BEIJING — For the first few minutes, Nastia Liukin didn't even know the score, unable to see anything beyond the "2" in front of her name.

When she finally did look, she figured she had to be seeing double.

"I turned to my dad and said, 'Dad, we got the same score.' He looked up and said, 'Oh yeah.' We started getting a little confused," she said.

They weren't the only ones.

The Olympic all-around champion finished with the same score as China's He Kexin in Monday night's uneven bars final, but He got the gold medal and Liukin went home with a silver thanks to gymnastics' convoluted tie-break system that sent everyone scrambling for the rulebook. Reading hieroglyphics might be easier than explaining why He won.

He's teammate, Yang Yilin, won the bronze.

"It's not correct. I believe it's correct to have two gold medals," International Gymnastics Federation president Bruno Grandi said. "But this is my modest opinion. The IOC is different."

Gymnastics used to give out duplicate medals at the Olympics. In a bit of irony, Liukin's father, Valeri, got one of his gold medals at the 1988 Olympics after tying teammate Vladimir Artemov on high bar. But the International Olympic Committee told the FIG to stop sharing medals after the Atlanta Games, and a tie-break system was implemented in 1997.

It's a complicated formula that is based on deductions from the execution mark and involves more math than the SAT. Even Liukin wasn't quite sure how the tie was broken — and that was after someone explained it to her.

"I'm not sure if anybody understands what the hell is going on," her father said.

The short answer is that He Kexin had 0.033 less in deductions when you apply the second tie-break formula.

For the long answer, grab a pencil and some scratch paper.

He and Liukin both finished with 16.725. They had identical 7.7 start values (the measure of a routine's difficulty) and they each had a 9.025 for execution after the highest and lowest of the six judges' marks were tossed out. The execution mark is based on the perfect 10 scale, and the first tie-break takes the average of the four deductions that counted. He and Liukin were still tied after that.

For the second tie-break, the three lowest deductions that counted are averaged. When that was done, He had .933 in deductions and Liukin had .966.

Got that?

"I'm a little disappointed I tied," Liukin said. "It wasn't like I got second by three-tenths or five-tenths. I had the same score. That's what makes it a little harder to take."

Not that Liukin will raise a fuss about it.