The problem is using the zscore or the number of standard deviations away from the mean as a measure of the performance.
First the history test, the z score or z=(x-mu)/sigma=(80-70)/10 = 1. This means that pauls score was exactly 1 standard deviation above the mean or approximately 84th percentile.
For the biol test the same procedure gives z=(75-70)/2.5 or 2. This is much better performance and the approxmiate percetile is 98.
The key point of this exercise is to know that if you have a normal distribution, then the z value (as computed above) give the position on the normal distribution which in turn gives the percentage of the population that is below this particular value