EDIT2: I made a last-minute change that was a mistake - % Change from 3 months prior can be found by subtracting the current month's C-S number from last month's number and multiplying by 3. It is still inaccurate, because it averages the current month with the previous 2. The sum of each group of 4 should be 0. The next attempt involves averaging the 3 months and, then removing the surrounding months, then subtracting the current month from last month and dividing by 3. This appears to be accurate for comparing the current month with 3 months prior. I had hoped that row 302 would accurately show monthly data, but instead it seems to still contain influence from previous months. EDIT: Sorry, I posted the wrong Redfin link earlier.I really love Redfin - their website and their field agents are great. But, they're trying to convince sellers that now (winter) is the best time to sell.
I crunched the overall market San Francisco Case-Shiller numbers and came up with the included image. I averaged the months between 1987 and 2010 and then did some other analysis on it. I tried to upload the Excel file to Google Docs, but it complained about some of the formulas. I'm amazed, as a potential first-time home buyer, how much seasonal ebb and flow the market has. What do you think? Can anybody think of a way to accurately reverse-engineer single-month averages from the 3 month moving average?