# spec.md — Speculative Categories, Frameworks and Evaluation

Deep dive reference. Read `plan.md` first if you are new.

Last reviewed: 30 July 2026
**Educational guidance, not personalised financial advice.**

---

## Part 1 — Category deep dives

### 1.1 ASX small caps and microcaps

The ASX is unusually rich in speculation because a large share of listings are pre-revenue resource
explorers. This creates a genuinely researchable market: continuous-disclosure obligations mean the
raw material exists, and most participants ignore it.

**What actually drives them**

- **Drill results.** The dominant catalyst class. Note that grade, width and depth all matter, and
  headline grams-per-tonne figures are routinely quoted from the best intercept in the hole.
- **Resource upgrades.** Movement from Inferred → Indicated → Measured under the JORC code is a real
  de-risking event with a real date.
- **Offtake agreements and government funding.** Converts a geology story into a commercial one.
- **Commodity price moves.** The explorer is a leveraged bet on the underlying commodity, which is
  itself often a bet on policy.
- **Capital raisings.** The most common cause of a sudden drop with "no news." Placements to
  institutions at a discount, with the shortfall hitting the market.

**Structural things to check every time**

- Cash position and quarterly burn rate (Appendix 5B on the ASX). Two quarters of cash means a raise
  is coming, and it will be at a discount.
- Shares on issue over five years. Tripled share count is a tripled hurdle.
- Options and performance rights overhang.
- Directors' history. The same names appear across serial shells.
- Average daily traded value. This is your real position-size constraint, not market cap.

**Current context (July 2026).** Small caps have been outperforming large caps, with the Small
Ordinaries index gaining ground and risk appetite extending into the speculative end. Resources,
mining and energy are the standout performers, with critical minerals — rare earths, antimony,
copper, uranium, lithium — attracting most of the flow. Capital is increasingly discriminating
between explorers who can show funding, technical progress and a credible path to cash flow, and
those who cannot. That discrimination is itself the signal: the theme being real does not make the
individual explorer real.

### 1.2 US small caps and microcaps

Deeper disclosure (EDGAR), deeper liquidity, and a far larger and faster retail crowd.

**What drives them**

- Biotech readouts — binary, dated, and brutal. A phase readout is a coin flip with a schedule.
- Earnings surprises and guidance changes in thinly covered names.
- Index inclusion and lock-up expiries — mechanical flows with known dates.
- Options positioning. Retail traded a record ~$6.8bn of options premium per day in June 2026, more
  than double the historical average, with roughly $1.9bn/day in semiconductors alone. That level
  of derivative activity means the tail can wag the dog, especially in smaller names.
- Short interest and float mechanics.

**Current context.** Retail is a persistent, dip-buying source of demand — buying roughly 3.5x
average volume on down days through 1H 2026, the strongest such behaviour on record. But the crowd
is also more concentrated and faster-rotating than in previous cycles: it has already cycled through
energy, silver, software, semiconductors, crypto ETFs and space names this year. Fast rotation means
narratives saturate sooner than they used to. Assume less runway than the historical analogues suggest.

### 1.3 Crypto narratives

The purest expression of narrative-driven speculation: no cash flows, no disclosure, 24/7 markets,
and reflexivity everywhere.

**Rotation structure.** Capital cycles roughly major asset → large alternatives → sector narratives
→ the memecoin tail, then back. Each full rotation historically runs 6–14 weeks. Late-cycle is
identifiable by the tail outperforming — when things with no pretence of utility lead, the rotation
is near its end.

**Evaluating a crypto narrative**

- Is there a real technical or regulatory change underneath, or only a name?
- Token supply schedule — unlock cliffs are dated, public, and reliably brutal.
- Concentration of holdings. A handful of wallets holding the float is not a market.
- Liquidity depth on actual venues, not headline volume figures.
- Whether the narrative has a *builder* base or only a *trader* base.

**Tier assignment:** major assets Tier 2, sector narratives Tier 2–3, memecoins Tier 4 without
exception.

### 1.4 AI and adjacent hype cycles

AI has stopped being one narrative and become a stack of them, each with its own cycle: compute,
power and cooling, networking, inference, agents, applications, and now quantum as the adjacent
"next" story.

