AdaBoost decision_function() outputs in binary classification with sklearn

As I understand it based on some study of the source code, I would expect, when using AdaBoost, that values obtained by calling decision_function() would be bounded between -1 and 1. This is because it's the weighed average of the probabilities. However, as you can see in the histogram below, the values seem to range from a little under -2 to a little over +2.

Why is this? Am I under some misunderstanding about how these values are calculated?