Build-In IPS policies: Differences?

Essentially: The first one are locked in and managed by Sophos. 
The others are free to design: They have ALL Rules included, but can be modified if you want to. 

Or you build your own rule set. By clicking on the other policies, you see what Pattern are loaded. 

I used the Sophos managed policies for years and never looked deeper into them.

Why are there different sub-settings like "browsers", "OS", "windows", "linux" when the last policy "All Clients" would catch any of the above?
Is that because of some kind of "speed optimizations"?

The main difference between LAN and DMZ is client vs. server.
Would it be a very bad idea to create a new IPS policy where no filters are configured at all?
Like this:

Hi,

the sub settings allow you to create a smaller footprint in memory by removing items you do not have on your network. The utm used to save large amounts of memory with fine tuning, it does not seem to have the same affect on the XG.

why would you create an ips policy with no entries?

ian

Hi,

the sub settings allow you to create a smaller footprint in memory by removing items you do not have on your network. The utm used to save large amounts of memory with fine tuning, it does not seem to have the same affect on the XG.

why would you create an ips policy with no entries?

ian

No, not with no entries, but only with one.

Like LAN to WAN and DMZ to WAN which at the end (last rule) include any for clients (L2W) or any for servers (D2W).
My goal would be a mixture of both policies, so not limiting the ruleset to clients or servers.

On UTM the IPS configuration (OS, applications, timeframe etc.) had an impact on the sum of rules that will apply.
I wonder if that is no issue on the Sophos Firewall, since the template rules automatically include everything.