**Quantum computing is the clearest current example of a mid-to-late-stage hype cycle.** Pure-plays
like IonQ, Rigetti and D-Wave have gained 50%+ since late March 2026, with the benchmark quantum
index up ~69% by end of May against the S&P 500's ~11%. A US Department of Commerce announcement of
$2bn in CHIPS Act funding in May was the policy catalyst. These companies have minimal revenue and
no profits; trailing gains in the hundreds to thousands of percent are a speculative premium on a
story about the 2030s, not a rerating of validated business models. Swings of 30–50% are routine.

**The general test for a thematic hype cycle:** is the *revenue* in the theme, or only the
*narrative*? Picks-and-shovels businesses in a hype cycle have real customers. Pure-plays often have
government grants and pilot programs. Both can go up; only one has a floor.

### 1.5 Collectibles

Slow, physical, and structurally different from everything above.

**Market context.** The global collectibles market is estimated in the $320bn–$600bn range depending
on definition, growing at a mid-single-digit CAGR. Interest is concentrated in established
categories — Pokémon, graded sports cards, LEGO sets, luxury watches. A $16.5m Pokémon card sale in
early 2026 crystallised the "collectibles as asset class" narrative in mainstream media, which by
the framework in `plan.md` is a **Stage 4 saturation signal**, not a green light.

**The frictions that eat returns**

- Round-trip costs of 10–20% (auction fees, buyer's premium, grading, shipping, insurance)
- Grading is a gate and a lottery — the same card at two grades can differ 10x in value
- Authentication risk, storage risk, physical damage risk
- Illiquidity measured in months
- Supply risk: overproduction in modern cards has diluted scarcity in lower tiers

**Regulatory watch.** Japan has moved to regulate its rapidly expanding Pokémon card market amid
concerns about speculation, transparency and consumer protection. Regulation arriving is a
late-cycle marker across every speculative category.

### 1.6 Prediction markets

Kalshi, Polymarket and the regulated venues. Combined monthly volume has gone from under $5bn in
September 2025 to roughly $21–24bn by April 2026, one of the fastest growth curves observed in
digital financial markets. Kalshi has overtaken Polymarket on volume and took roughly 83% of
notional volume across CFTC-approved exchanges through the 2026 World Cup final. Sports, politics
and crypto account for ~90% of volume on both platforms.

**Why they matter for a beginner.** Positions expire and resolve objectively, usually within weeks.
You cannot tell yourself a losing position is "long-term." That makes them the best available
calibration training: predict, size, resolve, score, repeat.

**Why they are not a compounding vehicle.** Capped upside per contract, fees, and — for the most
liquid markets — genuinely efficient pricing. Treat as a training ground and a sentiment data
source, not a return engine.

### 1.7 Emerging and frontier markets

Country and sector bets where currency risk, capital-control risk, custody risk and governance risk
stack on top of ordinary market risk. Narratives here are typically macro or policy driven —
elections, IMF programs, commodity terms of trade, sanctions relief. Long fuses, slow resolution,
and access is often the binding constraint for retail.

---

## Part 2 — Catalysts

A catalyst is a specific event that forces the market to re-price. "Sector momentum" is not a
catalyst; it is an excuse.

### 2.1 Catalyst taxonomy

| Type | Examples | Dated? | Typical reaction |
|---|---|---|---|
| **Scheduled binary** | Drill results, trial readouts, earnings, court rulings | Yes | Violent both ways |
| **Scheduled mechanical** | Index rebalances, lock-up expiries, token unlocks, options expiry | Yes | Predictable direction, size uncertain |
| **Policy** | Export controls, subsidies, defence budgets, regulatory approvals | Sometimes | Creates whole narratives |
| **Administrative rulemaking** | Reimbursement schedules, tariff schedules, licensing regimes, standards updates | Yes — statutory | Narrow, deep, and usually ignored |
| **Corporate action** | Takeovers, offtakes, spin-offs, capital raises | Rarely | Immediate and permanent |
| **Emergent** | Viral attention, influencer coverage, index-fund launch | No | Fast, shallow, often reverses |

**Added 29 July 2026.** Administrative rulemaking was split out of the policy row because it behaves
differently in the one way that matters most: the calendar is statutory rather than discretionary. A government
*deciding* something is undated; a government *publishing a rule* runs proposed rule → comment period → final
rule → effective date, every year, on schedule. The CMS Physician Fee Schedule is the clearest example — proposed
in July, comments close in September, final rule in the autumn, rates effective 1 January.

