惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

J
Java Code Geeks
H
Hackread – Cybersecurity News, Data Breaches, AI and More
Blog — PlanetScale
Blog — PlanetScale
G
Google Developers Blog
Microsoft Security Blog
Microsoft Security Blog
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
腾讯CDC
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
Jina AI
Jina AI
雷峰网
雷峰网
T
Tailwind CSS Blog
爱范儿
爱范儿
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
酷 壳 – CoolShell
酷 壳 – CoolShell
大猫的无限游戏
大猫的无限游戏
月光博客
月光博客
博客园 - 司徒正美
I
InfoQ
Engineering at Meta
Engineering at Meta
Vercel News
Vercel News
小众软件
小众软件
U
Unit 42
Google DeepMind News
Google DeepMind News
D
DataBreaches.Net

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
Building Velodrome: A Real-Time Multiplayer Typing Race Game
Daksh Saijwal · 2026-06-20 · via DEV Community

Building Velodrome: A Real-Time Multiplayer Typing Race Game

I built a full-stack typing game with live multiplayer, ghost racer replays, and per-key analytics. This post covers how I solved the hard parts.

Live demo | GitHub

The Idea

Typing games like TypeRacer are fun, but I wanted to add something different. Three things mattered to me:

  1. Ghost racer - your best run replays as an opponent in future races
  2. Per-key analytics - figure out which specific characters trip you up
  3. Real-time multiplayer - race friends live

The ghost racer was the hardest part. How do you actually record and replay a typing performance?

The Ghost Racer

Most typing games just track one number: how many characters you typed. But replaying a race means knowing when each keystroke happened. If I only store the final text, I lose all the timing info.

My solution was to record every keystroke with a timestamp:

// In useGame hook, during the race:
keylogRef.current.push({ 
  t: Date.now() - startRef.current,  // milliseconds elapsed
  len: value.length                   // cumulative characters typed
})

After the race ends, the array looks like:

[
  { t: 0, len: 1 },      // first char at 0ms
  { t: 50, len: 2 },     // second char at 50ms
  { t: 75, len: 3 },     // third char at 75ms
]

Then during a new race, every 500ms I check where the ghost should be at that elapsed time. I use binary search to find the closest keystroke:

// In the game loop while racing:
const elapsed = Date.now() - startRef.current
const ms = elapsed

// Find the keystroke that happened right before this time
let idx = ghost.keylog.findIndex(k => k.t > ms)
if (idx === -1) idx = ghost.keylog.length

// The character count at that point
const ghostLen = idx > 0 ? ghost.keylog[idx - 1].len : 0
const ghostProgress = (ghostLen / passageLength) * 100

setGhostProgress(ghostProgress)

This is O(n) per tick but the array is usually less than 5000 elements, so it's instant. No second game simulation needed, no manual interpolation, and it handles pauses and backspaces automatically.

The ghost gets saved to localStorage only if it's a personal best:

if (newWpm > existingGhost.wpm) {
  profile.ghost = { 
    wpm: newWpm, 
    keylog: keylogRef.current,
    at: Date.now() 
  }
  localStorage.setItem('velodrome.profile.v1', JSON.stringify(profile))
}

Next time you play, your ghost is there waiting. Beat it and it updates.

Per-Key Analytics

Overall accuracy is useless. You could be 98% accurate and still struggle with one specific key. You need to know exactly which characters slow you down.

I track which character caused each error while typing:

// In handleInput, when a new character is added:
if (value[i] !== passage[i]) {
  errorsRef.current += 1

  // Track which character we expected to type was wrong
  const expected = passage[i] === ' ' ? 'space' : passage[i]
  troubleRef.current[expected] = (troubleRef.current[expected] ?? 0) + 1
}

After 20 races, the data might look like:

{
  'e': 8,      // you mistyped 'e' 8 times
  'r': 5,
  's': 4,
  'space': 12  // spaces give you trouble
}

This matters because instead of seeing "98% accuracy" you see "I mess up spaces and the letter e". You can drill those specific keys.

Multiplayer with Socket.io

I had a bug where your own car wouldn't move during multiplayer races, but the opponent's car would move fine.

I was storing all racer data (you and opponents) in one state array:

const [racers, setRacers] = useState([])

// Problem: every time the players list updated, it reset progress to 0
useEffect(() => {
  setRacers(players.map(p => ({ progress: 0, wpm: 0, ... })))
}, [players])

The server only broadcasts opponent progress to other clients. Your own progress never comes back to you. So your car had nothing updating it.

The fix was to stop storing your car in state and just compute it every render from your live typing:

// Instead of storing racers in state:
const racers = players.map(p => {
  const isYou = p.id === socket?.id
  const oppData = oppProgress[p.id]  // only opponents stored in state

  return {
    id: p.id,
    progress: isYou ? progress : (oppData?.progress ?? 0),
    wpm: isYou ? wpm : (oppData?.wpm ?? 0),
    isYou,
  }
})

Now your car always reflects your live progress variable. Opponents update when the server broadcasts. No resets, no stale data.

It's a simple pattern but took debugging to find. The lesson: don't store derived data in state if it updates frequently.

Tech Decisions

Why Canvas for the race track? I could have used CSS progress bars. Canvas was definitely overkill, but it made the cars, dashed lanes, and checkered finish line look custom rather than templated. Recruiters notice that you used Canvas more than they notice CSS bars. It was 150 lines of Canvas code versus 20 lines of CSS. Worth it for a portfolio project.

Why localStorage instead of a database? For a resume project, localStorage works offline, needs no backend infrastructure, and users don't care if their stats disappear. Plus it gives me a good answer in interviews: "Currently uses localStorage for local stats. Next I'd add user authentication to sync across devices."

Why Socket.io? Real-time games need low latency. Socket.io gives bidirectional communication, auto-reconnect on network drops, and falls back to polling if WebSocket fails. It's the standard for multiplayer games.

What Could Be Better

Extract the magic strings. Room codes, phase names, event names are scattered throughout. Should be constants. End-to-end tests would help too, testing the full flow from creating a room to finishing a race. Currently I emit race progress on every keystroke which could be expensive at scale. Tournament mode with multiple rounds would be cool. The whole thing is desktop-focused when mobile would open it up more.

What I Learned

State management is subtle. Storing derived data that updates frequently leads to sync bugs. It's better to compute it. Timestamps are way cheaper than high-frequency updates. Recording one timestamp per keystroke beats sending progress updates 60 times a second. localStorage gets a bad rap but it's perfect for single-player stats and ghosts. Real-time feedback matters. Users notice lag. The 500ms update interval for opponent cars is noticeable but acceptable. And recruit for the architecture, not the UI. The ghost racer algorithm, per-key error tracking, and state machine are the resume-worthy parts.

How to Run It

Frontend on Vercel: https://velodrome-mu.vercel.app

Or locally:

cd frontend && npm install && npm run dev
cd ../backend && npm install && npm run dev

Links