When you enter estimates for features and track the time you spend on implementing them, FogBugz collects these into a historical track record. This is useful for you, as it helps you learn how to estimate features more accurately. It is also an important component of the Evidence-Based Scheduling algorithm (EBS).
To see your Estimate History report, choose Reports | My Estimation History.
You'll see a scatter plot chart showing a diamond for each completed case that had an estimate and time charged to it. The estimated time is shown on the x-axis and the actual time is shown on the y-axis. There's also a line approximating the best fit ratio. As you move your mouse around the chart, popups provide more detailed data. You can click on any point to browse to the case that it represents.
When you move your mouse towards the top right corner of the chart, you'll see tools for zooming and panning the chart.
The estimate history report also shows your 8 best and 8 worst estimates known to FogBugz. This may provide you with some useful information for future estimating.
Occasionally, you will find a case that is such an extreme outlier that it is messing up the schedules produced by EBS.
Normally, you should let those outliers stay in the system. For example, if, on rare occasion, a particular feature takes 10 times longer than estimated, that's not really an outlier: it's real data, which EBS should continue to use in predicting futures. There is a certain probability that some future feature will also take 10 times longer than expected, and EBS needs to know about that.
However, under rare circumstances, you may have outliers that do not represent real data. For example, you might have told FogBugz that you were working on a one-hour feature during the six months that you were on maternity leave, and that incorrect data is severely messing up EBS. In these cases, there is a "discard" icon that lets you throw out any data point, which will prevent it from being used in future EBS calculations.