0

Given a sample of items with known monetary values, say 10 cars, and a person rating each car in turn as being more or less valuable than their own, what would be the best way of estimating the value of that person's car?

The rating that they choose could be more complicated than just less/more valuable - it could be, for e.g., 'Much less', 'Less', 'Similar', 'More', 'Much more'. Whatever necessary within reason to get an accurate estimate...

Blob45
  • 11
  • 1
    This problem seems somewhat ill-defined, because analzing it quantitatively (or even qualitatively in most cases) would require that we know what value differential each person assigns to each of the possible ratings. So we would need the ratings to be numerical values (e.g., "$10,000 less valuable than my car"). Even then, we could estimate the comparative values but not the absolute values. – CloudyGloudy Jul 25 '17 at 17:09
  • Thanks. I agree that it could possibly be more accurate if a numerical rating were given, but in this particular scenario the answers need to be standardised. – Blob45 Jul 26 '17 at 07:54
  • That doesn't really make sense though...there's nothing "standardised" about it if the ratings are completely subjective -- that would make it impossible to find an answer. – CloudyGloudy Jul 26 '17 at 22:14
  • Agree it's impossible to find the answer, but what's needed is a model to find the best answer. It's ultimately subjective, but the aim is to approximate the most accurate valuation from the available data. – Blob45 Jul 27 '17 at 08:55

0 Answers0