PARROTS like a bet, just like humans, reckon scientists.
They can even grasp statistics — thought to be unique to great apes.
The discovery adds to the bird’s legendary intelligence and ability to talk.
Experiments found parrots outperformed monkeys and human infants in basic maths.
Tests on the kea, a parrot from New Zealand, showed they work out odds to choose which hand a token is in — so as to get a food reward.
They combine all the available data including where it is picked and who is handling it.
Lead author Dr Amalia Bastos said: ‘They can predict uncertain future events by filling in the gaps from incomplete information.’
The study, published in Nature Communications, mirrored previous trials in primates and human babies, and the birds did better.
Dr Bastos explained: ‘If you imagine I am placing my hand into a jar with mostly blue sweets and a few yellow and I take something from that jar, but you can’t see what is in my hand, you might guess I have taken a blue one.’
The biologist, an Oxford graduate now based at the University of Auckland, said kea could do the same.
Her team trained six parrots to associate black and orange tokens with a food treat or nothing, respectively.
In a series of tests their numbers were varied in two transparent jars, as a researcher offered them one from each, concealed in a closed fist.
The birds had to opt for a hand by tapping on it. They almost always preferred tokens from jars with a higher frequency of black ones — so they were more likely to win the snack.
Dr Bastos said: ‘They were choosing the hand based on the probability of it containing a black token, something only seen before in great apes.
‘Kea have a complex social structure where many live in a group. They come and go as they please. They need to remember the identities of individuals to interact with them.’
What’s more, kea also preferred a researcher who had previously demonstrated a ‘bias’ towards offering a higher number of black tokens.
Dr Bastos said: ‘We were very surprised to find kea can use social cues, even from humans, to make these judgments.
‘Kea were looking at the biased experimenter understanding they had a preference for a particular type of token, and then selecting this person even when both experimenters had the same populations of tokens.’
Dr Bastos said of her charges: ‘They are great fun to work with. They are some of the most intelligent birds I have ever met and they all have amusing personalities.’