# Catalyst Desk Playbook
### Small-account catalyst trading without buying the circus

**Version:** 1.0 · **Updated:** 2026-08-04  
**Price intent:** paid digital product  
**Not financial advice.** Process framework only. You can lose 100% of risked capital.

---

## Who this is for
- Small accounts ($100–$5k) trading **US micro / small / mid-cap** catalysts
- People tired of “smart money” screenshots and day-of headline chasing
- Options **or** stock — the process is the same; structure changes

## Who this is not for
- Large/mega-cap index gambling dressed up as alpha
- Guaranteed-win courses
- Chart-pattern religion with no primary fact object

---

## 1. Core idea

Edge is mostly **when + structure**, not ticker discovery.

A good candidate has:
1. A **fact object** (filing, IR PR, calendar date, clinical registry)
2. A labeled **entry window**
3. Rough **EV** (not vibes-only confidence)
4. A written **exit plan before entry**

Empty watchlist days are valid. Forced trades are how small accounts die.

---

## 2. Universe rules (copy these)

| Allow | Reject |
|-------|--------|
| Micro / small / lower-mid / mid | Large & mega (roughly >$10B) |
| US-listed, tradable | OTC junk / no primary docs |
| Liquid enough for your size | Toxic spread you can’t exit |
| Dated path in next ~1–15 sessions (sprint) | “Someday” story stocks |

Raise liquidity standards as market cap rises.

---

## 3. Two windows + one reject zone

### Window A — Pre-event
Intentional risk **before** a known print/date.
- Need date certainty + thesis scenarios + max loss
- Options: pay attention to IV crush

### Window B — Post-event pullback-continuation
After the first impulse finishes:
- Thesis still alive
- Pullback holds
- Volume/structure returns
- **Not** the first green spike

### Reject zone (Recipe D)
Hard pass when most of these stack:
- Headline already out
- Stock extended / vertical
- Short-dated OTM lottery mindset
- Social velocity lagging price

**Name it out loud:** late-chase. Reduce size or pass.

---

## 4. Fact object requirement

No primary object → **research only**. Never capital.

Valid objects:
- Company IR press release
- SEC filing (8-K, 13D/G, Form 4 cluster, S-3)
- Official calendar date (PDUFA, earnings, adcom, conference)
- ClinicalTrials.gov / registry timing update

Weak objects:
- Random Twitter thread
- Aggregator reprint with no link to primary
- “I heard”

---

## 5. Confluence recipes

Score by **count of independent concrete objects**, not story quality.

### Recipe A — Calendar coil (most useful)
1. Known date within 1–10 sessions  
2. Price not already vertical into the date  
3. Optional: multi-day call OI build  
4. Optional: SI/borrow context  
→ Choose Window A **or** planned stand-aside into Window B

### Recipe B — Ownership spark
1. Fresh 13D/13G or Form 4 open-market buy cluster  
2. Enough liquidity  
3. No fresh ATM poison  

### Recipe C — Binary biotech package
1. Dated readout on registry/IR  
2. Financing not desperate mid-hype  
3. Scenario tree (approve / delay / CRL) written before entry  

### Recipe D — Late chase (**reject**)
Headline out + extended + OTM short-dated + lagging social

### Recipe E — Micro trap (**reject / research only**)
Paid promo + ATM/S-3 + thin disclosures + social-only narrative

---

## 6. Entry window labels (required)

Every idea gets exactly one:
- `early`
- `pullback-continuation`
- `pre-event`
- `late-chase` (block or tiny only)
- `none`

Only early / pullback-continuation / clean pre-event get normal size.

---

## 7. EV beats confidence theater

Confidence without expected-value math is soft.

Minimum fields before size:

```text
debit_or_risk_usd
p_stop          # chance thesis fails / stop
p_2x            # chance of ~2R path
p_runner        # chance of ≥3R path
p_scratch       # small loss / flat
ev_R            # rough expected R
ev_edge         # positive | flat | negative
```

Sprint-grade formula (honest ranges > fake precision):

```text
ev_R ≈ (p_2x * 1.0) + (p_runner * 2.0) + (p_scratch * -0.15) + (p_stop * -0.60)
```

Gates:
- `ev_edge = negative` → no full size
- prefer `ev_R ≥ 0.15` for normal size
- `0–0.15` → tiny
- can’t estimate even a range → research only

