SIZING up his shot, Dengji Zhao's unblinking eyes move across the landscape. Behind him, his opponent emits a calculating hum.
Fifteen years after IBM's Deep Blue computer beat Garry Kasparov at chess, man and machine pitted their wits against each other again in Sydney on Wednesday. This time over Angry Birds.
''In chess the game is clearly defined, in Angry Birds you have no idea what will happen,'' said the contest organiser, head of the ANU's school of artificial intelligence, Jochen Renz.
Number cruncher … Dengji Zhao had several strategies to beat the computer. Photo: Janie Barrett
Teams had five weeks to design an artificial intelligence program to compete against humans in the wildly popular game, in which players launch birds into green pigs, often protected by obstacles.
There were 16 contestants, from as far away as Turkey and Brazil. The challenge proved more difficult than many expected.
One of the chief obstacles was getting computers to recognise pictures, while computing such a wide array of possible moves: ''If you make a shot that is only ten thousandths of a degree different, you get a completely different result''.
Each program had its own strategy. Some aimed directly at the pigs. Others ricocheted off obstacles.
Felix Andrews, a data miner, spent a month building a computer model that simulated possible paths for birds to take and chose the best among them.
''It used some heuristic guessing techniques,'' he said. But in the end it wasn't enough. ''My simulation didn't match the world of Angry Birds.''
Even the most sophisticated models tripped up. One was stuck on level 8 for most of an hour. ''They couldn't look, it was just too painful,'' Associate Professor Renz said.
But in the end, human intelligence triumphed. Zhao, playing Angry Birds for his first time, scored 565,000 points - 35,000 more than the leading computer.
''I had some strategies,'' he said. ''But sometimes good luck can give you something else''.
Associate Professor Renz believes that as such contests become more popular, computers will become more sophisticated - and successful. ''This was probably the last chance for humans to win,'' he said.