Two cautions that come with the category. A proposed rule is a **deadline for input, not a resolution**; values
routinely change between proposal and final rule, because the comment period exists precisely so that lobbying
works. And the resolution itself is often binary in an unhelpful direction: a rate left *unchanged* can move a
price as much as a rate cut, if the market was positioned for change.

**Amended 30 July 2026 — the sub-class the 29 July note got wrong.** Administrative rulemaking has two halves and
only one of them is dated:

| Sub-class | Timing | Route to price | Observed impact |
|---|---|---|---|
| **Scheduled rulemaking** | Statutory: proposed rule → comments → final rule → effective date | Anticipated, partly priced, resolves on publication | Moderate; sell-the-news risk high |
| **Unscheduled mid-cycle change** | None. An agency revises billing requirements, guidance, classifications or fee treatment between rule cycles | Arrives as a surprise to a market that had no date to prepare for | Larger, and asymmetric |

The evidence for the split arrived within a day of the category being written. Non-invasive bone growth stimulators
were reclassified by the FDA from Class III to Class II in April 2026; a Medicare billing change followed in May;
that change was **withdrawn by revised guidance on 1 July**, restoring prior reimbursement. The affected
single-code company repriced by roughly 10% and then about 16% — a far larger move than anything the annual fee
schedule produced. None of it was on a published calendar.

The implication is uncomfortable for anyone who likes dated catalysts: **within administrative rulemaking, the
larger moves live in the undated half.** A scheduled rule is researchable in advance and partly priced; an
unscheduled change is unpredictable and mostly unpriced. Those are opposite kinds of opportunity and they should
not be scored as one category. Practically, the researchable artefact is not "read the rule before the date" but
"maintain a standing map of which listed companies depend on which codes, so that when an undated change lands you
already know who it hits." The map is the durable work; the calendar is only one of the events that uses it.

A second, upstream lesson: the chain started with a **device-classification** decision, not a payment decision.
Regulatory catalysts propagate across agencies, and the agency that acts first is often not the one that sets the
price.

### 2.2 Grading a catalyst

Score each 1–5 and multiply out:

- **Specificity** — is there an actual event, or a vague expectation?
- **Date certainty** — a known date, a known quarter, or "soon"?
- **Materiality** — does it change the asset's value, or just the conversation?
- **Verifiability** — will an independent source confirm the outcome?
- **Asymmetry** — is the upside on a good outcome larger than the downside on a bad one?

Anything scoring low on date certainty deserves a smaller position and a hard time stop. "Soon" has
no expiry, and positions held against "soon" are where speculative capital goes to die quietly.

### 2.3 Sell-the-news

The single most common way a correct thesis loses money. If a catalyst is widely anticipated, its
outcome is partly priced before it happens. Good news then produces a flat or negative reaction
because the anticipating buyers become sellers on resolution.

**Practical implication:** decide before the catalyst whether you are trading the *anticipation* or
the *outcome*. They are different positions with different exits, and conflating them means you get
the worst of both.

---

## Part 3 — Risk tiers in detail

### Tier 1 — High risk, grounded narrative

Real revenue, real assets or a producing operation. Verifiable disclosure. Adequate liquidity.
Examples of shape: profitable small caps in a hot theme, established producers, mid-cap
takeover targets.
*Realistic downside:* −50% and slow recovery. *Position implication:* the only tier where a
merely-large position is defensible, and even then within a capped speculative sleeve.

### Tier 2 — Very high risk, hype-driven

Listed and liquid, but valued on a story about the future. Pre-revenue or thin revenue. Valuation
is untethered from current fundamentals by design.
Examples of shape: quantum pure-plays, space companies, advanced-stage explorers, major crypto
sector narratives.
*Realistic downside:* −70% to −90% when the narrative rolls over. *Position implication:* small,
with a written −50% plan.

### Tier 3 — Extreme risk, low liquidity

Microcaps. Thin order books, wide spreads, serial dilution, minimal independent coverage. The
screen price is an opinion, not an executable quote.
Examples of shape: early-stage explorers pre-resource, nano-cap biotech, thinly traded shells.
*Realistic downside:* −95% and unable to exit at the price shown. *Position implication:* assume
total loss; size so that total loss is uninteresting.

### Tier 4 — Meme-level speculation

The asset is attention. No cash flow, no asset, no disclosure obligation, often no identifiable
team. Value is entirely the belief that someone will pay more later.
*Realistic downside:* −100%, quickly, and it is the base case rather than the tail. *Position
implication:* money you have already mentally spent.

