Higher or Lower, played with Wikipedia pageviews

Two Wikipedia articles appear side by side. One shows how many times it was read last month. You guess whether the other was read more or less. Keep guessing until you get one wrong.

Why Wikipedia pageviews and not Google searches

Most higher or lower games use Google search volume. This one uses Wikipedia pageviews — the number of times people actually opened and read an article. The two measure different things, and they disagree more often than you would expect.

Plenty of topics get searched constantly but rarely read about in depth: products, brands, sports scores, anything people look up to get a quick answer. Other topics are searched rarely but read thoroughly when someone does arrive — historical events, scientific concepts, obscure biographies. Wikipedia pageviews capture curiosity and attention rather than commercial intent, so intuitions built on search-volume games tend to misfire here.

Where the numbers come from

Every figure comes from the Wikimedia Pageviews API, which publishes per-article view counts for every language edition of Wikipedia. The game uses English Wikipedia, counts only human visits rather than bots or crawlers, and covers one complete calendar month.

The data refreshes automatically on the third of each month. That means the answers change regularly — a streak built on memorising last month's numbers will not survive the update.

Which articles come up

Roughly ten thousand articles, drawn from Wikipedia's Vital articles lists. These are the subjects Wikipedia editors consider most essential to the encyclopedia, maintained collaboratively over many years. Using them instead of genuinely random articles avoids the problem that most of Wikipedia's seven million pages are short entries about villages, insect species and minor asteroids — technically random, but impossible to guess about and no fun at all.

How the difficulty works

• Early rounds pair two widely recognised subjects, with a large gap between their view counts.
• As your streak grows, the gap narrows and less familiar articles enter the pool.
• Pairs whose numbers are too close are never shown, so a loss is always a real mistake rather than a coin flip.
• No article repeats within a single run.

Tips

Recency beats fame. A long-dead composer usually loses to a moderately popular television series. Broad concepts like war or music tend to underperform against specific entities, because people search for the specific thing. Countries do well. Scientific concepts that appear in school curricula do far better than their cultural profile suggests. And anything that was in the news during the month in question can spike enormously.