Welcome to the Q&A Forum

Sometimes they do not choose to utilize the meta description and titles I believe because they may think something is more relevant to return based on the specific user query.

As far as I know, you cannot do anything about what gets displayed by the engine.

Your changes may appear as intended for some searches and not for others is what I am really saying.  It is the engine's prerogative as to what is displayed based on the user's query.

Drop them like they are hot.

1. You are correct, generally, Google will index when it makes sense for them to do so per their algorithms.  It is not in my experience driven by the time on the XML sitemap.  It could be loosely correlated, but not huge.

2. You should setup up 301 redirects on the auctions that have ended.  Don't 404 them, that is bad.  Redirect them to a page for new related auctions (probably will require some code to be whipped up).

3. You could try to utilize webmaster tools Fetch as Google to get the pages indexed, however, this is manual, not guaranteed that Google will index it and you only have so many requests available.

4. Spend some more time on getting organic links via articles and other content in the niche.  This will help on a number of levels not only in ranking but also attracting traffic from people interested in the product/market.

You can remove them via GWT as Chris has so aptly mentioned.  Best way to do IMO.

In a lot of respects, it measure the efficiency of your site.  How well the code is optimized as well as your server in processing that code.  It is possible to have a poorly coded site which performs OK because you compensate for the bad code/bad layout by having more horsepower.  The server is definitely correlated to the pagespeed rank.  It's likened to a big differential equation with 100 factors all playing a role in the final output, but several of the items contributing to the majority of the output behavior.

Check out some of the documentation on it: https://developers.google.com/speed/docs/insights/about

The raw horsepower of the server is only going to get you so far.  You would be much better off designing a system with a small server which is super efficient, than spend tons of money on the top of the line server which is using all of its power to process inefficient code, assets, etc.

I think this white board Friday may help you out a little

http://moz.com/blog/how-google-knows-what-sites-you-control-and-why-it-matters-whiteboard-friday

I would just start creating fresh content on the old site and wait for natural organic links to grow.  The traffic will come back over time.

So what you're really saying is that your click percentage is double what it was before...

I would only take the impression count with a grain of salt.  Also, take a look at all your analytics sources and look at ratios rather than just the raw #s; you will be able to get some more actionable data from ratio analysis.  Also, personally, and again this is a preference and some may like it or hate it, but I like statcounter.com for some analytics.  I have also used Kissmetrics which depending on what your business model is, can also be good (also shout out to CrazyEgg).  Worry much more about how many unique visitors are going to the site and your conversion rate (for whatever product it is).

If you focus on efficiency, you won't have to put as much input (unique visitors) in the top of the funnel.  Of course, this is easy for me to say because I have no practical advice on how to make your machine more efficient, but just that it is a good thing to focus on.

Yep, in terms of Google picking it up in the SERPs, they are only going to display 1 author (if appropriate).

So, it depends on whether or not that is your intended goal.

IMO, one site is enough to manage all by itself.

Having multiple sites, just from an administration perspective is very challenging in particular when you go over a certain amount.  You would have to have full time admins just to manage the sites, let alone the content of each site.

You would have to add up how much expense you have per unit of site versus how much net revenue (profit) you have directly from that site.  I think in a lot of cases the expenses per site are going to exceed the profits, but before making a sweeping generalization, do a study and make your best guesses before investing in the infrastructure, IT and employees.

The password protected stuff IMO won't help you SEO wise because it will likely be blocked to robots/search engines.

If it cannot be indexed, how will the search engines know it is there?

The best answer is, it depends.  What does your analytics tell you where the traffic is coming from and going to on the sites?