The rules that govern every Fortune Favors the Prepared analytical product, and a plain-language explainer of the intelligence-community frameworks they rest on. Operational discipline, anti-bias discipline, and tradecraft standards — what they are, and why they hold.
Fortune Favors the Prepared analytical products follow a defined set of operational, anti-bias, and tradecraft rules. The standards on this page govern every published product. They are not aspirational. The pre-delivery verification stack enforces them, the operator reviews them, and the reader can evaluate compliance directly from the product.
This document does two things. First, it explains the discipline in plain language — what analytic tradecraft is, and what the intelligence-community frameworks behind it actually say. Second, it records the FFTP house rules in three groups: Operational Discipline governs how analysis is presented to the reader; Anti-Bias Discipline governs how source material is evaluated and represented; Tradecraft Standards govern how analytical judgments are constructed, dated, and revised.
The goal is not to turn readers into intelligence analysts. It is to make our work auditable. When a product says a development is “likely” with “moderate confidence,” carries an alternate hypothesis, or flags a source as needing a framing caveat, those are not stylistic choices. They are the application of a discipline with a long professional pedigree. Understanding that discipline lets you judge our work on its merits — and apply the same thinking to information you encounter anywhere. Published products reference this page rather than repeating the standards inline; where a product carries a specific deviation from a standard, the deviation is documented in the product with rationale.
01What “Tradecraft” Means
Analytic tradecraft is the disciplined set of methods, standards, and habits that separate professional intelligence analysis from opinion, punditry, or a summary of the news. It is concerned less with what an analyst concludes than with how the conclusion was reached: whether the sourcing was sound, whether genuine alternatives were considered, whether uncertainty was expressed honestly, and whether the reasoning can be retraced by someone else.
Tradecraft exists because the human mind is predictably biased. It anchors on first impressions. It sees patterns in noise. It favors evidence that confirms what it already believes and discounts evidence that does not. It grows more confident as it accumulates detail, even when the new detail adds nothing. Structured methods and explicit standards are countermeasures. They force the analyst to slow down, externalize the reasoning, and test a judgment before publishing it.
It helps to separate two layers. Standards are the rules a finished product must satisfy — objectivity, proper sourcing, honest expression of uncertainty, consideration of alternatives, sound argumentation. Techniques are the structured methods an analyst uses to meet those standards — competing-hypotheses analysis, key-assumptions checks, indicators and warning, and others. Standards tell you what good looks like. Techniques are how you get there. The rest of this page walks through both, and then sets out the FFTP house rules that extend them.
02ICD 203 — The Analytic Standards
ICD 203 defines the standards every analytic product is expected to meet. It is the single most important reference behind this page and the source of the likelihood-and-confidence vocabulary we use. It is publicly available on the ODNI website, it is short, and it rewards a direct read.
ICD 203 establishes the standards by which analytic products — and the analysts who write them — are evaluated. Its most practically useful contribution is a set of tradecraft standards that translate directly into a working checklist.
The tradecraft standards
A sound analytic product, per ICD 203, does each of the following. It properly describes the quality and credibility of its sources. It properly expresses and explains uncertainty. It clearly distinguishes between reported information and the analyst’s own assumptions and judgments. It incorporates analysis of alternatives. It is relevant to the decision it informs and addresses the implications. It uses clear and logical argumentation. It explains any change from — or consistency with — prior judgments. It makes accurate judgments. And it uses visual information effectively where that conveys the point better than text.
Read those again as questions to ask of any analysis, including ours: How good are the sources? How uncertain is this? Which part is fact and which is the analyst’s inference? Were alternatives considered? Does the reasoning hold together? Has the judgment changed, and was that acknowledged? That is the checklist.
Likelihood versus confidence — the distinction that matters most
ICD 203 insists that two ideas, routinely confused, be kept separate and both stated explicitly.
These are independent dimensions. An analyst can be highly confident that something is unlikely. An analyst can have low confidence in a judgment that something is very likely. Collapsing the two — saying only “probably” while staying silent on how well-supported the judgment is — hides exactly the information a decision-maker needs.
The standardized likelihood ladder
So that “likely” means the same thing in every product, ICD 203 fixes a ladder of expressions. An analyst picks one term per judgment rather than improvising vocabulary.
The discipline is to choose one term, not to blend a word from the ladder with a separate numeric percentage that implies a precision the evidence cannot support.
