Migrating to V1.1 how to best use since_id and max_id to replace count for tweets/search


I was extremely happy with the results of the old API in terms of content returned via the count system. I am now migrating over to 1.1 and facing a number of issues. I am looking to request up to 500 tweets for anaylsis (not at the same time, 100 per request). Problem is with since_id and max_id I feel like I have lost control over the tweets im receiving.

My first approach is to get tweets with since_id left as blank but then after each iteration set since_id to last ID in a the result field. However this seems to be yielding significantly less tweets before and more worryingly I am getting an extremely high number of duplicate tweets from users who iv already received data from.

If your simply trying to get a pool of the most recent 500(or N) tweets how would you best use these params to get this sort of result?