Request RELAR Demo
Loading...
Request sample report. Request sample report.


Relar Confidence Report

This document provides an overview of the RELAR confidence score. It is intended to provide a non-technical description of the meaning and basis for the calculation of the score.
Generally, a RELAR report consists of a price prediction, a price range prediction, a time to sell prediction, and a time to sell range prediction. While many users focus on the price prediction number, in understanding the confidence score, it is important to remember that RELAR predicts both price and time to sell and expected ranges for both numbers. If the property does not achieve a score of 50 or better, there is no charge for the report.

First, consider an idealized scenario where there were 100 homes available for sale, all of which were completely identical in every respect. If all the homes are priced the same, the probability that any one home would sell to an interested buyer is 1%. The time that it would take, on average, to sell a house, is governed by the average rate at which buyers arrive to buy the houses. Eventually, all of the houses would sell, assuming that no other identical houses came on the market.

e, there is no such scenario in the real world. Instead, what we try to do is to create such a scenario by finding similar properties and then adjusting their sales prices based on how different they are from the property we are interested in. This produces a situation which differs from the ideal because the adjustments are imperfect and because the exact characteristics of the property may not all be known. In particular, where the properties differ from each other, we must try to infer what the adjustments should be from the behavior we see in the marketplace. The inference process leads to imperfections, or ranges, in the price adjustments. There are unknowns in the process, which are properly dealt with probabilistically.

As the price and sales time of each property is adjusted to match the subject, the price and time to sell of each property become, not numbers, but regions of probability. The size of each region depends on how well we know the physical and geographical characteristics of both the subject and the comp. Some of the factors which affect the adjustment are:

  • Gross Living Area
  • Bedroom/Bathroom/Room/Garage/Fireplace Count
  • Land or site area
  • Year built
  • Local Street Density/Local rooftop to rooftop distance
  • Distance to geographic features, including water, parkland, major streets, freeways, commercial development, airports, hospitals, etc.

In order to adjust the price of each property, we must know the relative importance of each factor to the others and the relative importance of each factor in the price of houses in the local market. These can be inferred from the data itself, but, as stated before, the inference process is not exact. The result must be dealt with probabilistically.

Once this is complete, we can calculate the expected ranges of price and time for the subject property. We can also then calculate a confidence score based on the probabilities that arise from our adjustment process and the uncertainties in it.

The confidence score represents the probability that the subject property will sell within the calculated price and time range. Thus, if the price range were $350,000 - $375,000 and the time to sell range were 75 days to 105 days, the confidence score is the probability that the subject property will sell within these ranges based on our ability to approximate the idealized scenario described in paragraph 3 above. If we calculated the 100 houses in that scenario and took the predicted price and time to sell for each house, and we had a confidence score of 80, we would expect that 80 of the 100 houses would fall within the given price and time to sell ranges.

The table below represents the low, medium & high confidence translations.


0-50 Unreliable We provide the report no charge and deem the report un-reliable
50-65 Low Confidence  
65-80 Medium Confidence  
80-100 High Confidence