P0: Data Integrity - Spot-Check Current Cards Against Real Data

[kevsilk597]

Objective

Before building automated verification, manually verify that our current cards contain accurate information. If they don't, we need to know now.

Current Cards to Verify

We have 6 active cards. Each must be manually checked.

Card 1: RIVALRY - NYK @ BKN (March 20)

Claims to verify:

• "History between these squads" — Is there actual historical data showing a rivalry pattern?
• Are NYK and BKN in our games table for March 20?

Verification steps:

-- Verify game exists
SELECT * FROM games WHERE game_date = '2026-03-20' AND home_team = 'BKN' AND away_team = 'NYK';

-- Check historical matchups
SELECT game_date, home_team, away_team, home_score, away_score
FROM games
WHERE (home_team = 'BKN' AND away_team = 'NYK')
   OR (home_team = 'NYK' AND away_team = 'BKN')
ORDER BY game_date DESC
LIMIT 10;

Card 2: TRAP_GAME - OKC @ PHI (March 23)

Claims to verify:

1. "OKC is 4-point favorite" — Check odds table
2. "just beat two ranked teams" — Check OKC's last 2 games
3. "76ers are 0-3 at home when favored teams roll in hot" — Complex query needed
4. "Thunder travel tonight. Sixers don't." — Check schedule

Verification steps:

-- 1. OKC spread
SELECT * FROM odds WHERE game_id LIKE '%okc%phi%' OR game_id LIKE '%20260323%';
SELECT * FROM game_odds WHERE game_id LIKE '%okc%phi%';

-- 2. OKC last 2 games
SELECT game_date, home_team, away_team, home_score, away_score
FROM games
WHERE (home_team = 'OKC' OR away_team = 'OKC')
AND game_date < '2026-03-23'
ORDER BY game_date DESC
LIMIT 2;

-- 3. PHI home record vs hot favorites (this is complex - document the exact query)
-- Need to define: what is "hot"? what is "favored"?

-- 4. Schedule check
SELECT game_date, home_team, away_team 
FROM games 
WHERE (home_team IN ('OKC', 'PHI') OR away_team IN ('OKC', 'PHI'))
AND game_date BETWEEN '2026-03-21' AND '2026-03-23'
ORDER BY game_date;

Card 3-6: SLEEPER Cards

Claims to verify:

• Each claims a player "is out"
• Verify against injuries table

SELECT player_name, team, injury_status, injury_date, return_date
FROM injuries
WHERE player_name IN ('Anthony Edwards', 'Jaylen Brown', 'Michael Porter Jr.', ...)
ORDER BY injury_date DESC;

Critical: If these players are NOT actually injured, the cards are hallucinations and must be deleted.

Verification Process

For each card:

1. Extract claims — List every factual statement
2. Write query — SQL to verify each claim
3. Run query — Execute against production database
4. Compare — Does DB result match card claim?
5. Document — Record result in this issue

Verification Report Template

### Card: [CARD_ID]

| Claim | Query | DB Result | Card Says | Match? |
|-------|-------|-----------|-----------|--------|
| OKC is 4-point favorite | SELECT spread FROM... | -4.5 | -4 | ⚠️ Close |
| 76ers are 0-3 at home... | SELECT ... | 1-2 | 0-3 | ❌ WRONG |

**Verdict:** PASS / FAIL / NEEDS REVIEW
**Action:** None / Fix card / Delete card / Fix data source

Immediate Actions Based on Findings

Finding	Action
Claim is accurate	Document the verification query for future automation
Claim is slightly off (rounding)	Acceptable if within 5%, document threshold
Claim is wrong	Delete the card immediately, investigate source
Claim is unverifiable (no data)	Delete the card, add data source or remove claim type

Requirements

1. Complete verification report for all 6 current cards
2. Delete any cards with false claims
3. Document the verification queries used
4. Identify any data gaps (claims we can't verify because data is missing)

Definition of Done

• All 6 cards manually verified
• Verification report posted in this issue
• Any cards with false claims deleted
• Verification queries documented for reuse
• Data gaps identified and logged as separate issues
• Confidence level: we can defend every stat in every live card