// From Engineer to Architect · The Field Guide

The Mental Models Atlas

Reference Map · scan as needed Khurram Saleem

From Engineer to Architect · The Field Guide 48 models · 7 categories

An architect makes decisions all day with incomplete information. The difference between a good one and a bad one is rarely more data. It is recognising, in the moment, which invisible force is bending your judgment, and having a name for it. This is the catalogue of those forces.

// the crux

The fourteen disciplines are how an architect behaves. These models are what you must recognise to behave that way. Learn to name them, and you start to see them coming.

// what is in here

Forty-eight models, biases, traps, and effects across seven categories, each in one line.
The Core 10 to master first, if you only have room for ten.
A deep-dive link wherever one of these already has a full essay on the site. Many already do.

An architect on a central platform with light-beams radiating out to labelled model-islands: a cognitive bias detector, a trade-off analysis engine, a SWOT analysis hub, a perspective observatory, a resilience fortress, a risk prediction engine, a learning academy, and a psychological safety garden. — The whole toolkit as one map: each model its own island, all wired back to where the decision gets made.

The same figure shown at three stages of growth along a rising path: an apprentice developer at a single laptop, a skilled software engineer working across several screens, and a software architect standing before a city-scale system diagram. — The same person at three altitudes: apprentice developer, skilled engineer, architect. What changes between the stages is how many of these models they can see in the room.

How to use this. It is a reference, not a read. Skim it once to learn the names, then come back to it before a hard decision and ask which of these is in the room. Start with the Core 10. Follow the deep-dive links where a model has its own essay. It is the toolkit behind the fourteen disciplines; the disciplines are how you behave, these are what you have to see first.

start here

// If you master ten

The Core 10

The highest-leverage models for an architect, drawn from across the categories below. If you only have room for ten, these are the ten. They shape not just technical decisions but people, incentives, risk, and the long game.

1

Cognitive Bias

Recognising that your own judgment is the first thing bending the decision.

Essay → Enemy Number One

2

First Principles Thinking

Breaking assumptions and reasoning up from fundamentals, not analogy.

Deep dive coming

3

Trade-Off Analysis

Every decision optimises one thing while sacrificing another. Name both.

Deep dive coming

4

Second-Order Thinking

The consequences of the consequences. And then what?

Deep dive coming

5

Systems Thinking

Seeing the whole ecosystem and its loops, not the components in isolation.

Deep dive coming

6

Survivorship Bias

Studying the systems that lived while the ones that died stay silent.

Essay → The Planes That Didn’t Come Back

7

Cobra Effect

An incentive that produces the exact opposite of what it intended.

Essay → The Cobra We All Bred

8

Incentive Design (Goodhart’s Law)

A measure that becomes a target stops being a good measure.

Deep dive coming

9

Reversibility & Optionality

Can this be undone, and does it keep your future choices open?

Deep dive coming

10

Probabilistic Thinking

Reasoning in odds and ranges, not false certainties.

Deep dive coming

On the left, the traditional software development lifecycle of requirements, design, development, testing, and deployment; on the right, the architect's vision of trade-offs, multiple perspectives, decision trees, resilience, and business outcomes. — What mastering all this is for: the shift from running the lifecycle to designing the systems and outcomes it serves.

the full atlas

// Category 1 · the mind

Cognitive Biases

Mental shortcuts that distort judgment without your permission. This whole category powers Discipline 1 – cognitive bias is enemy number one.

Confirmation Bias

Seeking evidence that supports beliefs you already hold.

Anchoring Bias

Over-relying on the first piece of information you heard.

Availability Bias

Judging likelihood by what is recent or memorable, not what is true.

Authority Bias

Assuming the senior person is automatically correct.

Halo Effect

Letting one good trait colour your whole judgment of something.

Hindsight Bias

Believing the outcome was obvious all along, once you know it.

Overconfidence Bias

Overestimating what you know or what you can do.

Loss Aversion

The fear of a loss outweighing the pull of an equal gain.

Status Quo Bias

Preferring the existing solution simply because it exists.

Framing Effect

The same facts, presented differently, changing the decision.

Dunning-Kruger Effect

The least competent over-rating their competence the most.

Bandwagon Effect

Believing it because the crowd believes it.

// Category 2 · the agenda

Conscious Biases

Biases we knowingly, or half-knowingly, choose to apply. The cognitive ones happen to you; these you let happen. They power Discipline 11 – convince with data, not ego.

Political Bias

Supporting the decision that increases your own influence.

Department Bias

Optimising for your own team at the system’s expense.

Vendor Bias

Preferring the familiar technology or vendor over the fitting one.

