Anyone Know Standard Error?

My data is currently following a 79-day Linear Regression Trendline. That is, the R-squared value is .96, so there is a very high correlaton between the data and the trendline. My question is this: How do I test the reliability of 1 and 2 Standard Errors above and below my data? I know that 1 Standard Deviation above and below the trendline should contain about 68% of my data and 2 Standard Deviations should contain 95%, but what about Standard Error Levels? Five days ago, when the data was following a 74-day Linear Regression Trendline, my data reversed quickly when it was 2 Standard Errors above the trendline--I want to know why.