### Tier migration

Tiers change with **structure**, not price. A Tier 3 explorer that signs a funded offtake and lists
a resource can migrate to Tier 2. A Tier 2 company that raises twice at successive discounts and
loses its liquidity migrates to Tier 3. Re-tier monthly; a silent tier migration downward is one of
the most common ways a position becomes unexitable without the holder noticing.

---

## Part 4 — Sentiment analysis

### 4.1 The five-point scale

| Reading | What it looks like | What it usually means |
|---|---|---|
| **Fearful** | Capitulation posts, "never again", volume dead | Late-stage decay; occasionally the base |
| **Cautious** | Technical questions, scepticism, small positions | Emergent stage — best risk/reward |
| **Neutral** | Low volume, factual discussion, little emotion | Latent or forgotten |
| **Excited** | Price targets, new accounts arriving, ramping up | Acceleration — trend intact, risk rising |
| **Euphoric** | "Can't lose", leverage talk, mainstream coverage | Saturation — the marginal buyer is uninformed |

### 4.2 Measure tone and volume separately

Discussion **volume** tells you how crowded a trade is. Discussion **tone** tells you which
direction the crowd leans. The interesting information is in the divergence:

- **Volume up, price up, tone constructive** → trend intact
- **Volume up, price flat/down, tone defensive** → distribution; someone is selling into the crowd
- **Volume down, price up, tone quiet** → accumulation, or simply illiquidity
- **Volume down, price down, tone hostile** → decay stage; the story is over even if holders aren't

### 4.3 Reading forums honestly

A ticker's own forum is structurally a room of holders. It is a valid instrument for measuring
sentiment and a worthless one for establishing facts. Specific things to weight:

- Are the *questions* getting more technical or less? Less technical = later stage.
- Are dissenting posts engaged with or shouted down? Shouting = saturation.
- Are there new accounts with strong price targets and no falsifiable claims? That is promotion.
- What is the ratio of "why this will go up" posts to "what would make this fail" posts?

### 4.4 Sentiment heatmap

The page renders a heatmap of current readings by theme. Its purpose is not prediction but
**crowding awareness** — if everything you are watching sits in Excited/Euphoric, you have no
diversification even if you hold twelve different names, because they will all sell off together.

---

## Part 5 — The evaluation framework

### 5.1 Seven-dimension score (0–100)

| Dimension | Weight | 0 points | Full points |
|---|---:|---|---|
| Narrative strength | 20 | Company-manufactured, no external driver | Externally driven, coherent, multi-source |
| Catalyst strength | 20 | Vague, undated, immaterial | Specific, dated, material, verifiable |
| Sentiment | 15 | Euphoric or dead | Constructive and building |
| Risk tier | 15 | Tier 4 | Tier 1 |
| Liquidity | 10 | Cannot exit at screen price | Deep book, tight spread |
| Hype-cycle position | 10 | Saturation or Decay | Emergent |
| Information availability | 10 | No filings, no coverage | Full disclosure + independent coverage |

**Reading the bands**

- **75–100** — genuinely interesting research subject; still speculative, still assume the tier's downside
- **55–74** — worth a written thesis and a watchlist slot
- **35–54** — watch only; the narrative is real but the entry conditions are not
- **0–34** — record it in history for pattern-learning; not a research subject today

**What the score is not.** It is not expected return, not a probability, and not a recommendation.
A 90-scoring Tier 4 asset is still an asset you should assume goes to zero. The score's real
utility is *comparative and longitudinal*: the change in a position's score from entry is the
earliest, cleanest exit signal available.

### 5.2 The ten-minute triage

Before any deep work, answer these. Any "no" ends the analysis.

1. Can I name the narrative in one sentence without using the word "potential"?
2. Is there a specific, dated catalyst in the next 90 days?
3. Can I state what would prove me wrong, observably?
4. Is there enough liquidity that my intended position is a small share of daily volume?
5. Do primary sources exist — filings, announcements, regulator publications?
6. Is the sentiment somewhere below euphoric?
7. Do I know who is on the other side of this trade and why?
8. Have I written the −50% plan?

### 5.3 Pre-mortem

Before entering, write the story of how this position lost 90% of its value. Not a list of risks —
a narrative, in past tense, with a sequence of events. The specific failure paths you can imagine
in advance are the ones you will recognise early enough to act on.

