One other option you might want to add is being able to toggle blend probes in the manager.  I had a bare scene that worked great but my game scene was still seeing high draw calls/batches.  It stemmed from having a light probe group.

ya that definitely fixes the draw call/batching numbers.  Honestly I can't really tell before from after, you probably know more about what to look for specifically.  It looks fine to me.

Before:

After:

hmm ok so maybe I spoke too soon, or maybe my system is just high end enough where it's hard to notice the performance difference.

So I don't think there is any actual performance difference between good/bad above.  Looking closer I see that in forward it's showing the draw calls as batched into one batch.  In deferred it doesn't show them as batched draw calls, but I don't see any difference in framerate or in the CPU profiler between the two. I think this is just a quirk of how the profiler shows stuff.  There might be some actual differences, but they don't appear to be differences that effect anything.  At least not on PC.

So good here means setpass = draw calls = batches.

Bad means setpass is low but draw cals and batches are high (and both the same)

manager forward cam deferred = good

manager deferred cam deferred = bad

manager deferred cam forward = good

manager forward cam forward = good

I'm getting my info from the profiler. You know it could be a quirk of the profiler.  I know setpass calls are supposedly draw calls now, but the profiler is not consistent there if that's completely accurate.  Because in forward setpass = drawcalls but in deferred that is not the case.

I'm assuming it's fixable so I'll just keep an eye out.  When it's working correctly it's great, just what I was looking for.

strange.  So if I set billboard setting to billboards only, something goes wrong and it shows both, shoots draw calls through the roof it hit 11k lol.

The regular Gaia Demo does the same thing for me.  Switch cam/manager to deferred and draw calls jump.

oh ya unity 2017.1.0p2