p-value for panelists correctly identifying both sets

We've been running the attached sensory analysis test at work for the past few years and using the attached "Chart 1" to determine the p-value. I would like to walk through the math that created the chart, but I'm not sure where to start.
For example, I tried to work out the case that 6 out of 10 panelists correctly identify both sets. According to the chart, the p-value would be 0.0197, or highly significant. This is as far as I was able to get:
Expected value = 0.25 (Probability of correctly guessing both sets)
Observed value = 0.60 (actual value in this example)
chi-squared value = 0.49
Significance level = 5% (assumed)

Plugging this into a chi square distribution chart, I'm getting a p-value between .1 and .9, which is completely different from "Chart 1". Where am I going wrong?