Principle
A model that explains historical data too closely will fail on new data.
Mechanisms
Pending editorial review.
A model that explains historical data too closely will fail on new data.
Pending editorial review.
Pending editorial review.
Pending editorial review.
Pending editorial review.
Complex models often mistake noise for signal.
Inputs (Triggers)
Pending editorial review.
Outputs (Behaviors)
Pending editorial review.
Behavioral Signature
You memorized the test. You did not learn the subject.
Examples
- A sales forecast that perfectly matches last year but fails next quarter.
- Complex models often mistake noise for signal.
Pending editorial review.
Original analysis from The Incentives Lab — how this element behaves inside real payoff structures.
Why this element matters to incentive design
Most organizations meet this element as a personnel problem. It is not one. The mechanism underneath it is straightforward: a model that explains historical data too closely will fail on new data. You can recognize it in the field by its signature: you memorized the test. You did not learn the subject. Every element in the Cognition dimension changes the perceived payoff of an action before the action happens, which is exactly where incentive design has leverage.
How it gets exploited
Left undesigned, complex models often mistake noise for signal. It is amplified whenever complex models often mistake noise for signal. Inside organizations that shows up as complex models often mistake noise for signal. The pattern is the same one Goodhart's Law describes: the measurable proxy attracts the effort, and the purpose behind it quietly loses funding.
How the Lab designs around it
The redesign move is to test models on out-of-sample data and prefer simplicity. Watch for it at the boundaries: handoffs, promotions, incident reviews, and budget cycles are where this element gets its power.
Famous Experiments
Pending editorial review.
Design Principles
- Test models on out-of-sample data and prefer simplicity.
Measurement Approaches
Pending editorial review.
Evidence
Pending editorial review.
Pending editorial review.
The Perverse Incentive Lens™
How this behavior is exploited — and how to redesign around it.
- Test models on out-of-sample data and prefer simplicity.
Pending editorial review.
Pending editorial review.
Interactive Mini Network
Click any neighbor to re-center the graph and follow the threads of connection.
Knowledge Graph Neighbors
Auto-linked to the rest of the Human Behavior Taxonomy by family, domain, dimension, and shared keywords.
Your capacity to learn, unlearn, and relearn faster than the environment changes.
Malcolm Knowles's principles of adult learning: self-direction, experience as resource, readiness, problem-centered orientation, intrinsic motivation.
Amishi Jha's research: attention is a finite, depletable, but trainable cognitive resource — short daily mindfulness measurably improves it.
Approach a situation as if you knew nothing.
Unusual or bizarre material is more memorable than common material.
Hierarchy of cognitive learning: Remember → Understand → Apply → Analyze → Evaluate → Create. Benjamin Bloom, 1956 (revised 2001).
Wallace J. Nichols: proximity to water reliably shifts the nervous system toward a calmer, more creative state.
Resourcing past memories with the state you needed at the time, so old triggers stop firing the old response.
Grouping information into familiar units to improve memory and processing.
Compressing many small patterns into one larger unit your mind treats as a single object.
Anchor a peak resourceful state to an imagined spot on the floor; step into it on demand before high-stakes moments.
Associating a neutral stimulus with a meaningful one until the neutral one triggers the response.
Where Overfitting is cited in the corpus
Essays, field guides, and diagnostics from The Incentives Lab that apply this element.
- EssayGoodhart's Law in the Real World
How measurable proxies capture judgment.
- EssayThe Perverse Incentives Hiding in Your KPIs
Cognitive shortcuts turned into scorecards.
- EssayAI Agents Inherit Your Incentives
How this element propagates into automated systems.
- ReferenceThe Periodic Table of Human Behavior
The full 1,267-element map this page belongs to.
- CourseIncentives 101
The free ten-part primer on reading a payoff structure.
Questions about Overfitting
- What is Overfitting?
- Overfitting is a model that explains historical data too closely will fail on new data. It sits in the Cognition dimension (COG) of the Human Behavior Taxonomy™ as element HBT-COG-0510, within the Learning family. The core principle: a model that explains historical data too closely will fail on new data. In incentive terms, it matters because it changes the payoff people perceive before they choose — which means it can be designed for, or exploited.
- What is an example of Overfitting?
- Complex models often mistake noise for signal. The Incentives Lab catalogs everyday, organizational, and historical instances of this element on its Human Behavior Taxonomy™ page (HBT-COG-0510).
- How is Overfitting exploited?
- Complex models often mistake noise for signal.
- How do you design around Overfitting?
- Test models on out-of-sample data and prefer simplicity.
- Which behavioral dimension does Overfitting belong to?
- Overfitting is classified in the Cognition dimension (COG) of the Human Behavior Taxonomy™, family "Learning", class "Mental Model". Its permanent identifier is HBT-COG-0510 and its evidence grade is B.