🧭 How it works
PriceGuess is an educational price-guessing game across gold, stocks, crypto, real estate, FX, and commodities. This page explains how puzzles are generated, how scoring works, and what limitations to keep in mind — so there is no mystery about what you see on screen.
1. How puzzles are assembled
Each mode draws from its own dataset. Three principles are shared across all of them.
- •Snapshot based: Numbers are captured at build time from public records or representative values — not a live market feed. Expect a lag vs. live quotes.
- •Public sources: We reference public sources like Kitco, CoinGecko, Yahoo Finance, the Bank of Korea, and the EIA. No proprietary feeds are used.
- •Educational framing: Rather than raw numbers, each item pairs the price with a short historical or contextual note so players learn the surrounding story.
2. Per-mode flow
A single numeric puzzle rotates every midnight (KST). You have six tries and must land within 5% error.
Left value shown, right hidden — pick higher or lower. One miss ends your streak.
Stack correct higher/lower calls inside a 60-second clock using the same pool.
Identify a ticker or predict direction from a demo chart simplified from real historical paths.
Pick 3 assets against an AI's picks and compare their recorded 1-year returns — no real trade takes place.
Compare USD-converted Big Mac prices between two countries and learn PPP intuition.
3. How scoring works
All scores, streaks, and stats stay in your browser's localStorage. Nothing is sent to a server and there are no accounts.
- •Numeric modes (Daily, CEO Salary, Oil, Crypto, Tuition, Flight, …): within ±5% counts as correct.
- •Higher/lower modes (H/L, Time Attack, Big Mac): exact-match selection.
- •Chart Quiz (direction mode): pick the correct ticker or up/down direction.
- •Price Sense Battle: compare weighted return from your mix against the AI's using the stored 1-year return figures.
4. Ads vs. content
Ads shown via Google AdSense have zero influence on answers, scoring, which assets appear, or when data is refreshed. No advertiser can cause a particular asset to show up more often or a hint to be hidden.
5. Known limitations (important)
- •Not a live quote. Anything after the build date is not reflected.
- •Real estate, salary, shopping, and flight items are representative examples — not specific listings or personal figures.
- •Collection, rounding, and unit conversion can introduce roughly ±1–5% error — corrected when confirmed.