Most weekends, somewhere around fifty competitions run worldwide, and nearly every one of them ends the same way: a 3×3 final, five solves per person, and a winner. One question, asked of all 15,251 finals ever held: how fast do you have to be to win one?
First, the measure. A final is decided on an average of five: each finalist does five timed solves, the best and worst are discarded, and the three in the middle are averaged. The chart above takes every competition final on earth, year by year since 2004, and finds the median winning average: the mark that beat half of that year's winners. Not the World Championship, not the record solves that make the news. The ordinary Saturday, in the middle of the pack.
In 2008 that number was 13.56 seconds, across 109 finals. A club solver averaging fourteen seconds today would have walked into half the trophies on earth. By 2015 the median had fallen to 10.01. In 2025 it was 7.44, across 2,170 finals, and 7.32 is where this year's figure currently sits. The chart above is the whole story in one shape: to win the same local competition, you must now be almost exactly twice as fast as in 2008, and there are twenty times as many finals to win.
How completely has the old standard been overtaken? Of the 2,170 finals held in 2025, exactly three were won with an average slower than the 2008 median. The typical champion of 2008, frozen in time and dropped into last year, would have won three competitions out of 2,170.
of 2025's 2,170 finals were won at a pace that would have been an ordinary winning time in 2008.
At the sharp end the standard has reached a level that borders on cruelty. The fastest average of five ever produced by somebody who still lost the final is 3.85 seconds, by Yiheng Wang (王艺衡) at Weihai Open 2026. Until 10 January 2026, that average was faster than the world record itself. On the day, it was worth second place: Xuanyi Geng (耿暄一) won the final with 3.74.
It is honest to add what the median hides. The slowest winning average of 2025 was 17.8 seconds and the fastest was 4.03, which means the spread between one Saturday and another spans a factor of four. Small competitions in new cubing countries are still won at a pace a dedicated beginner can reach in months. The door into competitive cubing has not closed; what has changed is how far behind it the room now extends.
One footnote in the data, because the series is otherwise relentless: the median fell every single year from 2004 to 2025 with one exception, a rise of nine hundredths from 2021 to 2022. That is not a pause in the sport's progress. It is an echo of the pandemic, when the few competitions still running were concentrated in the fastest scenes on earth, a story told in full elsewhere in this edition.
The decline shows no sign of a floor. The 2026 figure is on course to be the lowest yet, and every year of finals so far has asked more of its winners than the year before. Whoever wins the median Saturday of 2035 will make 7.44 look the way 13.56 looks now.
Median of winning averages across all 3×3 competition finals per calendar year, WCA results export of 6 September 2026. Recomputed nightly.