Confidence levels
Confidence is expressed in three levels. The essential requirement is that the level be auditable: a reader should be able to look at the cited sourcing and reasoning and see why the analyst chose it.
| Level | What it rests on |
|---|---|
| High confidence | High-quality information and sound reasoning — multiple credible sources, well corroborated, with few or no significant gaps. |
| Moderate confidence | Credibly sourced and plausible, but not fully corroborated, or reasoning that involves some assumptions. A reasonable basis, with room for error. |
| Low confidence | Information that is scant, questionable, fragmented, or resting largely on single sourcing or inference. The judgment is offered, but flagged as weakly supported. |
When different parts of a single judgment rest on different evidentiary footing, the honest move is split confidence — rating each element separately rather than averaging them into one misleading number. A finding might carry high confidence that an event occurred, moderate confidence about its scale, and low confidence about who was behind it. Naming the split tells the reader precisely where the analysis is solid and where it is not.
02ICD 203 — The Analytic Standards
ICD 203 defines the standards every analytic product is expected to meet. It is the single most important reference behind this page and the source of the likelihood-and-confidence vocabulary we use. It is publicly available on the ODNI website, it is short, and it rewards a direct read.
ICD 203 establishes the standards by which analytic products — and the analysts who write them — are evaluated. Its most practically useful contribution is a set of tradecraft standards that translate directly into a working checklist.
The tradecraft standards
A sound analytic product, per ICD 203, does each of the following. It properly describes the quality and credibility of its sources. It properly expresses and explains uncertainty. It clearly distinguishes between reported information and the analyst’s own assumptions and judgments. It incorporates analysis of alternatives. It is relevant to the decision it informs and addresses the implications. It uses clear and logical argumentation. It explains any change from — or consistency with — prior judgments. It makes accurate judgments. And it uses visual information effectively where that conveys the point better than text.
Read those again as questions to ask of any analysis, including ours: How good are the sources? How uncertain is this? Which part is fact and which is the analyst’s inference? Were alternatives considered? Does the reasoning hold together? Has the judgment changed, and was that acknowledged? That is the checklist.
Likelihood versus confidence — the distinction that matters most
ICD 203 insists that two ideas, routinely confused, be kept separate and both stated explicitly.
These are independent dimensions. An analyst can be highly confident that something is unlikely. An analyst can have low confidence in a judgment that something is very likely. Collapsing the two — saying only “probably” while staying silent on how well-supported the judgment is — hides exactly the information a decision-maker needs.
The standardized likelihood ladder
So that “likely” means the same thing in every product, ICD 203 fixes a ladder of expressions. An analyst picks one term per judgment rather than improvising vocabulary.
The discipline is to choose one term, not to blend a word from the ladder with a separate numeric percentage that implies a precision the evidence cannot support.
Confidence levels
Confidence is expressed in three levels. The essential requirement is that the level be auditable: a reader should be able to look at the cited sourcing and reasoning and see why the analyst chose it.
| Level | What it rests on |
|---|---|
| High confidence | High-quality information and sound reasoning — multiple credible sources, well corroborated, with few or no significant gaps. |
| Moderate confidence | Credibly sourced and plausible, but not fully corroborated, or reasoning that involves some assumptions. A reasonable basis, with room for error. |
| Low confidence | Information that is scant, questionable, fragmented, or resting largely on single sourcing or inference. The judgment is offered, but flagged as weakly supported. |
When different parts of a single judgment rest on different evidentiary footing, the honest move is split confidence — rating each element separately rather than averaging them into one misleading number. A finding might carry high confidence that an event occurred, moderate confidence about its scale, and low confidence about who was behind it. Naming the split tells the reader precisely where the analysis is solid and where it is not.
04Structured Analytic Techniques
The Primer catalogs the named, repeatable methods analysts use to meet the standards in ICD 203.
Structured analytic techniques are step-by-step methods that externalize and discipline reasoning. They make thinking visible — to the analyst, to a reviewer, and to the reader — and they create friction against cognitive bias. They are usually grouped by what they do.
Diagnostic techniques — testing the current thinking
Key Assumptions Check
The analyst lists every assumption the current line of analysis depends on, then stress-tests each one. Is it well-founded? What happens to the judgment if it is wrong? This is often the highest-value technique, because faulty assumptions are a leading cause of analytic failure.
Quality of Information Check
A deliberate re-examination of the sourcing behind the key judgments — reliability, access, corroboration, and the possibility of deception. It is the working counterpart to the ICD 206 sourcing standard.
