Acca Tracker

Track accumulator (acca) betting slips across football, basketball, and tennis — parses slip photo or text, checks live scores every 15 minutes, and reports bet status with overall acca health and cash-out context.

Acca Tracker — Multi-Sport Accumulator Monitor

Track accumulator (parlay) bets across football, basketball, and tennis by monitoring live scores and reporting whether each leg is still alive. Works with any betting slip — photo, screenshot, or typed text.

When to Use

• User shares a photo/screenshot of a betting slip (accumulator/parlay/bet builder)
• User wants live score updates for their bets
• User says "track my acca", "monitor my bet", "is my bet alive", "check my slip"
• Slips can mix sports (e.g., 2 football + 1 basketball + 1 tennis)

How It Works

text

User sends slip photo
  → Agent parses legs (vision_analyze or text extraction)
  → Agent creates a cron job (*/15 * * * *)
  → Cron searches live scores every 15 min
  → Cron delivers report: per-leg status + acca health
  → Auto-cleans up after all matches finish

Step 1: Parse the Slip

From a Photo

Use vision_analyze with the image and this question:

text

Extract all betting legs from this slip. For each leg return:
- Match (Team A vs Team B) or Event (Player A vs Player B for tennis)
- Sport (football/soccer, basketball, tennis)
- Competition/League
- Date and kickoff time
- Bet type (exact wording from slip)
- Decimal odds

Also extract: total odds, stake amount, max return, any bonus info.

Sport Detection

Auto-detect sport per leg based on:

• Tennis: single names/players, WTA/ATP/Grand Slam, "set" in bet type, no team names
• Basketball: NBA, EuroLeague, BBL, "quarter", "points", "rebounds"
• Football/Soccer: everything else (default)

Tag each leg with its sport — the cron prompt uses sport-specific APIs.

From Text

If user types their slip, extract the same fields. Normalize bet type wording to match the standard types in references/bet-types.md.

Step 2: Build the Slip Summary

After parsing, confirm with the user before starting tracking:

text

📋 PARSED SLIP — 5 legs

 1. Arsenal vs PSG (⚽ UCL, 21:00) — Arsenal W (1.55) — 64.5%
 2. Bayern vs Inter (⚽ UCL, 21:00) — Bayern W (1.40) — 71.4%
 3. Lakers vs Celtics (🏀 NBA, 02:30) — Lakers W (1.80) — 55.6%
 4. Djokovic vs Alcaraz (🎾 ATP, 14:00) — Djokovic W (2.10) — 47.6%
 5. FSV Schöningen vs Lohne (⚽ RL, 18:00) — BTTS No (2.65) — 37.7% ⚠️

Stake: €10 | Combined odds: 30.12 | Max return: €301.20
Riskiest leg: #5 (37.7% implied)

Start tracking? (yes/no)

Calculate implied probability per leg: 1 / odds * 100. Flag any under 40% as ⚠️ gamble leg.

Step 3: Create the Cron Job

Use the cronjob tool:

text

action: create
name: acca-tracker-{short-id}
schedule: "*/15 * * * *"
repeat: 48
deliver: origin
prompt: {see template below}

Schedule Adjustments

Scenario	Schedule	Repeat
Single match day	*/15 * * * *	48
High-stakes acca	*/10 * * * *	72
Multi-day (Tue + Wed)	*/15 * * * *	192
Tournament bracket	0 */2 * * *	84

Cron Prompt Template

text

You are tracking a MULTI-SPORT accumulator bet. Check live scores NOW and report.

SLIP DETAILS:
{paste all legs with sport tags, bet types, odds, and win conditions}

Total odds: {total} | Stake: {stake} | Max return: {max_return}

INSTRUCTIONS:
1. For each leg, determine scores using the scores.sh helper script:
   - Run in terminal: bash scripts/scores.sh {DATE} Soccer
   - Run in terminal: bash scripts/scores.sh {DATE} Basketball
   - Run in terminal: bash scripts/scores.sh {DATE} Tennis
   - Output: league|home|away|home_score|away_score|status (or league|event|result|status for tennis)
   - Match results to legs by team/player names
   - If script returns NO_DATA, fall back to web_search as last resort
2. For each leg determine:
   - Current score
   - Match status: Not Started / Live (minute/set) / HT / FT / Postponed / Abandoned
   - Sport-specific score format (football: goals, basketball: points, tennis: sets+games)
   - Bet status: ✅ WON / ✅ WINNING / ⏳ PENDING / ❌ LOST / ❌ DEAD
3. Calculate overall acca status:
   - ALL SAFE ✅ = no legs lost
   - STILL ALIVE ⏳ = no legs lost, none confirmed won
   - ACCA DEAD ❌ = at least one leg LOST
4. If acca dead: name the killing leg, explain why
5. Calculate "legs alive / won / lost / pending / total"
6. If all FT: state "TRACKING COMPLETE" and summarize

SPORT-SPECIFIC BET EVALUATION:
- Football: standard 90min + stoppage (see references/bet-types.md)
- Basketball: Match Winner = team with more points at FT. Overtime counts.
  Over/Under = total combined points of both teams.
- Tennis: Match Winner = player who wins required sets (2/3 or 3/5).
  Set betting = exact set score. Games over/Under = total games in match.

