Youth Hockey Analytics

The Problem

Every fall, a youth hockey league has to sort dozens of teams into competitive tiers. Get it wrong and kids spend the season losing 9–0 or winning 9–0, and nobody learns anything. The decision usually comes down to a weekend of placement games and a room full of volunteers going on gut feel.

The data to do better exists. It's just spread across scoring apps, rating sites, and last season's results, and nobody has the time to pull it together before the deadline.

The Solution

We built a placement board for a regional league: 33 teams, 30 placement games over three days. It pulls game results straight from the league's scoring system, blends them with last season's team ratings, and estimates how strong each team really is.

For each team, the board shows its record, how it did compared to what was expected, the odds it belongs a tier up or down, and where the cut lines between tiers fall. League organizers set how many tiers they want. The board shows the fairest split, and why.

A Coaching Tool, Too

The same analysis turned out to be just as useful on the bench. Coaches use it to see how their team stacks up before a game, which opponents are stronger than their record suggests, and whether a lopsided score was a bad day or a sign of a real gap. It turns "I think we're about a B team" into a number they can plan practices around.

Should There Be a New League?

A group of local clubs asked a bigger question: families were driving over an hour for league games. Could a regional league closer to home work? We pulled team counts for every club within about an hour's drive, broken out by age group and level, and mapped which clubs already play in which leagues.

The answer: there are enough teams to fill a league at 10U, 12U, and 14U. The harder question is the rules, not the numbers. Staying sanctioned by the national governing body matters, because going independent would lock out member clubs. The research became a short brief for the people who will make the pitch.

How We Thought About It

AI did the heavy lifting: gathering results from several sources, matching team names that don't line up across systems, and running the numbers. The final call stays with people. The board makes a recommendation and shows its work. It doesn't place a single team on its own.

The Result

• 33 teams placed using game data instead of gut feel alone
• Clear cut lines and move-up/move-down odds for every team
• A scouting and planning tool coaches use during the season
• 130+ teams across two dozen clubs mapped for a regional league feasibility brief