Indicators / Signposts of Change
The analyst defines, in advance, observable events that would signal a situation is moving toward one outcome rather than another. Pre-defined indicators guard against rationalizing new information into an existing view, and they turn a static judgment into a monitoring framework.
Contrarian techniques — challenging the current thinking
Devil’s Advocacy
An analyst deliberately builds the strongest possible case against the prevailing judgment — not to be contrarian for its own sake, but to expose weak evidence and untested assumptions in the consensus view.
What-If Analysis and Red-Teaming
What-If analysis assumes a surprising event has already happened and works backward to explain how, surfacing pathways the main analysis discounted. Red-teaming models how an adversary would actually think and act, rather than projecting the analyst’s own assumptions onto them.
Imaginative techniques — generating new thinking
Alternative Futures and Scenarios
Rather than predicting a single outcome, the analyst develops several plausible futures driven by the key uncertainties. This is valuable for highly uncertain, long-horizon questions, where a single point prediction is unreliable and a range of prepared-for possibilities is more useful.
Analysis of Competing Hypotheses
Analysis of Competing Hypotheses — ACH — deserves its own treatment. It is the structured technique most directly tied to the ICD 203 requirement to consider alternatives, and it is the discipline behind our standing rule that every assessment carry genuine alternate hypotheses. It originates with Richards Heuer’s work in Psychology of Intelligence Analysis.
The central insight is a reversal. Intuitive analysis tends to start with the most plausible explanation and then look for evidence that supports it — a process wide open to confirmation bias. ACH inverts this. Instead of asking which hypothesis has the most evidence for it, ACH asks which hypothesis is least contradicted by the evidence. The analyst sets out to disconfirm and eliminate, not to confirm.
- Identify the hypotheses. Generate a full set of mutually exclusive explanations — enough to cover the realistic possibilities, including ones that seem unlikely at first.
- List the evidence and arguments. Assemble all relevant evidence, including the absence of evidence where something would be expected.
- Build the matrix. Place hypotheses across the top and evidence down the side. For each cell, mark whether that item is consistent, inconsistent, or not applicable with that hypothesis.
- Refine the matrix. Remove evidence consistent with every hypothesis — it has no diagnostic value. What matters is the evidence that distinguishes.
- Draw tentative conclusions by trying to disprove. Focus on inconsistency. The hypothesis with the least evidence against it is the strongest — not the one with the most evidence for it.
- Test sensitivity. Identify the few items of evidence driving the conclusion, and ask what happens if any of them are wrong or are deception.
- Report conclusions. Present the relative likelihood of all the hypotheses, not just the leading one, with confidence stated per ICD 203.
- Identify indicators for the future. Define what later evidence would confirm or overturn the conclusion, turning the analysis into a monitoring tool.
05FFTP House Rules — Operational and Anti-Bias Discipline
The frameworks above were designed for classified intelligence consumers. The FFTP product is designed for the household practitioner. Both audiences deserve rigorous analytical product; the rules below adjust the discipline to the audience. The first two groups govern how analysis is presented and how source material is evaluated.
Operational Discipline
How analysis is presented to the reader — the standards governing the operational utility of the product.
No Gatekeeping
Our readers get the same operational picture a federal practitioner would see. We do not withhold operational detail behind credentialing or reader-comfort considerations. PREP-CON assessment is not softened for comfort. If the operational basis supports a given level, the DTR states that level.
No Dunning-Kruger Enablement
We do not publish confidence our sources cannot sustain. Every confidence level is auditable against the cited sources. Split confidence is applied when different aspects of a finding sit at different evidentiary levels. False precision is a failure mode; we name it and reject it.
No Fear Marketing
Preparedness products are sold on operational value, not on amplified threat. The analytical standard prohibits inflating threat to drive sales or attention. If the threat is real, the analysis carries it. If it is not, the analysis does not invent it. Subscription supports the analytical work; the household work is what matters.
Operator Audience Standard
Written for the practitioner who needs to make decisions on the analysis — not for the news consumer, the policy advocate, or the entertainment reader. The standard is operational utility. If a paragraph does not help the reader make a better decision, it is cut.
Household-as-Unit Standard
Analysis serves the household as the unit of resilience. Aggregate national-level analysis is means, not end. Specific household actions calibrated to household situation and time horizon are the deliverable. The household decision is the test of whether the analysis is complete.
Self-Disclosure of Limits
Gaps acknowledged in the sweep manifest are surfaced in the analytical product, not buried. “Information not available” is a valid finding when corroboration cannot be obtained. The reader can see what we know, what we do not know, and where the boundary sits.