Personal Preference Bias

Choosing a technology because you happen to like it.

Recency Bias

Favouring whatever worked the last time you tried it.

// a conscious bias in the wild

“I know Kubernetes isn’t needed here, but it’s good for my career growth.”

That is closer to a conscious bias than a cognitive one. Nothing about the judgment is mistaken – the engineer is right that it is not needed. The danger is that it will be dressed up as a technical argument in the review, and the real reason will never be said out loud.

// Category 3 · the spiral

Decision-Making Traps

Patterns that pull a group deeper into a bad decision. They power Discipline 8 – reduce decision turnaround, and know when to say no.

Sunk Cost Fallacy

Continuing because of what you have already spent, not what is left to gain.

Escalation of Commitment

Doubling down on a losing path because backing out feels like defeat.

Groupthink

Consensus arriving so smoothly that nobody did the critical thinking.

False Consensus Effect

Assuming everyone quietly agrees with you because nobody objected.

Planning Fallacy

Underestimating the time and effort, every single time.

// Category 4 · the ripple

Systems Thinking Effects

The ones an architect feels most. Consequences travel, incentives bite back, metrics rot. These power Discipline 4 – perspective over perception, the 360 degree view.

Butterfly Effect

A small change creating a wildly disproportionate outcome.

Cobra Effect · Core 10

An incentive producing the exact opposite of what it intended.

Goodhart’s Law · Core 10

A measure that becomes a target stops being a good measure.

Campbell’s Law

The more a metric is used to decide, the more it distorts behaviour.

Second-Order Effects · Core 10

The consequences of the consequences. And then what?

Law of Unintended Consequences

Deliberate actions producing results nobody asked for.

Feedback Loops

Behaviours that reinforce themselves, or quietly balance themselves out.

Tragedy of the Commons

A shared resource depleted because no single actor owns its health.

// Category 5 · the failure

Risk & Failure Models

How systems actually break, and how we fool ourselves about it beforehand. These power Discipline 9 – inverted thinking: ask what could go wrong first.

Survivorship Bias · Core 10

Studying the systems that lived while the ones that died stay silent.

Black Swan Event

A rare, high-impact event that looks obvious only in hindsight.

Normalcy Bias

Assuming the disaster will not happen because it has not yet.

Single Point of Failure

One component whose failure takes the whole system down with it.

Swiss Cheese Model

Several small weaknesses lining up just so to let a failure through.

// Category 6 · the people

Leadership & Organizational Models

The forces that decide whether the truth reaches the decision. These power Discipline 12 and 13 – let the best idea win; be ego-free, blame-free, collaborative.

Psychological Safety

People speaking up with a doubt or a dissent without fear.

Principal-Agent Problem

The decision-maker and the executor pulling on different incentives.

Incentive Misalignment

A reward structure quietly driving the wrong behaviour.

Diffusion of Responsibility

Everybody assuming somebody else owns it, so nobody does.

Peter Principle

People rising until they reach the role they are no longer good at.

// Category 7 · the craft

Architecture-Specific Mental Models

Often more valuable than any technical pattern. These power Discipline 2 and 3 – trade-off analysis, and knowing what pragmatism actually means. Six of the Core 10 live here.

Trade-Off Analysis · Core 10

Every decision optimises something while sacrificing something else.

First Principles Thinking · Core 10

Break the assumptions and reason up from fundamentals.

Reversibility · Core 10

Can this decision be undone, and how cheaply?

Cost of Change

How expensive future modifications will be, designed in now.

Optionality · Core 10

Preserving future choices instead of spending them early.

Second-Order Thinking · Core 10

Thinking past the immediate outcome to what it sets in motion.

Systems Thinking · Core 10

Viewing the whole ecosystem, not the individual components.

Constraints Thinking

Using limits as a forcing function. Innovation often starts here.

Nobody recognises all forty-eight in the moment. That is not the point. The point is that the architect who can name even a handful of these, out loud, in a room full of smart people heading confidently toward a bad decision, is worth more than the one with the deepest stack of technical patterns. The patterns change every few years. These do not. Learn the names. Then learn to say them gently, at the right time.

// carry forward

This is the toolkit. The character that wields it is the other half of the field guide. Start with the discipline that governs all the rest: cognitive bias is enemy number one, and how to decide on merit, not momentum.

// read next

From Engineer to Architect · Article 1

Cognitive Bias Is Enemy Number One

The first model in the Atlas, given the full story: why your own judgment is the bias you cannot see, the four that sit on every architect's shoulder, and how to decide on merit, not momentum.