William Wolf shows us the value of filtered statistics:
Wolf only had 700 complaints, but 166,900 records were estimated for return. He is looking much worse than reality shows.
So, what is happening is that there are 3 possible employee results for complaints. It is rather simple. CE is taking the total amount of records(500,701) and dividing by 3 assuming that all 3 will have roughly the same amount of records. We see that along with the estimated number of records being the same, the execution plan operators are the same. For such a variation in amount of records, there must be a better way.
I rarely create filtered statistics, in part because I don’t have a good idea of exactly which values people will use when searching. But one slight change to Wolf’s scenario might help: having a filter where name = Sunshine and a filter where name <> Sunshine (or name is null). That might help a case where there’s extreme skew with one value and the rest are much closer to uniformly distributed.