REPORT FORMAT (code blocks only):
🏟️ ACCA LIVE REPORT — {time}

Leg | Match                          | Score        | Status | Bet        | Result
----|--------------------------------|--------------|--------|------------|----------
 1  | ⚽ Arsenal vs PSG              | 2 - 0 (67')  | 67'    | Arsenal W  | ✅ WINNING
 2  | 🏀 Lakers vs Celtics           | 98 - 102     | Q4     | Lakers W   | ❌ DEAD
 3  | 🎾 Djokovic vs Alcaraz         | 6-4, 3-2     | Set 2  | Djokovic W | ✅ WINNING

📊 1 winning / 0 won / 1 dead / 1 pending (3 total)
❌ ACCA DEAD — Leg 2 killed it (Lakers trailing 98-102 Q4)

DATA: If no score found, say so explicitly. Never guess scores.
If 2+ hours past kickoff with no data: "unverified — likely finished"

Step 4: Report States

All Alive

text

🏟️ ACCA LIVE — 21:34 CET

 1 | Arsenal vs PSG       | 1 - 0 | 55' | Arsenal W | ✅ WINNING
 2 | Bayern vs Inter      | 0 - 0 | 55' | Bayern W  | ⏳ PENDING
 3 | Luton vs Northampton | 0 - 0 | 55' | BTTS No   | ✅ WINNING

📊 1 winning / 0 won / 0 lost / 2 pending
💰 ⏳ STILL ALIVE

Acca Dead

text

🏟️ ACCA LIVE — 22:18 CET

 1 | Arsenal vs PSG       | 2 - 1 | 78' | Arsenal W | ✅ WINNING
 2 | Bayern vs Inter      | 1 - 2 | 78' | Bayern W  | ❌ DEAD
 3 | Luton vs Northampton | 0 - 0 | 78' | BTTS No   | ✅ WINNING

📊 1 winning / 0 won / 1 dead / 1 pending
❌ ACCA DEAD — Leg 2 killed it (Bayern losing 1-2)

Remaining matches still playing — updating for interest.

Final — Won

text

🏟️ ACCA FINAL — 23:45 CET — 🎉🎉🎉

 1 | Arsenal vs PSG       | 2 - 1 | FT | Arsenal W | ✅ WON
 2 | Bayern vs Inter      | 3 - 1 | FT | Bayern W  | ✅ WON
 3 | Luton vs Northampton | 0 - 0 | FT | BTTS No   | ✅ WON

📊 3/3 LEGS WON
💰 ACCA WON — Stake: €10 | Return: €210.40

Tracking complete.

Cash-Out Context

When some legs won and others pending, estimate cash-out value:

text

Won legs combined odds: 1.55 × 1.40 = 2.17
Remaining legs implied probability: 0.565 × 0.541 × 0.377 = 0.115
Estimated value: €10 × 2.17 × 0.115 = €2.49
Cash-out range: ~€2.00 – €3.50 (rough estimate)

Only include when: at least 1 leg WON, at least 1 PENDING, acca NOT dead. Always frame as "estimated".

Data Source Strategy

See references/data-sources.md for full API details.

Quick reference:

• PRIMARY: scripts/scores.sh {date} {sport} — handles all transport, all sports
• FALLBACK: execute_code with Python/urllib (same API, different transport)
• WEB SEARCH: last resort when script returns NO_DATA
• TENNIS NOTE: Player names in strEvent (not strHomeTeam/strAwayTeam), coverage tournament-dependent

Bet Types

See references/bet-types.md for the full list of 18+ bet types with scoring logic, decision tree, and edge cases (void, ET, red cards, handicaps).

Step 5: Cleanup

• After all matches FT, deliver final summary then stop
• Auto-ends after repeat count
• User can say "stop tracking" to remove cron early
• Multi-day accas: increase repeat to 192

Limitations

• 15-min intervals — goals/points missed between checks, not real-time
• Low-tier football leagues — limited coverage, may only get HT/FT
• JS-rendered sites — Flashscore/Sofascore can't be scraped directly
• Tennis player names — TheSportsDB stores in strEvent, parsing required
• Bookmaker-specific rules — some bet types vary by bookmaker (e.g., retirement handling)
• Telegram cron delivery — if reports fail to arrive, check last_delivery_error. Try explicit telegram:chat_id:thread_id format. Reports still generate at ~/.hermes/cron/output/{job_id}/.md

Tips

• Always confirm slip with user before starting
• Include team nicknames in search queries
• When acca dies, keep tracking for interest
• Flag "near misses" (hit post, disallowed goal) if mentioned in results
• Name each cron differently for multi-slip support
• Tennis coverage on TheSportsDB is tournament-dependent — may return zero events between tournaments. Always have web_search as fallback.
• Basketball events are rich on TheSportsDB — covers NBA, EuroLeague, BBL, LKL, LNB, Japanese B1, etc. ESPN is backup only.

ClawHub Publishing

• Current slug: sports-acca-tracker (old football-acca-tracker also exists, no delete cmd)
• Publish command: clawhub publish <path> --slug <slug> --name "Name" --version X.Y.Z --tags "tag1,tag2"
• After publish, skill is hidden until security scan completes (usually minutes)
• Verify with: clawhub inspect <slug>
• ClawHub has no unpublish/delete — slug collisions mean choosing a new slug, not overwriting