The use case would be as follows:

 Let’s say I have two hosts called host1 and host2.  I create two policies with a precedence of 100 but the first policy has an agent selection criteria of “Agent Host Name starts with host1”.  The second policy has an agent selection criteria of “Agent Host Name starts with host2”.  Suppose that I have 600 host specific policies and I used the above strategy where they have the same precedence but different agent selection criteria.  Is there any technical or performance consideration for adopting this strategy?  Does this slow anything down or cause any problems? 
 

Generally speaking we have adopted a top down strategy.  Policies with the largest numeric precedence establish baselines and default configurations.  Policies with the lowest numeric precedence override those more general policies.  Host specific policies are overrides where we only put information that is specific to a particular host that is different from the baseline.  I just want to know if you have any knowledge about stacking many policies on the same precedence create a performance or behavioral issue.

There would not be any performance concerns. Precedence (for the most part) is an Agent side impact.
As an example of this, BMC has tested 20k policies against 20k agents (think 1 agent – 1 policy). In that test, we used the same precedence # for each policy and did not encounter any performance or policy application issues.