@markr,
I feel as though the argument shifted during the discussion. Here was the first report:
______________
"The good news was that overall, we picked a wine’s price point (low, medium or high) in slightly more than half of all tastings (53.5% of all answers correct).
However, as a group we were not statistically significantly better than random under several different statistical models – in local parlance that means: if we sat and happily drank all the wine, didn’t give a hoot about the scores, and ask Ranger and Ringo [OUR DOGS] to scribble “a”, “b” and “c” in the boxes, the results would have been indistinguishable from ours (assuming their writing was legible, and that they hadn’t been sneaking tastes when we weren’t looking)."
_______________
When I disagreed with this, saying that random selection should only produce 33% correct, there was this response:
_______________
"Given a series of "independent" tastings, the whole group including the canines, have a 2/3 chance of getting at least one wine correct in a sample of three. So, with 14 of us hard at work, at least 9 of us should have got 1 wine correct in each flight. Sadly for even the proudest wine tasters, we did statistically no better than that in any of the four flights."
________________
Judging in that manner, not enough weight is given when someone correctly identified all 3 wines. I'd be interested in your comments.