Standardised scores are typically used to allow comparisons across different data sets (or across different time periods). They do this by re-basing sets of raw data on the basis of their averages (means) and standard deviations.
For example, let's assume pupils took 2 test papers, one with a maximum possible 175 marks, the other out of, say, 95 marks. How do we compare overall performance? Well, we could easily compute each individual's percentage. But, if someone scores 75% on each paper, does that mean the relative performance is identical? To answer that, standardisation looks at the distribution of all the data (i.e., takes account of all scores available).
Standard deviations are measures of the distribution of the data. So, if the data is normally distributed, about 68% of the observations would fall within +/- one standard deviation of the mean. (About 95% of all the data would be within +/- 2 standard deviations.) Therefore, for any score, we can work out how far it is from the mean in terms of standard deviations. Armed with this, we now have a better basis for comparison. On paper 1, 75% may place the pupil in the top 10% of all pupils whilst, on the second paper, 75% may place them only in the top 25%. (Indeed, from the actual score, we can determine the exact percentile of any mark.)
From your example above, it looks like the mean is 100 (normal to rebase the mean to 100); the standard deviation is 15 (giving a range from 85 to 115 = 100-15 & 100 + 15). Below 85 places the data more than one standard deviation below the mean (basically bottom 20%), greater than 115 places the data more than one standard deviation above the mean (top 20%).
Hope this helps!