About Cradl

AI-powered NCAA lacrosse analysis

What is Cradl?

Cradl is an AI-powered analysis tool for NCAA Division I Men's Lacrosse. We combine real-time odds, three seasons of game data, box score stats, and a custom Elo model — then feed it all to AI to produce sharp, data-driven matchup breakdowns.

Cradl is not a sportsbook — we don't accept wagers or pay out winnings. All analysis is for informational and entertainment purposes only.

How It Works

• Live odds — DraftKings-preferred spreads, totals, and moneylines via The Odds API, updated every 30 minutes
• Historical stats — three seasons (2024–present) of NCAA lacrosse results, scores, and box scores
• Statistical model — generates predicted spreads, totals, and win probabilities for every matchup
• AI analysis — Claude synthesizes all data into a single breakdown with edge ratings and picks for spread, total, and moneyline

The Model

Every analysis is built on a multi-layer statistical model that evaluates both teams before AI generates the final breakdown:

• Power ratings — Elo-style ratings (1500 = average) for every D1 team, updated after each game with opponent-adjusted margins, per-team home advantage, and 50% season regression toward conference average
• Spread prediction — converts the power rating gap into an expected margin, adjusted by Process Quality differential
• Total prediction — uses venue-specific offensive and defensive splits (home offense at home, away defense on road) with a pace adjustment relative to the league average
• Process Quality (PQ) — a 0–100 composite score from opponent-adjusted box score stats: faceoff %, ground balls, clear %, shot accuracy, and extra-man offense %
• Venue splits — goals scored and allowed at home vs on the road for each team
• Opponent-adjusted stats — how a team performs relative to what their opponents typically allow, not just raw averages
• Strength of schedule — opponent win %, RPI, and record broken down by tier (elite, solid, weak)
• Trends — win streaks, scoring trajectory (last 3 vs season average), O/U trends, margin distributions, and volatility
• Line movement — how odds shift from open to current, signaling where money is going
• Head-to-head history — prior matchups and scoring margins from the last three seasons

The model compares its predicted spread and total against the market line and grades the gap as an Edge — None, Slight, or Moderate. AI then explains the why: whether supporting signals (splits, trajectory, PQ, rest, tier record) back that edge.

The Edge Read

• Edge — how much validated advantage the model sees on a market: None, Slight, or Moderate. Strong is reserved — never shown until a category earns it from live graded results.
• Receipts — the real win–loss record for picks at that Edge this season, shown right next to it; “still accruing” until enough games have graded.
• Picks Tracker — every AI pick is logged and graded against actual results so you can track accuracy over time

Data Sources

• The Odds API — real-time moneyline, spread, and total odds from licensed sportsbooks (DraftKings preferred)
• ESPN — historical game scores, team records, schedules, and live scores
• NCAA / Sidearm Sports — box scores including player stats, faceoffs, ground balls, and goalie saves
• Anthropic Claude — AI analysis engine that synthesizes all model outputs and data into readable insights

Built by Blaine McMahon

Cradl is a solo project built by Blaine McMahon — a computer engineering grad and former lacrosse player at UMass Lowell who wanted better tools for analyzing NCAA matchups. If you have questions or feedback, reach out at [email protected].