Presentation Patterns — Master Index

Structured reference taxonomy primarily based on Presentation Patterns: Techniques for Crafting Better Presentations (Neal Ford, Matthew McCullough, Nathaniel Schutta, 2013), with supplementary patterns and reinforcements from Presentation Zen (Garr Reynolds, 2nd ed., 2012), Resonate: Present Visual Stories that Transform Audiences (Nancy Duarte, 2010), and Make It Stick: The Science of Successful Learning (Peter C. Brown, Henry L. Roediger III, Mark A. McDaniel, 2014), and The Whole Brain Business Book (Ned Herrmann, 2nd ed., 2015), plus a set of vault-derived patterns observed across the speaker's corpus. Contains 83 named patterns and 28 antipatterns organized by presentation lifecycle phase. See the Sources section at the end of this file for full citations.

This is the primary entry point. The agent reads this file first, then drills into individual pattern files only when detailed descriptions, scoring criteria, or combinatorics are needed.

How to Use This Index

1. During creator Phase 2 (Architecture): Read the phase-grouped lookup table below to find which patterns are relevant at the current phase, then read the pattern files for the ones you want to present as options.
2. During vault Step 3 (Analysis): Scan against observable patterns only — skip patterns marked observable: false (pre-event logistics, physical stage behaviors, post-event follow-up, hidden authoring processes, and external systems that the acquired artifacts cannot directly prove). Every observable entry declares evidence_channels, and every detection needs source-located evidence through one of those channels. For entries with an evidence-source contract, score only when an available source satisfies that contract; otherwise record the entry as not_evaluable rather than guessing or treating it as absent. A polished outcome is not evidence that a named preparation process occurred.
3. During creator Phase 4 (Guardrails): Read all antipatterns and compare against the outline. Render [RECURRING] lines only from Phase 4's emitted recurring_antipatterns records. Use [CONTEXTUAL] for the current-outline scan, which remains enabled without history. Skip unobservable antipatterns.
4. During creator Phase 6 (Publishing / Go-Live): Surface unobservable patterns as a go-live preparation checklist — these are actions to take before, during, and after delivery that the vault cannot score retroactively.

Decision Labels, Confidence, and Binary Pattern Score

Every catalog entry defines three non-arithmetic decision labels. They are criteria for classifying the observed evidence, not point values:

• Strong signal: the evidence meets that entry's Strong signal criterion.
• Moderate signal: the evidence meets that entry's Moderate signal criterion but not its Strong signal criterion.
• Absent: qualifying inspection finds no detected instance under the entry's Absent criterion. For antipatterns, the explicit antipattern present and antipattern not present qualifiers preserve direct polarity.

Runtime confidence records evidence certainty and has exactly three valid values: weak, moderate, and strong. strong maps to the entry's Strong signal decision criterion and must also satisfy any strong_evaluable_from gate. moderate maps to the entry's Moderate signal decision criterion and satisfies the base evaluable_from gate. weak means direct, source-located but incomplete positive evidence that satisfies the base gate. It never means speculation, a failed source gate, not_evaluable, or not_applicable. medium is invalid and is never an alias for moderate.

Detection polarity remains binary regardless of confidence. Each detected pattern contributes +1 and each detected antipattern contributes −1 to pattern_score; weak, moderate, and strong detections have the same arithmetic weight. Absent, undetected, not_evaluable, and not_applicable outcomes contribute zero. Equivalently, pattern_score = count(patterns_detected) - count(antipatterns_detected).

Evidence-Source Contract

Some patterns depend on motion, artifact differences, or ordered visual states that a transcript or flattened slide cannot expose. Those entries declare three frontmatter fields:

• evaluable_from: one or more OR alternatives that can support evaluation. A string is a singleton alternative; a nested list is an all-of alternative whose sources must all have been inspected.
• evidence_requirements: positive facts the named source must establish before scoring.
• not_evaluable_when: precise source limitations that prohibit a present/absent guess.

