Start 5

Inspiration

Inspired by 82-0.com, Start Five aims to challenge players to build the best five-man lineup with not the biggest names or highest scorers, but the best actual lineup.

Basketball fans do this kind of reasoning constantly. You weigh stars against fit, shooting against defense, size against speed, and roles against talent. Start Five turns that instinct into a fast game.

What it does

Start Five gives the player a random set of NBA players and asks them to build a starting five. Once they lock in their lineup, the game evaluates the team through a simulator and gives feedback on how strong the lineup is.

The goal is to make the player think like a coach or GM:

Do you take the best player available, even if the fit is awkward? Do you need more shooting? Can your defense survive? Are you building a balanced lineup or just collecting names?

The fun comes from making a choice, seeing how the simulator judges it, and immediately wanting to try again.

How we built it

We built Start Five as a Reddit Devvit Web app. The game runs inside Reddit, so players can launch it directly from a post without needing to install anything or leave the platform.

The core pieces are:

• A randomized player pool
• A lineup builder for selecting five players
• A scoring/simulation system that evaluates the lineup
• A lightweight game loop built for quick replayability

The simulator looks at the lineup as a whole rather than treating the team as a simple sum of player ratings. A good team needs roles, spacing, defense, and star power. In rough terms, the scoring model is closer to:

$$ TeamScore = Talent + Fit + Balance + MatchupStrength $$

The exact result depends on how the chosen players work together, not just who has the highest individual rating.

Challenges we faced

The hardest part was making the simulation feel fair.

Sports games can break quickly if the model rewards the wrong things. If raw talent matters too much, the game becomes obvious. If fit matters too much, superstar players feel artificially weak. We had to tune the system so that great players still feel great, but lineup construction still matters.

Another challenge was scope. There are a lot of features we want eventually: more modes, more player pools, better explanations, deeper comparisons, and maybe social features later. For the hackathon, we focused on the core loop first: random pool, pick five, simulate, replay.

What we learned

We learned that a simple sports idea gets interesting fast once you force real tradeoffs.

A five-player lineup is small enough to understand immediately, but complex enough that there is rarely one obvious answer. The same player can be amazing in one lineup and awkward in another. That made the project more about judgment than trivia, which is exactly what we wanted.

We also learned a lot about building inside Reddit with Devvit: how to package the app, test it, publish it, and design around the constraints of a platform-native experience.

Built With

• css
• devvit
• html
• javascript
• nba-api
• react
• typescript