Overview

I built the player rating and live draft system for Stonewall Sports from scratch, turning a three-person subjective scoring process into a multi-dimensional, data-driven platform that produces competitive teams and runs in a fraction of the time.

Previously, three board members independently rated each player on a 1–4 scale with no defined criteria, leaving each rater to supply their own interpretation of what the number meant. Scores were averaged and rounded, carrying individual bias directly into draft rankings with no way to reconcile the gaps. Most seasons produced one dominant team, two middling ones, and one that rarely won. Average game score delta sat at 2.9 points, and roughly 20% of players couldn’t be identified by name during the draft itself.

I designed and built the system across two phases. The first used Google Sheets to validate the rating model: six skills across three weighted categories, scored on a 1–7 Likert scale, with Offense and Defense combining into a Base Score and Psych functioning as an exponential multiplier rather than an additive input. Raters worked in siloed tabs with no visibility into each other’s scores. Season 3 migrated to a purpose-built web application, where raters authenticated with bcrypt-hashed credentials, scored players with auto-save and CSV import for returning-player continuity, and admins monitored progress in real time. The live draft interface ran on captains’ phones, with real-time pick updates across all devices, quota-enforced eligibility filtering, and profile photos pulled from linked social accounts.

The result is a system that removes subjectivity from team construction, enforces competitive balance across the entire player pool, and runs a full draft in 15 minutes.

Impact

• Average game score delta dropped from 2.9 to 1.7 within the first two seasons
• Draft duration reduced from ~60 minutes to 15 minutes in Season 3
• Player identification issues dropped from ~20% to 0% after profile photo integration
• System adopted and run successfully for 3 consecutive seasons
• Rating data carries forward across seasons via CSV import for returning players

Tools Used

Google Sheets · Next.js · React 19 · TypeScript · PostgreSQL · Prisma ORM · Cursor AI · LeagueApps · Render

Beyond the competitive improvements, the platform is built to grow. New seasons, expanded player pools, or additional sports divisions can be onboarded without rebuilding the core infrastructure.

TECHNICAL DOCUMENTATION

For a full breakdown of the system architecture, tools, and implementation: