Skip to main content

When generating an automated valuation for a property, how are Confidence Scores and forecast standard deviation values related?

Information about confidence scores and fsd.

Updated today

The forecast standard deviation (also known as “fsd”) for HouseCanary’s AVM is a statistical estimate of model uncertainty. It is a quantity used to create the upper and lower bounds on the value estimate. The value range represents the quantity such that the range will actually capture the subsequent arm's-length sale price approximately 68% of the time (one standard deviation).

The fsd for a property and the Confidence Score are representations of the same number and concept, but are presented using a different format. Confidence Scores are calculated by taking 1 minus the fsd, and then converting that result to a percentage. As an example:

If a property has an fsd of .12, then the Confidence Score is 88%

  • (1 - .12) = 0.88

  • 0.88 * 100 = 88%

HouseCanary also buckets confidence scores into three main buckets:

  • High Confidence - properties that have a confidence score greater than 85.0%

  • Medium Confidence - properties that have a confidence score between 85.0% and 70.0%

  • Low Confidence - properties that have a confidence score less than 70.0%

Both the Confidence Score and the fsd are indicators of how confident (or, inversely, how uncertain) we are about a property’s value based on our AVM model.

Did this answer your question?