MHB Determining asset value using historical stats

[jcbn82]

Hey guys, nice to meet you all! My name is Jonathan and I'm from Queensland Australia.

I'm currently working on a project which is seeking to find the most accurate method to value an asset based on historical stats. I think the best way to describe would be to provide a theoretical example:

I am a widget investor, and every day there are thousands of widgets bought and sold in any particular marketplace. As a value investor I am looking to purchase only those widgets that are available for less than their true value. Widgets are a desirable commodity and generally increase in value over time.

To determine a formula for true value (or as close as possible), I am able to access the following data for each widget that is for sale and those recently sold:

- last sale amount
- days since manufacture
- days since last sale
- number of arms
- number of motors
- number of wheels
- colour
- brand of widget
- general condition of widget (1 to 10)

Using this data, I am hoping to be able to create a formula that will give me a conservative estimate of value. I am not sure how to best set the coefficients and each variable, and to be honest am not sure where to start.

Any pointers in the right direction would be greatly appreciated. Thanks!

 

[Ackbach]

I'm not much up on this method, but it seems to me that multiple linear regression might be useful to you. Jameson and I like Serena are excellent at statistics - they might be able to help you out more.