### Sanity killers (haircut EV hard)
- Deep OTM + post-headline + rich IV
- Move-to-breakeven ≳ 35–45% with no secondary path
- Binary hold-through with no crush plan
- Illiquid chain (spread tax)

---

## 8. Signal stack (score 0–2 each)

| Signal | What “2” looks like |
|--------|---------------------|
| Fact quality | Primary + dated + specific |
| Calendar path | Hard date inside your hold window |
| Entry window | Early / clean pullback-cont |
| Structure / m2BE | Realistic move to make money |
| Multi-day OI build | Clear build **into** event (not same-day spike) |
| Price structure | Coil/base, not vertical |
| Social velocity | Early acceleration, not lagging hype |
| Dilution risk | Clean / low near-term dilute |
| Liquidity | You can enter and exit the plan |
| Heat fit | Fits account budget |

High confidence + weak signal sum = cut confidence or block.

---

## 9. Sizing (small account)

Default heat caps (adjust to pain tolerance):
- One idea: ≤35–40% buying power (options premium at risk)
- **Stub accounts (<~$250):** prefer **shares**, deploy more of the book, skip lottery calls
- Never average down short-dated OTM losers
- One symbol gets one thesis budget — not a new budget per fill

| ev_R | conf | size |
|------|------|------|
| < 0 | any | no enter |
| 0–0.15 | ok | tiny |
| ≥ 0.15 | ≥70 | normal |
| ≥ 0.15 | ≥85 + clean | upper cap only |

---

## 10. Exit plan before entry (non-negotiable)

Write these **before** you click buy:
- Hard invalidation (price, filing, or thesis break)
- Profit scales (example: trim at ~2R / ~3R, runner rules)
- Time stop (especially short-dated options)
- Binary rule: hold through print? (default no unless precommitted)
- “No average down” unless **new** fact object

### Thin tape rule
Don’t panic-exit on one empty bid or a single spike mark.
Stops **arm → confirm** (second check / a few minutes) unless thesis is clearly dead.

---

## 11. Options-specific geometry

For sub-$5 names and short-dated calls:
- State % move-to-breakeven explicitly
- Deny deep OTM lotto unless conf is exceptional **and** catalyst sits inside expiry
- Same-week calls need fast confirmation; don’t donate theta out of hope
- Multi-day OI build = confirm only; same-day post-spike flow = noise

---

## 12. Daily operating loop (30–45 min)

1. **Calendar** next 1–10 sessions (PDUFA, earnings, adcoms, conferences)
2. **Hunt** fresh IR/8-K in universe (not only pet tickers)
3. **Label** each name: permission = `research_only` | `watch` | `ready` | `blocked`
4. **Score** only the top 1–3 with fact + window + rough EV
5. **Trade or pass** — pass is success
6. **Journal** one decision object even on no-trade days

### Decision object template

```text
decision: ENTER | TRIM | EXIT | HOLD | NO_TRADE
symbol:
reason:
fact_object:
entry_window:
confidence:
ev_R / ev_edge:
signal_note:
size:
limit:
exit_plan:
gates_passed: yes|no
block_reason:
```

---

## 13. Common failure modes (steal these scars)

1. Triple-buying short-dated OTM losers in one symbol  
2. Treating conf scores as EV  
3. Buying first spike after AdCom/FDA headline  
4. Holding premium to zero because “binary soon”  
5. Micro promo + ATM as “asymmetric upside”  
6. Research theater: beautiful notes, zero deploy when a clean equity setup exists  
7. Account too small for options sprint — still forcing options  

---

## 14. Stub-capital mode (<~$250)

When the account can’t express multi-contract options:
- Prefer **long equity** on one catalyst name
- Deploy most of BP only when fact + window + invalidation exist
- Options only if debit is a small slice and geometry is clean
- Goal: stop being 100% cash out of fake perfectionism, without Recipe D suicide

---

## 15. Ethics / realism

- This is a **process**. It does not remove variance.
- Full loss on a single idea must be acceptable at your stated size.
- Past pattern language ≠ future profits.
- If you need the money for rent, do not use this framework with money you can’t lose.

---

## Quick wall checklist (print this)

- [ ] Fact object + URL  
- [ ] Universe OK (not large/mega)  
- [ ] Entry window labeled (not late-chase)  
- [ ] EV rough math done  
- [ ] Signal stack not empty  
- [ ] Size fits heat  
- [ ] Exit plan written  
- [ ] Not averaging a loser  
- [ ] Not Recipe D / E  

If any hard box fails → **no trade**.