I would use linear programming to formulate the recipes and product mix. My gut tells me you'd end up losing money having 8 different beers on tap, but that's why you run the numbers to be sure. My gut also tells me you'd lose money making any amount of lager.
I don't see your logic here. Having more beer on tap is likely to increase sales, not keep them flat. People will try different things and come back more often if there is more variety. If it is licensed for it, you can also have guest taps to make up the difference if you can't keep your product on.
There are a lot of variables that go into it obviously, but I would not advise anyone to open a pub with fewer than 8 taps. I would probably suggest 16 or 20, with about half guest taps from other local-ish beers and possibly something like Rolling Rock or Yeungling depending on the expected demographic. You want people to try your beer, but you have to get them in the door first.
You can cut down on the number of fermenters to save $$, but make sure the building and equipment can handle more so you can expand easily. Unless you can get a good deal on used ones of course. Make sure the glycol can handle it, the hoses are long enough, the walk in is big enough, etc.
Obviously there will be differences between markets, so look around at the local places and see what they have on tap and how many taps. Around here it's weird if a place has BMC on tap, but in other places that's all you can find. You've got to know the demo. If most places don't serve craft beer they you'll need to spend a lot of time educating your customers and it will probably be best to have some familiar beers on tap for them.