Anti-Bias Discipline: Source-Symmetric Skepticism
How source material is evaluated and represented — the standards governing analytical neutrality.
Source-Symmetric Skepticism
The same evidentiary standard applies to all state actors. Iranian state media, Israeli state and military sources, US state and military sources, and Chinese state sources are all flagged when cited. Single-source state-media claims are classified LOW CONFIDENCE absent multi-stream corroboration regardless of which state is the source. State framing is reported attributed; it is not adopted as DTR voice. Loaded language is stripped from DTR narrative unless quoted with attribution. The same skepticism applied to adversary state media is applied to allied state media. This standard does not produce false equivalence — it produces calibrated confidence.
Neutrality Across Political Spectrum
Products serve households across the political spectrum. Partisan framing is excluded. Political intent attribution is out of scope. Cause analysis (why a decision was made) is treated separately from impact analysis (what the decision produces). Specific policy responses are assessed on operational characteristics rather than partisan framing. Analyst opinions on political questions are not represented in published products.
No Anonymous Sourcing for Load-Bearing Claims
Every load-bearing claim has a named, citable source with a live hyperlink. Anonymous sourcing is acceptable for color and context; it is not acceptable for the analytical spine. “Sources say” framing is not used as a substitute for evidence we can show the reader.
Disconfirming Evidence Requirement
Every analytical judgment carries both supporting and disconfirming evidence in separate sections. The reader can evaluate the strength of the judgment by weighing both sides directly. One-sided argument is not analytical product; it is advocacy. The disconfirming section is not a defensive caveat; it is part of the analytical work.
Two-Alternates Rule
Every analysis carries at least two explicit alternate hypotheses. Each alternate carries its own supporting evidence and its own disconfirming evidence on equal footing with the primary analysis. An alternate without supporting evidence is a strawman and is rejected. An alternate without disconfirming evidence is unfalsifiable and is rejected. The reader can compare three lines of analysis directly: the primary and the two alternates, each with its own evidence trail. Where genuinely distinct lines of analysis exceed two, additional alternates are added rather than collapsed into composites.
06FFTP House Rules — Tradecraft Standards
How analytical judgments are constructed, dated, and revised — the standards governing analytical hygiene. These rules are the working expression of ICD 203 and ICD 206 inside the FFTP production pipeline.
Estimative Language Discipline
Per ICD 203. Likelihood and confidence are separate dimensions; both are stated explicitly. Likelihood uses standardized terms (almost certainly, very likely, likely, roughly even chance, unlikely, very unlikely, almost no chance). Confidence uses HIGH, MODERATE, or LOW. They do not collapse into each other. Forward-looking judgments without estimative language are flagged for revision.
Reporting Distinguished From Judgment
Per ICD 203. The reader can always tell what is reported fact, what is an analyst assumption, and what is analyst inference. Reported intelligence carries its source; assumptions are labelled as assumptions; the analytical judgment is set apart from the reporting it rests on. Presenting an assumption as if it were reported fact — assumption laundering — is a pre-delivery fail. SOP §0.113-C enforces.
Customer Relevance and Implications
Per ICD 203. Every analytical finding states what it means for the reader’s decision. Analysis that does not change a household or practitioner decision is cut. The implication, not just the fact, is the deliverable. This is the analytic-standards expression of the Operator Audience and Household-as-Unit standards above. SOP §0.113-C enforces.
Change From Prior Judgment
Per ICD 203. When a judgment shifts from a prior product — a confidence change, a likelihood change, a directional reversal — the change is stated and explained, not buried. A carry-forward entry whose scored judgment differs from the last cycle carries an explicit change line. Burying a shifted judgment is a failure mode the standard exists to prevent. SOP §0.113-C enforces.
Effective Visual Information
Per ICD 203. Charts, maps, condition banners, and geographic indices are used wherever they convey the operational picture more clearly than text. Visual severity encoding matches the written confidence rating. SOP §0.103 and §0.76 enforce.
Source Characterization
Per ICD 206. Every load-bearing claim carries a named, citable source characterized by tier, access, and any bias or datedness that bears on credibility. The FFTP Master Source Registry — its three-tier classification, its adversary-state-media framing caveat, and its aggregator-stripping policy — is the working implementation of ICD 206. Single-source state media is classified LOW CONFIDENCE and attributed, never adopted as product voice.
