The method

How Whole Board measures intelligence

Most online IQ tests ask 20 pattern puzzles and hand you one number. Real intelligence tests, like the WAIS used by psychologists, measure several separate abilities and combine them. Whole Board works the same way.

Intelligence has more than one part

The most widely accepted model of intelligence today is called CHC theory (named after the researchers Cattell, Horn and Carroll). It splits intelligence into broad abilities. Someone can be a fast visual thinker with an average vocabulary, or the other way around. Most major modern IQ tests, such as the Woodcock-Johnson, were built on or updated to fit this model.

Each Whole Board category measures one of these abilities. Researchers use short codes for these abilities. Here's what each one means (EF, executive function, comes from cognitive psychology rather than CHC):

How your scores are calculated

1. Category scores

Your raw result in each category (for example, 6 out of 8 grids solved) is compared to how people typically score. It's converted to the standard IQ scale, where 100 is exactly average and every 15 points is one standard deviation. A 115 means you scored higher than about 84% of people in that ability.

2. Your overall IQ

Once you've completed at least 3 categories, we combine your best score in each into one overall IQ. We don't just average them. Abilities are related to each other, so averaging would squeeze everyone toward 100. Instead we use a standard statistical correction for combining related scores, which keeps the overall score on the IQ scale.

Your overall score comes with an estimated 90% range. That's the band your true score most likely falls in. It gets narrower with every category you add, so finishing all eight gives you the most precise result.

3. Calibration

When a category launches, its scores start from estimates based on research into how adults perform on similar tasks. As real people take each test, we recalibrate using their first attempts only, so practice from retakes doesn't distort what "average" means.

War of Minds: leaderboards and retakes

• Boards at every level. Players rank on their home board: neighborhood, city, state and the whole US. Places rank too: neighborhoods inside a city, cities inside a state, states inside the country. A place's score is the average of its 10 best verified players, and it needs 5 to appear.
• Age groups (18–24, 25–34, 35–44, 45–54, 55+) get their own boards at every level.
• Verified and 18+. Ranked boards need a one-time ID check. Under 18 or outside the US, you can still take every test and play private matches.
• Three time frames: this week, this month and the season (January 1 to December 31). Each one is live until it ends. Then its top 3 players and places go into the Hall of Fame for good, and it starts fresh.
• Every category has its own board, ranked by each player's best ranked score this season.
• Ranked attempts are free: your first try, then one free retake per category every 7 days. Everyone gets the same number of shots, so the leaderboard can't be bought.
• The overall leaderboard ranks everyone who has completed at least 3 categories by overall IQ.
• The Crowning. When the season ends on December 31, the #1 state, each state's #1 city and each city's #1 neighborhood are crowned. Each crowned territory gets a letter from us. Verified players who live there, and every player on a season podium, get one congratulation email on January 1.
• The 125 line. Verified players who reach an overall IQ of 125 or higher get a private email invite to what we're building next (you can turn invites off any time).
• Private matches (1 v 1, 2 v 2, 1 v 2, 1 v 3) give friends and family the exact same questions. Teams score the average of their players. Matches never affect rank.
• Want more practice? Practice retakes are $0.99 and can be taken any time. They update your history, trend chart and practice best, but not your rank.

Fair play: human-only

• Fresh questions every attempt. Pattern, Number, Spatial and Logic puzzles are generated from rules when you press start, so there's no answer key to look up. Word Power draws a new mix from a rotating bank.
• Server-side timers. The clock runs on our server from the moment you press Begin, with tight limits per question.
• AI-resistant. Questions are drawn as images, leaving the tab is detected, and automated browsers are flagged.
• Scored on our server. Your browser only sends what you picked. Scores can't be edited.

Read the full Fair Play rules and why they matter for building a new human baseline in the age of AI.

How accurate is it?

Whole Board uses the same kinds of tasks as professional tests and standard scoring methods, and it gets more accurate as more people take it. Like every online test, it can't control your environment: sleep, distractions and screen size all affect results. For an official diagnosis used in school or medical decisions, you'd still need a test given in person by a licensed psychologist.

Expert review: where we're headed

We want Whole Board held to the standard professionals use, not just our own word. Our plan, in order:

1. Publish the method. How every category works and how scores are calculated (this page).
2. Build real norms. Recalibrate "average" from real players' first attempts.
3. Independent expert review. Invite licensed psychologists and psychometricians (the specialists who design and validate tests) to review our questions, scoring and fairness, independently from us.
4. Publish the numbers. Share reliability and validity results, including what didn't hold up, and fix it.

Today, Whole Board has not yet been reviewed by licensed psychologists. When it is, we'll say who reviewed it and what they found. Until then, it's a well-built estimate, not a clinical assessment.