Two outcome-specific gates use the same OR/all-of grammar. Every bundled observable entry declares both fields explicitly; omission/default behavior exists only for external or legacy catalogs:

• strong_evaluable_from: sources that can support a confidence: strong detection. Legacy loaders default it to evaluable_from when omitted; moderate and weak detections always use the base gate.
• absence_evaluable_from: sources that can establish a valid undetected outcome. Legacy loaders default it to evaluable_from when omitted. When this gate is not satisfied, an undetected entry is not_evaluable; a valid positive detection takes precedence and suppresses that contradictory absence requirement. A satisfied role gate permits an undetected outcome but does not prohibit an explicit not_evaluable when a catalog disqualifier or artifact capability failure still prevents evaluation.

An explicit absence_evaluable_from: null is different from omission: it permits source-gated positive detections but never authorizes an absence. Use it when a talk artifact can expose the behavior but cannot prove that an external prerequisite or unobserved contingency created a fair opportunity for the behavior to occur.

Logical prerequisites that complete permitted artifacts can establish use a required pair of applicability fields:

• not_applicable_when: one or more stable condition_id + description objects.
• applicability_evaluable_from: the same OR/all-of source grammar used by the evidence gates; its source coverage applies to every condition on the entry.

Every applicability condition must receive a source-located assessment. An explicit null absence gate permits source-gated positive detections but never authorizes an absence; entries with an applicability contract require a source-located applicable/not-applicable assessment, and only an applicable assessment plus complete absence-gate coverage yields undetected. An established condition yields not_applicable; incomplete applicability or absence coverage yields not_evaluable. This keeps inapplicable talks out of recurrence denominators instead of silently counting them as pattern absences.

Applicability declarations preserve the catalog's intended future evidence semantics, but a source-role label does not manufacture a missing modality. The current runtime leaves a native-deck, delivery-video, or comparison-based applicability alternative unsatisfied unless a versioned capability/alignment receipt proves the exact screen, audio, audience, stage, session-boundary, playback, or comparison facts the condition requires. Unsupported applicability therefore yields not_evaluable, never a forced applicable or not_applicable assessment.

The evidence-source enum is deliberately small:

An evidence-source gate and evidence_channels are intersecting constraints, not alternative vocabularies. Every exact source named by a detection must have a source-located citation through a channel that the entry permits and that can locate that source. Passing a source gate never expands the entry's allowed citation channels, and a metadata citation may corroborate a detection but never widens the source gate.

For any nested gate alternative, list every underlying source plus the source_comparison marker in the return's global inspected evidence_sources. For a positive comparison detection, additionally cite evidence_source: source_comparison and include evidence_sources_used with exactly the qualifying underlying group. For an undetected absence outcome, gate satisfaction is derived only from the global inspected-source list; there is no detection object or evidence_sources_used field. The comparison label is never a gate member or singleton alternative. Alternatives remain OR: for example, delivery_video can be a self-contained singleton alternative beside an all-of [static_slides, transcript] alternative.

For the current bundled catalog capability generation, absence gates are stricter than the general grammar: they are either null, singleton transcript, or singleton static_slides. native_deck, delivery_video, and nested/comparison alternatives may still support positive detections, but cannot establish absence until a versioned capability/alignment receipt is defined and audited.

For a positive detection, record the qualifying value as evidence_source alongside the concrete evidence. For an undetected outcome, apply the effective absence gate described above. Append the pattern ID, best available source, and reason to pattern_observations.not_evaluable when that absence gate is unsatisfied or a not_evaluable_when condition still prevents judgment. Do not add it to either detected array, do not infer absence, and exclude it from pattern_score. observable: false remains a separate catalog-wide rule: those entries are skipped and surfaced as go-live actions, not emitted as per-talk not_evaluable observations.

---

Pattern Catalog

Prepare Phase (20 patterns + 4 antipatterns)

Build Phase (41 patterns + 10 antipatterns)

Deliver Phase (22 patterns + 14 antipatterns)

---

Phase-Grouped Lookup Table

Which patterns to surface at each creator phase:

Phase 0: Intake

Audience/context patterns — shown as context-setting reminders.

• know-your-audience
• required
• the-big-why
• proposed
• emotional-state

Phase 1: Intent

Planning patterns — inform spec construction.

• narrative-arc
• fourthought
• triad
• required
• proposed
• mentor
• opening-punch
• walk-around (audit load-bearing claims for register spread)
• abstract-attorney (antipattern — warn about abstract drift)

Phase 2: Architecture

All structural patterns — the main Pattern Strategy selection moment.

