Model-Free Methods

Sometimes you can't build a credible model of the project — it's too novel, too tangled, too unlike anything before. You can still learn what works, straight from experience, with no theory at all.

Model-free methods take the opposite bet to Chapter 15. Instead of learning how the environment works and planning against it, they learn which actions are good directly from the stream of experience — bypassing the model entirely. There's no transition model to estimate, no planning step, and crucially no model error to propagate. You act, you observe how it turned out, and you adjust your sense of which actions pay off. That's it.

Figure 1. Model-based methods (top, faded) take the detour through a learned model and a planning step. Model-free methods (bottom) map experience straight to the value of actions — simpler, and immune to model error, at the cost of needing more experience.

1Temporal-difference learning: don't wait for the end

The engine of model-free learning is the temporal-difference update, and its insight is worth dwelling on. You don't wait until a project finishes to judge whether a decision was good. The instant things turn out better or worse than you expected, you nudge your estimate of that situation's value toward the reward you just saw plus your revised estimate of where you've landed. You learn from the surprise, immediately, rather than waiting for the final tally. This "bootstrapping" — updating an estimate using other estimates — is what lets model-free methods learn continuously from an ongoing stream.

2Q-learning and its relatives

Q-learning applies this to action values: it learns, for each situation, how good each available action is, updating those values from observed transitions. It has a useful property — it can learn the value of the best policy even while behaving more cautiously to explore, which matters when experimenting for real carries risk. Its on-policy cousin evaluates the policy you're actually following. For the enormous state spaces of real problems, these values are represented with the function approximation of Chapter 7 — the pairing that produced the well-known deep reinforcement-learning systems.

Key idea

Model-free learning skips the theory and maps experience straight to "which actions pay off here". It needs more experience than a model-based approach, but it carries no model to be wrong — a fair trade when the environment defies clean modelling.

What it means in practice

When a project or programme is too novel or complex to model credibly, you can still improve — by learning directly which responses tend to work out in which situations, accumulated across real outcomes. The temporal-difference habit is the transferable one: revise your read on how a situation is going the moment it beats or misses expectations, instead of withholding judgement until the post-mortem. It is learning from experience without needing a theory of why — slower to converge, but honest about how little you can model.

Handbook application: from concept to controlled practice

Purpose. This expanded section turns the original page into a practical handbook. It preserves the supplied material and adds a repeatable way to apply, check and review Model-Free Methods. It does not replace a contract, legislation, a controlled standard, competent engineering judgement or specialist advice.

The operating aim is to convert the subject into a governed decision, owned work, usable evidence and a reviewable outcome. Read the original explanation first, then use the workflow and checks below to convert knowledge into evidence.

Use Model-Free Methods as a decision instrument rather than an administrative form. The subject terms—learning, temporal-difference, q-learning, model-free, experience—need an explicit connection to the project objective, business value and stakeholder commitments. Before completing the artefact, write one sentence stating who will use it, what decision it supports and when that decision is required.

Apply a disciplined information model. Separate facts supported by evidence, forecasts derived from a method, assumptions awaiting validation, constraints that limit choice, risks that may occur, issues that already exist and actions assigned to people. Each material entry should have an owner, date, status and next review point. Where probability or impact scores are used, define the scale so different reviewers interpret it consistently.

A baseline is useful only when changes are visible. Give the artefact an identifier, version, approval state and effective date. Define which changes require reapproval, how superseded versions are retained and where supporting evidence is stored. During reviews, focus on exceptions, decisions and trends rather than reading every field aloud. Record the decision and rationale, not merely that a meeting occurred.

Close the loop beyond delivery. Confirm acceptance criteria, unresolved items, transferred responsibilities and operational ownership. Where benefits are expected, identify the outcome measure, baseline, target, observation period and owner who remains accountable after the project team disbands. Lessons should describe the condition, consequence and reusable action; a generic statement such as “communicate better” cannot improve the next project.

Step-by-step operating method

Clarify the decision. Name the outcome, sponsor, affected stakeholders and decision that this work must enable.
Set boundaries. Record scope, assumptions, constraints, dependencies, tolerances and escalation conditions.
Plan the evidence. Define deliverables, measures, owners, due dates and acceptance criteria before execution.
Control delivery. Compare actual performance with the baseline, assess changes and manage risks and issues explicitly.
Close the loop. Confirm acceptance, transfer ownership, capture lessons and track benefits beyond handover.

Completion and governance protocol

Start with a short drafting workshop involving the accountable owner and the people who hold the evidence. Complete high-consequence fields first: objective, scope, owner, baseline, acceptance, dependencies and escalation. Mark unknowns as assumptions or actions rather than hiding them behind vague prose. Circulate a review draft, resolve conflicting interpretations, baseline the approved version and place the next review date in an owned schedule.

Information type	Minimum useful content	Review test
Outcome	Observable change and intended recipient	Not merely a deliverable or activity
Measure	Definition, baseline, target, frequency and source	Two reviewers would calculate it the same way
Ownership	One accountable role plus contributors and approver	Authority matches responsibility
Uncertainty	Assumption, risk or issue with response and trigger	Status reflects current reality
Control	Version, approval, review date and change rule	Current baseline is identifiable

Common failure modes and recovery actions

1. Watch for

Producing a document with no named decision or accountable owner.

Recovery: Return to the governing definition or requirement and restate the decision in one sentence.

2. Watch for

Mixing risks, current issues, assumptions and actions in one unstructured list.

Recovery: Separate evidence from assumption, assign an owner and set a date for validation.

3. Watch for

Measuring activity or output while leaving the intended outcome undefined.

Recovery: Run a small counterexample, boundary test, pilot or independent check before proceeding.

4. Watch for

Accepting changes without evaluating effects on value, scope, schedule, cost and risk.

Recovery: Record the consequence, decision and rationale, then update the controlled baseline.

5. Watch for

Closing the project at delivery even though benefit ownership has not transferred.

Recovery: Escalate when the issue affects safety, compliance, acceptance, material value or an agreed tolerance.

Review checklist

Which decision or commitment does this artefact support?
Who owns each action, risk, acceptance and post-project benefit?
What is the baseline and what variance triggers escalation?
Where is the evidence that the result was accepted and transferred?
Are mandatory requirements distinguished from recommendations and illustrative values?
Are sources, assumptions, units, dates and versions recorded closely enough to reproduce the decision?
Have safety, legal, ethical, stakeholder and operational consequences been considered at the appropriate level?
Is there a named owner and a trigger for review, escalation, change or retirement?

Questions for deeper application

What is the most important distinction a practitioner must preserve when applying Model-Free Methods?