Your chess history should be something you can keep. In Scoresheet, both players sign the game’s scoresheet with their Nimiq keys. Your Scoresheet rating is calculated from those signed records. You can export your history, verify the signatures, and recompute your rating without our API. That’s the reason for the wallet integration: a record of your games that you can verify yourself. There’s a full chess app behind it: - Play and review. A custom TypeScript engine and plain-English explanations of mistakes, based on board analysis—not an LLM. When the reviewer cannot support an explanation, it falls back to the numerical evaluation. - Practice on your own. 5,000 offline puzzles, a daily puzzle, Puzzle Storm, and 12 tactical themes. Online Solo Rated Runs offer server-countersigned results without needing an opponent. - Run tournaments. Round-robin and Swiss formats, with standings calculated in the browser from verified, signed games. It’s built for Nimiq Pay, with wallet detection, light and dark themes, five languages, and testing on real Android devices. The project includes 731 unit tests and 590 automated browser checks. One test blocks every request to our API and checks that independent verification still works. Play: https://chess-liard-sigma.vercel.app Explore the code: https://github.com/Pratiikpy/scoresheet

Two weeks into Scoresheet. Before we wrote a single line of app code, we spent 17,625 lines understanding what already works. We tore apart Lichess and Chess.com as products, lila as an architecture, Stockfish, Fishtest, Lc0, Maia and Chessformer on the engine side — then went deeper into how players actually analyse games, how puzzles are selected, and how a chessboard should behave on a phone. 80 research notes came first. Every note answers the same nine questions. Every claim has a source URL or is explicitly marked NOT VERIFIED. The nine questions are the same for every note, in the same order: what they do · why it works · what they do badly · what we should copy conceptually · what we can do better · what it would technically take · what could break · what we can uniquely do because of Nimiq That template is in the repo. Check it yourself. What came out of it: • Game review that explains why, not just “-2.4”. It tells you the piece you left hanging, the pin you walked into — then gives the position back so you can find the better move yourself. • Lichess-style puzzles: Daily, Endless, Storm, Streak and 12 themes. 5,000 puzzles bundled into the app, playable completely offline. • Your own mistakes become puzzles — but only after the position passes our soundness checks. • Bots that make human-like mistakes, inspired by the Maia approach, instead of simply being weakened engines. • A board built for us: drag, tap-tap, cancellable premoves, in-place promotion, keyboard play, and full square announcements for accessibility. Even chessground, the board used by Lichess, ships without ARIA. • Tournaments without arbitrary draws. Round-robin up to 8 players, Swiss beyond that, with the standings explicitly showing which tiebreak rule decided each position. And then there’s the part that only makes sense because of Nimiq. Both players sign the result with their wallets. The rating is derived from those signatures. We don't issue it. We can't revoke it.

Spent this week tearing apart Upwork, Fiverr, Freelancer.com and Contra — their apps, their fees, and about 600 of their one-star reviews. Sharing what I found and what I'm building, because I'd rather be told it's wrong now. What the reviews actually say. The loudest complaints aren't about features. They're paying just to apply (Connects, gig fees, paid memberships — one reviewer called it a platform that "makes people beggars"), getting locked out of money you already earned, and withdrawal fees eating a small balance. And the structural one: a live job on Freelancer.com — a full service website, $126 average bid, 135 people bidding. 134 of them worked for free. Who it's for: working freelancers. Fiverr sellers losing 20% per order and waiting two weeks to clear. Upwork freelancers buying Connects for jobs they don't win. Designers, editors, writers, developers. What chit is. The whole app is one object — a chit, one line of text. - Money attached = I need this. - Empty hands = I can do this. That's the entire interface. No post-a-job vs browse-gigs choice, no proposals inbox, no bidding screen. The amount you attach configures everything else — how many people can attempt it, the deadline, whether the money is held or sent instantly. A small amount opens a race. A large amount becomes one private hire with the money held until you accept. Same screen. Four things that only work because of Nimiq: 1. Everyone whose work was good gets paid — not just the winner. Paying five people 50c each costs essentially nothing on NIM. On card rails the fees alone exceed the payment, which is precisely why no platform does this. 2. We never hold your money. It goes straight to your wallet. Nothing to freeze, and withdrawal doesn't exist as a concept. 3. Nobody builds a profile. A listing is written when money moves, never by filling a form — so your profile is computed from work you were actually paid for. No bio, no gig editor, nothing to fake. And you can take it with you, signed and verifiable off-platform.

Week 1 of Cycle 2 is officially done! If you’re still in the idea stage, don’t rush into the next 3 weeks of building just because you have something you could build. This is the perfect time to pressure-test your idea. Ask yourself: 💡 Who is my Mini App built for? 💡 What problem does it solve? 💡 Why would they come back? And remember, you don’t need to build the most beautiful or technically impressive Mini App to win. You need to build something useful. We have a Validate Your Idea channel specifically for this. Share your concept, get feedback from other builders and the competition team, and strengthen your idea before committing weeks to it. ℹ️ Scoring reminder: We’ve updated the Cycle 2 scoring to put more emphasis on the Mini App itself, with greater weight on the idea, build, and overall product experience. The four scoring categories remain the same, but the weight on marketing has been adjusted based on feedback from Cycle 1. 📖 Scoring guide: miniappscompetition.com/scoring ⏰ Submission deadline: September 18 at 23:59 UTC 📤 Submission portal: miniappscompetition.com/submissions/cycle2 You still have 3 weeks. Before you lock in, make sure you’re building the right thing, then use the remaining time to make it great. Where are you at with your Mini App? Share an update below. 👇

what type of product is highly encouraged for cycle 2??