Swoops v0.1.0 Release Notes
Welcome to the first beta release of Swoops! Version 0.1.0 marks the beginning of season 0 and the introduction of our first gameplay mode: head-to-head contests.
What’s new in v0.1.0
We’ve released our first gameplay mode: head-to-head contests! Users can now head over to the Court (our game lobby), find a list of games to join, submit their lineup, and directly play against an opponent.
These initial games are blind contests where each user doesn’t know about the other’s lineup until both lineups have been submitted. It’s also an asynchronous mode, enabling users to play at their own leisure without needing to be at the keyboard together.
We have the most advanced basketball simulator in the world and this release makes use of that. After trialing our first version of the simulator on SwoopsGM, we took lessons from that body of work and created the current iteration of our simulator.
Written in python by our team of data scientists (who’ve all worked for an NBA team or the NBA itself), it is a core part of our product and what we think differentiates us. Built on tens of thousands of data points sourced from real NBA data, our simulator is event based and makes use of various machine learning libraries including pandas and scikit.
Every game has a play-by-play event log attached to it. We include this (1) to show some of the inner workings of our simulator and (2) to give users a resource of data to analyze how their game went. We’re going to tweak this section the most to make it more readable and interactive in the future.
This is our first big release and we have some rough edges that the team is going to be polishing out. Here’s what’s on our radar for the next release:
- Some users are reporting that their player intermittently disappears from their locker room. Others are reporting seeing players that they recently sold still appear. We’re looking to release a patch early next week to resolve this.
- Running averages on players (e.g. ppg, rpg, etc) sometimes drift - we’re monitoring this closely. There may be instances where these stats take a couple minutes to update after the completion of a game.
- If both teams are unnamed, you get this awkward scenario where it is "UNNAMED VS UNNAMED". We have a solution in mind which involves using the user's last four digits of their wallet address.
- The play-by-play log shows our players’ long IDs. We’ll deploy a fix soon to remove these and show the actual players’ names.
- Our game is currently a desktop online experience. We made this choice partly because of UX considerations (crypto wallets on mobile are still maturing) and partly because a desktop resolution lends itself to a better environment to show a lot of numbers and data. It is on our roadmap to deliver a mobile experience.