Object select pass versus render layers

[SonyBoy]

Trying to figure out if there is a way to use object selects to avoid having to create render layers. If I have one object occluding another, I can send both to their own object select passes and get a clean element with alpha for the occluding object. However, the occluded object's render element will have a matte knocking out areas that are occluded by the other object. Is there a way to avoid this? That is, can I send the occluded object to an object select element and have it render as if the occluding object is not there?

In instances where you want, e.g., extreme moblur on the occluding object this would be ideal. I'm wondering if this is fundamentally impossible though, since sample rays cannot "know" what is behind the occluding object, hence normally the need for a separate render pass?

[hermit.crab]

Sorry to say, it cannot be done with the Object Select pass. Though bothersome, render layers seem like a viable solution.

[SonyBoy]

OK, thanks for at least confirming. And am I correct in saying this is simply because of sample rays being unable to "see" occluded objects?

[hermit.crab]

I guess the complexity of the underlying code can be summed up in the words "unable to see", yes.

[SonyBoy]

Thanks for coming down to my level