And it is only a small fraction of total exchange volume.
According to Bitcoinity (http://bitcoinity.org/markets/list), the total traded in the last 24 hours on all exchanges was 484k. Bitstamp's 25k is a little more than 5% of the total exchange volume. 3600 BTC is only 0.75% of the total exchange volume, and don't forget that 3600 assumes that all mined bitcoins are sold.
Anyway, my point is not that the price is unaffected by newly mined bitcoins. Of course it is. It's basic inflation of the money supply. I just don't think it is the major cause of the steep decline over the last 9 months.
Mining has increased the number of Bitcoins over the last year from about 12 million to about 13.5 million. The increase is about 13%, so from a mathematical standpoint, inflation can account for only a decline in the price by about 11%.
Regarding your 11% comment, I don't think that's how it works. Example: If you were to take 11% of the value of bitcoin's market cap today and dump it on all exchanges at once, it would SURELY drive the price to zero. That's because the size of the cumulative exchange orderbooks is just a fraction of what's sitting in cold storage etc. So for every 1% of coins dumped it will have more than 1% effect on the price, unless done over a loooooooooong period of time such that the order books are thoroughly replenished between sales.
His calculation comes from the assumption that the market cap of bitcoin will stay fixed while the number of bitcoin will increase. Since there will be more bitcoin overall, the price per bitcoin will go down by 11% if the number of bitcoin goes up 13%