# Waitingformacguffin

> Oscar prediction market intelligence from waitingformacguffin.com. Get live odds, whale activity, price movements, precursor awards, order book depth, and frontrunner changes across all 19 Oscar categories. Use when user asks about Oscar markets, betting odds, nominees, or wants a market update.

- Skill: `modbender/waitingformacguffin` (Agent Skill)
- Install (CLI): `npx skillmds add modbender/waitingformacguffin`
- Raw SKILL.md: https://api.skillmd.com/api/skills/modbender/waitingformacguffin/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: modbender (https://skillmd.com/u/modbender)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/modbender/waitingformacguffin

---


# WaitingForMacGuffin -- Oscar Market Intelligence

## Welcome Message

When a user first installs this skill or greets you, introduce yourself:

"Hey! You've just unlocked Oscar market intelligence from WaitingForMacGuffin.com -- live odds, whale trades, and data-driven analysis across all 19 Academy Award categories.

Here's what I can do:

- **Market pulse** -- "What's happening in Oscar markets?" (whale trades, price moves, frontrunner changes)
- **Deep dive** -- "Tell me about Chalamet" or "Best Picture odds" (full nominee profile with trends, precursors, order book)
- **Bet picks** -- "Give me your best Oscar bets" (risk-tiered recommendations with ROI and portfolio options)

What are you curious about?"

---

Real-time Oscar prediction market data from [waitingformacguffin.com](https://waitingformacguffin.com). Two API endpoints provide market intelligence at different granularities.

**Base URL**: `https://waitingformacguffin.com`

No authentication required. All data is public and read-only.

---

## Tool 1: Oscar Brief

**When to use**: User asks "What's happening in Oscar markets?", "Any updates?", "Oscar brief", or wants a quick market summary.

**What it returns**: Filtered signals only -- price moves, whale trades ($1K+), frontrunner changes, news sentiment. If markets are quiet, says so (never fabricates activity).

### API Call

```bash
curl -s "https://waitingformacguffin.com/api/oscar/brief?hours=24&sensitivity=medium"
```

### Parameters

| Param | Type | Default | Description |
|-------|------|---------|-------------|
| `hours` | number | 24 | Lookback period (1-168) |
| `sensitivity` | string | "medium" | "low" (>7pt moves, >$5K trades), "medium" (>3pt, >$1K), "high" (>1pt, >$500) |
| `categories` | string | big 6 | Comma-separated category slugs. Omit for big 6 (best-picture, best-director, best-actor, best-actress, supporting-actor, supporting-actress) |

### Response Structure

```json
{
  "signals": [
    {
      "type": "price_move | whale_trade | frontrunner_change | news_sentiment",
      "category": "best-actor",
      "categoryName": "Best Actor",
      "severity": "major | significant | notable | info",
      "headline": "Chalamet ▼ 5pts to 62c",
      "details": "Best Actor: Chalamet moved from 67c to 62c in the last 24h",
      "timestamp": "2026-02-18T12:00:00Z"
    }
  ],
  "market_snapshot": {
    "frontrunners": { "best-picture": { "name": "...", "price": 45 } },
    "whale_trade_count_24h": 7,
    "overall_sentiment": "quiet | active | volatile"
  }
}
```

### How to Present Results

- Lead with the `overall_sentiment` and `whale_trade_count_24h`
- List frontrunners with prices
- Show signals grouped by severity (major first)
- If `signals` is empty, say "Markets are quiet -- no significant moves"
- Use severity icons: major = !!!, significant = !!, notable = !, info = i

### Example

```bash
# Default brief (24h, medium sensitivity, big 6 categories)
curl -s "https://waitingformacguffin.com/api/oscar/brief"

# Last 48 hours, high sensitivity, all categories
curl -s "https://waitingformacguffin.com/api/oscar/brief?hours=48&sensitivity=high&categories=best-picture,best-director,best-actor,best-actress,supporting-actor,supporting-actress,best-cinematography,best-original-screenplay,best-adapted-screenplay,best-international-feature,best-film-editing,best-costume-design,best-original-song,best-original-score,best-production-design,best-sound,best-documentary-feature,best-makeup-hairstyling,best-visual-effects"
```

---

## Tool 2: Oscar Research

**When to use**: User asks "Tell me about Chalamet", "Should I bet on X?", "What are the Best Picture odds?", or wants detailed research on a specific nominee or category.

**What it returns**: Deep dive with odds, 7-day trend, precursor wins, whale activity, order book depth + slippage, news, and a data-driven assessment.

