Currently capped to 1000, is there a way to increase the number of results returned or is it simply down to chopping up the dataset with restrictBy conditions?
Also, is it normal for the results to return evenly distributed by the specified criteria?
For example, I'm searching for mp3 devices, with restrictBy=price[0,100] to get results with prices ranging from $0 to $100. This range seems to affect the immediate results page rather than the dataset as a whole, so each page I would get products spanning the full range of prices, and the same would apply if I move forward by increasing startIndex. Is it not possibly to mimic Google Shopping's style where you'd have the pagination and the price range would increase the further pages you go along?
Currently the limit for results is 1000, but as you rightly say, you can chop them up with restrictBy.
The price restriction is not exactly as you describe. There appears to be differential treatment of the first N results (where N is around 200) in order that they are as relevant as possible for the search term, so for a query like "ipod" actual ipods would be returned rather than ipod cases etc, which would be cheaper on the scale.
I and others can probably give you a way better explanation on the user group.