Checkpoint and Background Writer Statistics

We're trying to pull a few various metrics from the postgres catalogs to analyze database performance, and included is buffers_checkpoint and buffers_clean from pg_stat_bgwriter.

One of our clients has an lru_max_dirty setting of 1000 and a bgwriter_delay of 200, but we're still seeing much higher growth in the buffers_checkpoint than buffers_clean, with buffers_checkpoint increasing at about 20 times the rate that buffers_clean is.  We don't see very much growth in the maxwritten_clean, though - something like one every couple of days.

So the questions is, why are there still so many more checkpoint buffers than clean buffers written?  Does the buffers_checkpoint value include more than we think?  If not, can we tune things further?