With the mastery_and_novelty domain available:

• Signature tier: Current-cohort mastery_levels.signature patterns
• Contextual-history tier: Current-cohort regular/occasional patterns matching context
• New to You tier: Exactly current-cohort mastery_levels.never_tried / never_used_patterns, filtered by relevance; never not_yet_observed or a newest-talk first detection
• Shake It Up tier: Exactly 1–2 current-cohort never-tried provocations

When that domain is unavailable, suppress all four history-derived tiers and use a flat relevant list from this current taxonomy without usage or novelty claims.

Structural patterns relevant here:

• narrative-arc, triad, talklet, expansion-joints, lightning-talk, takahashi, cave-painting
• a-la-carte-content, context-keeper, bookends, intermezzi
• unifying-visual-theme, vacation-photos, infodeck
• brain-breaks, greek-chorus, retrieval-beat, walk-around
• live-demo (if demo-driven mode)
• sparkline (if persuasive mode — alternative to narrative-arc)
• call-to-adventure (architectural placement of sparkline turning point 1)

Phase 3: Content

Build patterns — applied during outline writing.

• foreshadowing, backtracking, brain-breaks, bookends, intermezzi, coda
• breadcrumbs, context-keeper, a-la-carte-content
• live-demo, lipsync, emergence
• leet-grammars, entertainment, breathing-room
• display-of-high-value, mentor, seeding-the-first-question
• make-it-rain, lightsaber, echo-chamber
• opening-punch, screen-blackout
• call-to-adventure, call-to-action, new-bliss, star-moment, inoculation, master-story
• concrete-before-abstract
• guess-first, retrieval-beat
• walk-around

Phase 4: Guardrails

Antipatterns as warnings — scanned against the outline.

All 28 antipatterns, with two flag types:

• [RECURRING]. Emit one line per Phase 4 recurring_antipatterns record.
• [CONTEXTUAL] — detected in the current outline; always enabled and not a claim that the issue is historically new

Phase 5: Slides

Visual/construction patterns — applied during slide generation.

• defy-defaults, composite-animation, vacation-photos, analog-noise
• takahashi, cave-painting, charred-trail, gradual-consistency
• exuberant-title-top, invisibility, soft-transitions
• traveling-highlights, crawling-code, emergence
• preroll, crawling-credits, bookends, intermezzi, breadcrumbs
• screen-blackout
• guess-first, retrieval-beat (question/answer slide construction — the deck must withhold)
• second-look (two-layer legibility — requires a destination; without one it is ant-fonts)
• cookie-cutter, ant-fonts, fontaholic, floodmarks, photomaniac, laser-weapons (antipatterns)

Phase 6: Publishing

Delivery prep patterns — final checklist before the talk.

• preparation, carnegie-hall, seeding-satisfaction, seeding-the-first-question
• posse, shoeless, emotional-state, weatherman, stakeout
• echo-chamber, red-yellow-green
• social-media-advertising, preroll, live-on-tape
• call-to-action (go-live preparation — confirm asks are concrete and immediately executable)
• spaced-followup (post-event — the only checklist item that fires after the talk)
• second-look (confirm the shownotes destination exists and the deck is actually published)

---

Vault Dimension → Pattern Mapping

Reverse lookup: which patterns relate to each of the 14 rhetoric analysis dimensions.

---

Denominator-Safe Catalog Status

All 81 observable entries have explicit positive, strong, and absence outcome gates. Of those, 37 also declare source-located applicability conditions. Only 16 currently permit absence; 65 are deliberately positive-only because their available role labels do not intrinsically prove the timing, rendering, screen, audio, audience, stage, session, or playback capability needed for an exhaustive negative judgment.

Every other observable entry declares absence_evaluable_from: null. For the static lane, this authority depends on a complete, separately declared readable PDF; a video-extracted, sampled, deduplicated, or cropped PDF can still support a positive detection but never joins an absence denominator.

---

Unobservable Patterns — Go-Live Checklist

These patterns involve pre-event logistics, hidden authoring processes, physical stage behaviors, post-event follow-up, or external systems that the current ingress artifacts cannot directly prove. The vault cannot score them. Instead, they surface during creator Phase 6 (Publishing / Go-Live) as a preparation checklist.