Date Coherence
Never claim “deadline now passed” or “T+N past” for dates that are mathematically in the future. The stale-string hunt is daily, not optional. The pre-delivery date-coherence audit verifies that every dated claim is consistent with the cycle publication date. SOP §0.85 enforces.
Stale-String Hunt
Claims valid in prior cycles but superseded by new information are hunted, not allowed to persist. The stale-strings blacklist is the institutional memory; the daily hunt verifies zero hits across all product source files. SOP §0.86 enforces. A pattern caught once enters the blacklist permanently until proven unstale.
Primary Source Sweep
Every cycle requires a fresh read of primary sources before claims that depend on them. The cited timestamp is explicit in the sweep manifest. No carryover claims without re-verification. Authoritative sources (SPC, CISA KEV, WASDE, SWPC, USDA, EIA, IEA, BLS, Federal Reserve, FDA) are read at source rather than via aggregator.
Reporting Period Anchor
The masthead reporting period is anchored to the actual last 24 to 36 hours. Stale multi-day labels are forbidden. “Since invasion” framing is not used where “this cycle” framing is required. The reader knows exactly what window the analysis covers. SOP §0.45 enforces.
Pre-Delivery Verification Gate
Present-files is gated behind a verification stack. Confident-sounding declarations in chat are not a substitute for running the checks. When the operator asks “are you sure,” the gate re-runs; it does not recall from memory. SOP §0.75 enforces.
Visual Inspection Standard
100 percent of pages of every product are viewed before delivery. Sampling is not acceptable. Hooks do not catch rendering failures (empty cells, column overflow, layout regression, content bleeding past box border); only visual inspection does. SOP §0.76 enforces. Coverage is documented in the §0.83 checklist with concrete page counts.
Self-Criticism and Revision
Analytical judgments are subject to revision as new information emerges. Today’s HIGH CONFIDENCE may be tomorrow’s MODERATE CONFIDENCE; the product acknowledges this rather than defending past calls. Revised judgments are stated cleanly; the prior judgment is not hidden. Readers can audit the analytical trail across cycles.
Cross-Reference Discipline
Claims that depend on other claims explicitly cross-reference them. The reader can trace the dependency chain. Internal coherence is enforced through the pre-delivery audit. Inconsistent claims across sections of the same product are surfaced before delivery.
07Why This Matters for the Practitioner
None of this is academic. The standards exist because analytic failure has real costs, and the costs land on whoever relied on the analysis. For the household or organization making preparedness decisions, the same discipline is what separates a useful threat picture from noise.
Three habits carry over directly to any reader. First, demand that uncertainty be expressed honestly — both how likely something is and how solid the basis is — and be skeptical of confident-sounding analysis that never says how it might be wrong. Second, ask which alternatives were considered; a single-explanation account that never engages a competing one is advocacy, not analysis. Third, check the sourcing — whether load-bearing claims trace to named, credible sources, and whether adversary or interested sources are flagged rather than passed off as neutral fact.
When you read a Fortune Favors the Prepared product, you should expect to see all three. That is the point of publishing our standards alongside this explainer: the work is meant to be audited, not taken on faith.
Primary Sources & Further Reading
All items below are unclassified and publicly available; consult the issuing organization’s own site for the current version.
Intelligence Community Directive 203, Analytic Standards — Office of the Director of National Intelligence. The defining document for analytic standards, estimative language, and confidence ratings.
Intelligence Community Directive 206, Sourcing Requirements for Disseminated Analytic Products — Office of the Director of National Intelligence. The companion directive on source characterization and referencing.
A Tradecraft Primer: Structured Analytic Techniques for Improving Intelligence Analysis — CIA Center for the Study of Intelligence. This page draws on all nine ICD 203 tradecraft standards — source quality, expression of uncertainty, reporting distinguished from judgment, analysis of alternatives, customer relevance, clear argumentation, change from prior judgment, accurate judgment, and effective visual information — together with the ICD 206 sourcing requirements.
Richards J. Heuer Jr., Psychology of Intelligence Analysis — CIA Center for the Study of Intelligence. The foundational text behind Analysis of Competing Hypotheses.
Heuer & Pherson, Structured Analytic Techniques for Intelligence Analysis — a practitioner catalog of the methods summarized here.
For supply-chain and critical-infrastructure work, FFTP additionally references NIST Special Publication 800-161 and the CISA ICT Supply Chain Risk Management resources.
Recommended citation: “Analytical Standards and Tradecraft, Fortune Favors the Prepared, v2.0.” Stable URL: https://fortunefavorstheprepared.com/analytical-standards-and-tradecraft/