### API Call

```bash
curl -s "https://waitingformacguffin.com/api/oscar/research?query=Chalamet"
```

### Parameters

| Param | Type | Default | Description |
|-------|------|---------|-------------|
| `query` | string | (required) | Nominee name, film title, or category slug. Supports fuzzy matching. |
| `include_orderbook` | boolean | true | Include order book depth and slippage analysis |
| `budget_for_slippage` | number | 500 | USD amount for slippage calculation (100-100000) |
| `category` | string | (optional) | Category slug to narrow disambiguation |

### Query Resolution

The query is fuzzy-matched automatically:
1. **Category slug or name**: "best-picture" or "Best Picture" returns category overview
2. **Exact name**: "Timothee Chalamet" (case-insensitive)
3. **Substring**: "Chalamet" finds "Timothee Chalamet"
4. **Diacritics-normalized**: "Timothee" matches "Timothee"
5. **Film title**: matches against the film database
6. **Typo correction**: "Chalmet" resolves via Levenshtein (edit distance <= 3)

### Three Response Modes

**1. Nominee deep-dive** (`mode: "nominee"`) -- single match:

```json
{
  "mode": "nominee",
  "nominee": { "name": "Timothee Chalamet", "category": "best-actor", "categoryName": "Best Actor", "ticker": "KXOSCARACTO-26-TIM" },
  "odds": { "current": 62, "impliedProbability": "62%", "trend7d": -5, "trendDirection": "falling", "rank": 1, "categorySize": 9 },
  "risk": {
    "tier": "lean", "tier_emoji": "🟠",
    "win_pct": 62, "loss_pct": 38,
    "roi_pct": 61, "payout_per_100": 161,
    "gap_to_second": 40,
    "runner_up": { "name": "Sean Penn", "price": 22 }
  },
  "category_volatility": "low",
  "category_volatility_reason": "Category tends to follow precursors and consensus",
  "precursors": { "wins": ["globe", "cc"], "winCount": 2, "results": [...] },
  "whaleActivity": { "tradeCount": 3, "totalVolumeUsd": 20200, "sentiment": "mixed", "directionRatio": 0.59, "recentTrades": [...] },
  "orderBook": { "bestAsk": 62, "depthAtBest": 847, "slippageAnalysis": [{ "budgetUsd": 500, "avgFillPrice": 62.4, "slippagePct": 0.6, "assessment": "healthy" }] },
  "news": [{ "title": "...", "source": "THR", "sentiment": "negative" }],
  "assessment": { "summary": "...", "edgeIndicator": "strong_value | fair_value | overpriced | uncertain", "risks": [...], "catalysts": [...] }
}
```

**2. Category overview** (`mode: "category"`) -- query is a category:

```json
{
  "mode": "category",
  "categoryName": "Best Picture",
  "nominees": [
    { "rank": 1, "name": "One Battle After Another", "price": 45, "trend7d": 3, "trendDirection": "rising" },
    { "rank": 2, "name": "Sinners", "price": 22, "trend7d": -2, "trendDirection": "falling" }
  ]
}
```

**3. Disambiguation** (`mode: "disambiguation"`) -- multiple matches:

```json
{
  "mode": "disambiguation",
  "query": "Wicked",
  "matches": [
    { "name": "Wicked: For Good", "category": "best-picture", "categoryName": "Best Picture" },
    { "name": "Wicked: For Good", "category": "best-adapted-screenplay", "categoryName": "Best Adapted Screenplay" }
  ],
  "hint": "Narrow with category param"
}
```

When you get disambiguation, ask the user which category they mean, then re-call with `&category=best-picture`.

### How to Present Results

**Nominee deep-dive** -- present in this order:
1. Name, category, and ticker
2. Odds: current price, implied probability, 7d trend (with arrow), rank
3. Precursors: list wins with award names
4. Whale activity: trade count, total volume, directional sentiment
5. Order book: best ask, depth, slippage at the user's budget
6. News: relevant headlines with source and sentiment
7. Assessment: summary, edge indicator, risks and catalysts

**Category overview** -- present as a ranked table with price and trend.

**Disambiguation** -- list the matches and ask user to pick a category.

### Examples

```bash
# Nominee deep-dive
curl -s "https://waitingformacguffin.com/api/oscar/research?query=Chalamet"

# Category overview
curl -s "https://waitingformacguffin.com/api/oscar/research?query=best-picture"

# With custom slippage budget
curl -s "https://waitingformacguffin.com/api/oscar/research?query=Chalamet&budget_for_slippage=2000"

# Narrow disambiguation
curl -s "https://waitingformacguffin.com/api/oscar/research?query=Wicked&category=best-picture"

# Skip order book (faster)
curl -s "https://waitingformacguffin.com/api/oscar/research?query=Chalamet&include_orderbook=false"
```

---

## Available Categories

`best-picture`, `best-director`, `best-actor`, `best-actress`, `supporting-actor`, `supporting-actress`, `best-cinematography`, `best-original-screenplay`, `best-adapted-screenplay`, `best-international-feature`, `best-film-editing`, `best-costume-design`, `best-original-song`, `best-original-score`, `best-production-design`, `best-sound`, `best-documentary-feature`, `best-makeup-hairstyling`, `best-visual-effects`

## Slippage Assessment Scale

| Level | Slippage | Meaning |
|-------|----------|---------|
| healthy | <= 1% | Clean fill, safe to size up |
| moderate | 1-3% | Acceptable for most bets |
| thin | 3-7% | Consider splitting into smaller orders |
| dangerous | > 7% | Order book too thin, risk of bad fill |