Do not include these in vault scoring or the speaker profile's pattern_profile.

Pre-Event Preparation

Artifact and Publishing Evidence

During Delivery

Post-Event

Antipatterns to Avoid (unobservable)

---

Summary Statistics

• Total entries: 111 (83 patterns + 28 antipatterns)
• Observable (vault-scorable): 81 (62 patterns + 19 antipatterns)
• Positive source-gated: 81
• Absence source-gated: 16
• Applicability-gated: 37
• Positive-only (absence disabled): 65
• Unobservable (go-live checklist): 30 (21 patterns + 9 antipatterns)
• Prepare phase: 24 (20 patterns + 4 antipatterns)
• Build phase: 51 (41 patterns + 10 antipatterns)
• Deliver phase: 36 (22 patterns + 14 antipatterns)

Sources

• Ford, N., McCullough, M., & Schutta, N. (2013). Presentation Patterns: Techniques for Crafting Better Presentations. Addison-Wesley. — primary taxonomy source.
• Reynolds, G. (2012). Presentation Zen: Simple Ideas on Presentation Design and Delivery (2nd ed.). New Riders. — supplementary source; reinforces ~17 existing patterns and contributes the opening-punch and screen-blackout patterns plus three refinement subsections folded into existing patterns: "Hara Hachi Bu — The 90–95% Finish Line" (in breathing-room.md), "Plan Analog Before Going Digital" (in concurrent-creation.md), and "The Elevator Test" (in the-big-why.md).
• Vault-derived patterns — observed across the speaker's own corpus rather than sourced from a book: delayed-self-introduction, three-part-close, progressive-reveal, anti-sell, meme-as-argument, second-look. Mechanism support is cited per-file where a literature basis exists (e.g. second-look cites Loewenstein's information-gap account of curiosity); the evidence is field observation.
• Brown, P. C., Roediger, H. L., III, & McDaniel, M. A. (2014). Make It Stick: The Science of Successful Learning. Belknap Press / Harvard University Press. — supplementary source; the catalog's only retention-science source. Contributes three new patterns (guess-first, retrieval-beat, spaced-followup) and one new antipattern (nodding-room), plus five refinement subsections folded into existing patterns: "Rehearse by Retrieval, Not Rereading" (in carnegie-hall.md), "The Consolidation Pause" (in brain-breaks.md), "Learning Styles Are a Myth" (in know-your-audience.md), "Smile Sheets Do Not Measure Retention" (in red-yellow-green.md), and "Do Not Make It Hard to Read" (in analog-noise.md, which corrects a retracted disfluency claim the file previously carried).
• Herrmann, N., & Herrmann-Nehdi, A. (2015). The Whole Brain Business Book (2nd ed.). McGraw-Hill. — supplementary source; contributes walk-around and the golden-rule antipattern from Ch. 8 (Communicating Across Thinking Styles) and Ch. 13 (Influencing and Getting Buy-In). The catalog imports the four-question Walk-Around procedure and the A/B/C/D quadrant vocabulary as a recognizable handle; it does not import the Whole Brain Model's physiological claims, the HBDI instrument, audience typing, or the book's gender-differences section. See walk-around.md Related Reading for the boundary and for the replicable premise (Elaboration Likelihood / need for cognition) the pattern actually rests on.
• Duarte, N. (2010). Resonate: Present Visual Stories that Transform Audiences. Wiley. — supplementary source; reinforces ~20 existing patterns and contributes seven new build-phase patterns (sparkline, call-to-adventure, call-to-action, new-bliss, star-moment, inoculation, master-story) plus six refinement subsections folded into existing patterns: "Adopting the Stance — Planning Implications" (in mentor.md), "The Big Idea — Statement Format" (in the-big-why.md), "Numerical Narrative — Making Numbers Land" (in vacation-photos.md), "Screening with Critics — Beyond Copyediting" (in peer-review.md), "Murder Your Darlings — The Pre-Delivery Cut Pass" (in crucible.md), and "The Three Contrast Types — Engine of the Middle" (in sparkline.md).