[squid-users] Why TCP_MISS with simple request/response and aggressive refresh_pattern?

I'm trying to force Squid to cache the results of queries against the
Bing API, which are requests to this URL:

http://api.bing.net/json.aspx?AppId=XXX&Query=hello+world&Sources=Web&Version=2.0&Market=en-us&Adult=moderate&Web.Offset=0&Web.Count=5&Web.Options=DisableHostCollapsing

The responses look like:

HTTP/1.0 200 OK
Content-Length: 4292
Content-Type: application/json; charset=utf-8
X-Akamai-TestID: ce56c7cb038c487183fcc36abc4bd6d9
Date: Fri, 28 Jan 2011 03:03:34 GMT
Connection: keep-alive

{"SearchResponse":{"Version":"2.0","Query":{"SearchTerms":"hello
world"},"Web":...

However, when making requests to that URL with:

http_proxy=http://localhost:3128 wget -q -O/dev/null http://api.bing.net/...

The following entries are added to access.log and store.log,
respectively (I set "debug_options ALL,2"...after trying ,9 I was
quickly overwhelmed, and I'm not sure what the right level is to set
this to):

1296185204.853 96 127.0.0.1 TCP_MISS/200 4606 GET
http://api.bing.net/json.aspx? - DIRECT/173.223.232.104
application/json

1296185204.853 RELEASE -1 FFFFFFFF AC065F42901342C31D5FFDD834A921C9
200 1296185204 -1 -1 application/json 4292/4292 GET
http://api.bing.net/json.aspx?

Spent some time Googling and reading the docs/book/lists, but didn't
find anything. Any hints? Thanks in advance.

-- 
Yang Zhang
http://yz.mit.edu/