## Edge Indicator Meanings

| Indicator | Meaning |
|-----------|---------|
| strong_value | Multiple bullish signals, price may be undervalued |
| fair_value | Signals balanced, price reflects available data |
| overpriced | Risk signals outweigh catalysts |
| uncertain | Mixed or insufficient signals |

## Important Notes

- Odds are in cents (1-99), representing implied probability percentage
- Whale trades are $1,000+ single transactions
- Precursor awards (DGA, SAG, BAFTA, etc.) historically correlate with Oscar outcomes
- Order book data is from Kalshi prediction markets
- Assessment is data-driven and heuristic, not financial advice

---

## Bet Recommendation Mode

### Intent Detection

Switch to **bet recommendation mode** when the user's query matches any of these patterns:
- "Give me bets", "best bets", "sure things", "safe bets", "high confidence picks"
- "What should I bet on?", "Where should I put my money?"
- "Best picks for $X", "How to bet $100 on Oscars"
- "Build me a portfolio", "conservative picks", "aggressive bets"
- Any query that explicitly asks for **recommendations**, **picks**, or **what to bet**

Stay in **informational mode** for:
- "Tell me about Chalamet" (deep dive, no recommendation framing)
- "What are Best Picture odds?" (category overview)
- "Oscar brief" / "What's happening?" (market pulse)
- Simple lookups, category overviews, or disambiguation

### How to Build Bet Picks

1. Use the **Oscar Brief** to identify frontrunners across categories
2. For each pick candidate, call **Oscar Research** to get the full `risk` object
3. Present each pick using the format below

### Per-Pick Presentation Format

For each recommended pick, present as a structured tree:

```
{tier_emoji} **{Name}** -- {Category}
├─ Price: {current}c ({win_pct}% win / {loss_pct}% loss)
├─ ROI: ${payout_per_100} back on $100 bet (+{roi_pct}%)
├─ Gap: {gap_to_second}pts ahead of {runner_up.name} ({runner_up.price}c)
├─ Precursors: {winCount} wins ({wins list})
├─ Whales: {sentiment} ({totalVolumeUsd} volume)
├─ Volatility: {category_volatility} -- {category_volatility_reason}
└─ Verdict: {1-sentence assessment summary}
```

### Risk Tier Table

Always show this legend when presenting 2+ picks:

| Tier | Emoji | Win % Range | Meaning |
|------|-------|-------------|---------|
| Near lock | 🟢 | 85%+ | Highest confidence, lowest ROI |
| Strong favorite | 🟡 | 70-84% | Solid pick, moderate ROI |
| Lean | 🟠 | 45-69% | Has edge but real downside |
| Toss-up | 🔴 | <45% | High risk, high reward |

### Language Rules

- **Never say "sure thing"** for any pick priced below 85c
- **"Lock" or "near-lock"** only for 85c+ (🟢 tier)
- Always state **explicit percentages** -- "67% chance to win" not "likely"
- Always state the **loss probability** -- "33% chance you lose your $100"
- Frame ROI in dollars: "$149 back on a $100 bet" not just "49% ROI"
- Include the **volatility caveat** for high-volatility categories: "Supporting categories are historically unpredictable -- even favorites get upset"

### Comparison Table

When presenting **3 or more picks**, always include a summary comparison table:

```
| Pick | Tier | Price | Win% | ROI | Gap | Precursors |
|------|------|-------|------|-----|-----|------------|
| Name | 🟢   | 89c   | 89%  | +12%| 72  | 5 wins     |
| Name | 🟡   | 74c   | 74%  | +35%| 45  | 3 wins     |
| Name | 🟠   | 55c   | 55%  | +82%| 20  | 2 wins     |
```

### Portfolio Suggestions

When users ask for portfolio-style recommendations or "how to bet $X", offer tiered portfolio options:

**Conservative (lowest risk)**
- Only 🟢 near-lock picks
- Lower total ROI but highest confidence
- "If you want to sleep easy"

**Balanced (recommended)**
- Mix of 🟢 and 🟡 picks
- Good ROI with solid confidence
- "Best risk/reward tradeoff"

**Aggressive (highest ROI)**
- Best ROI picks from 🟡 and 🟠 tiers
- Higher potential return, real chance of losses
- "Swing for the fences"

Example portfolio format:
```
**Balanced Portfolio -- $100 budget**
| Pick | Tier | Allocation | If Win |
|------|------|-----------|--------|
| Name | 🟢   | $40       | $45    |
| Name | 🟡   | $35       | $47    |
| Name | 🟠   | $25       | $45    |
| **Total** | | **$100** | **$137** (+37%) |
```

### When NOT to Use Bet Mode

Even if the user asks about betting, stay informational if:
- They ask about a **single specific nominee** ("Should I bet on Chalamet?") -- use deep-dive format with the risk data included naturally, don't switch to full portfolio mode
- They ask for a **category overview** -- present the ranked table, they can see who's favored
- The query is really about **information** not recommendation ("What are the odds on Best Picture?")