---

## Part 6 — The six-week learning pathway

### Week 1 — Understanding speculation
*Concepts:* speculation vs investing; why base rates matter; sizing arithmetic; total-loss framing.
*Exercise:* write your speculative capital number and per-position cap. Take zero positions.
*Output:* a one-page written risk policy.

### Week 2 — Understanding narratives
*Concepts:* the five lifecycle stages; how narratives are manufactured vs how they emerge; the
saturation test.
*Exercise:* pick three live narratives. Track them daily for a week: stage, evidence, who is
talking, whether good news is still moving prices.
*Output:* a narrative register with dated entries.

### Week 3 — Understanding catalysts
*Concepts:* the catalyst taxonomy; specificity and date certainty; sell-the-news.
*Exercise:* build a 90-day catalyst calendar for ten watchlist assets. Predict the direction of
five of them in writing, then check.
*Output:* a dated catalyst calendar and five scored predictions.

### Week 4 — Building a research routine
*Concepts:* source tiering; primary vs secondary vs sentiment; watchlist hygiene.
*Exercise:* run the full daily/weekly cadence from `plan.md` §6 for one week without missing a day.
Cull your watchlist to twelve names.
*Output:* seven daily logs and one weekly review.

### Week 5 — Tracking sentiment
*Concepts:* tone vs volume; divergence patterns; crowding; reading forums as instruments.
*Exercise:* score sentiment daily on ten assets. At week's end, compare your scores to price action
and find every divergence.
*Output:* a sentiment heatmap and a written divergence analysis.

### Week 6 — Evaluating opportunities
*Concepts:* the seven-dimension score; the ten-minute triage; the pre-mortem.
*Exercise:* fully score five live ideas. Write complete journal entries for the top two, including
pre-mortems and −50% plans. Paper-trade both.
*Output:* five scorecards and two complete position journals.

**Graduation test:** given any speculative idea, state within ten minutes its narrative stage, next
dated catalyst, liquidity constraint, risk tier and score — then explain precisely what would make
you wrong.

---

## Part 7 — Sandbox scenarios (SIMULATED)

These are teaching simulations. They are fictional and are never mixed into the live watchlist.

**Scenario A — Narrative birth to death (16 weeks).**
Week 0: a government announces export restrictions on a critical input (*Latent* — three specialist
newsletters connect it to four tickers). Week 3: forum mentions rise 400%, price +60%, discussion is
technical (*Emergent* — best risk/reward, hardest to act on). Week 6: mainstream finance media
covers the theme; two companies announce name changes and pivots into it; price +240% (*Acceleration*).
Week 10: an ETF launches; a company posts an excellent result and the stock closes **down 4%**
(*Saturation* — the tell). Week 16: volume down 70%, price −65% from peak, forum hostile (*Decay*).
*Lesson:* the saturation signal was the flat reaction to good news, not the price peak.

**Scenario B — Catalyst emergence and sell-the-news.**
A microcap has drill results due "in Q3." Price drifts up 80% into the window on anticipation.
Results land: genuinely good grades, genuinely narrow widths. Stock closes −22%.
*Lesson:* the anticipation trade and the outcome trade are different positions. The result was good
and the position still lost, because the anticipating buyers were the sellers.

**Scenario C — Silent tier migration.**
A Tier 2 pre-revenue company raises at a 20% discount, then again four months later at a 30%
discount. Share count +65%. An institutional holder exits. Average daily traded value falls 80%.
No single day looks dramatic. The asset is now Tier 3 and the position cannot be exited at the
screen price.
*Lesson:* re-tier monthly on structure. The dangerous change was invisible on the price chart.

**Scenario D — Sentiment/price divergence.**
Mentions rise for three straight weeks while price grinds sideways and tone shifts from technical
questions to defensive reassurance. Two weeks later, price −40%.
*Lesson:* rising volume with flat price and defensive tone is distribution.

**Scenario E — Regulation as a late-cycle marker.**
A collectibles category runs for eighteen months. Mainstream press declares it an asset class after
a record sale. A regulator announces consultation on consumer protection and market transparency.
Prices peak within a quarter.
*Lesson:* regulation arrives after the crowd does. It is a stage marker, not a validation.

---

## Related files

- `plan.md` — the beginner's plan and behavioural framework
- `skills.md` — system architecture, agents, workflow and limitations
- `history.md` — append-only log of flagged opportunities
- `index.html` — the live, daily-updated page

---

*Educational guidance, not personalised financial advice. Speculative assets can lose all of their
value. Verify everything independently.*