How is the Case-Shiller index calculated anyway? If they take same-house sales, would they take a May 2000 sale at $350K, and then look at the most recent sale in October 2006 at $600K, and figure that house gained in price? How do they figure out the peak, and subtract the price of the October 2006 house from the actual peak?

Interesting question. It’s an index, so they can use statistics over whatever was sold in a period (quarter), but how do they normalize/relate the data so that they have meaning relative to the mix of sales in other quarters ?

At least we understand the median and can try to understand it’s flaws and sometimes counter-intuitive trends (e.g. market spiking because high end sales continue, while low-end sales evaporate as the market started to soften).