when deciding which arbitration or potential FA to keep, what are the most important stats to pay attention to?

Don't look at stats. Look at ratings. The first thing I look at is vL and vR ratings for both pitchers and position players. Then I look at the batting eye and contact ratings for position players. I look at the pitch ratings for the pitchers. Look at the makeup of the player, that will tell you how likely they are going to continue developing and/or sustain their ratings as they age.

Agreed, though I wouldn't ignore the stats necessarily. If you have run the team for a couple years and players have underachieved what you would expect them to do every year (without a good reason such as poor defense or catching causing a pitcher to do worse), it is safe to assume that they will continue to under perform your expectations. That doesn't mean you don't necessarily sign them, but it may affect how much you're willing to pay to hope they suddenly get better.

what are the target numbers to "expect" good results?

70? 80?

Philogenemay is referring to a world in its first season, so there are no stats to reference.

is a great defensive CF with no bat, worth 2.5 mil?

My guess is you are talking about Jerome Pressley and I'd say no.

I would agree with trop on the specific point. In general though, when looking at a player with high defense, vs. low offense, I'd suggest trying to quantify what you give up vs. what you get. You can look at any league's players under Report > Player Search to see what stats players with ratings like yours have generated.

If you are talking about Pressley, I'd say he'll prevent a few more runs than the typical CF, but his offense will create MUCH fewer than you should tolerate.

Your SS on the other hand should be gold glove caliber, and likely more than offset with his glove, what he'll lack in his bat.

I think it's all situational. I have a team where my SS and CFer aren't good hitters, but both are top quality SS level defensively. I play them because I have plenty of offense even with 2 dead spots in the lineup. I am .080 ahead of any other team in the league in OPS, and having those guys in the lineup makes me feel better about my defense even if I have no statistical proof it makes me a better team.

Having a quality defensive player at a key position or two might be worth it depending on the makeup of the rest of the team.

Stick around awhile and you'll see a lot of threads like this. "Is my guy's defense good enough to makeup for a weak bat" You'll get lots of varied opinions. Bottom line is you need to win games, whether you win 12-10 or 3-1 doesn't matter.

So, when considering 2 starters for a lineup spot, estimate how many runs they will create and how many runs their defense will prevent. Focus on the net and you'll probably have a leg up on your competition.

Exactly. No one can tell you what you're comfortable with on your team.

Exactly. That's why I said it's situational. I don't even do the same thing with all of my teams...it depends on what options I have.

2.5M in concept for a pure defensive specialist (at CF, Id want that to be 90+ glove and range), with really poor bat, is a little high. If I was trying to really put emphasis on ML level payroll and wanted a great defensive backup, then it could work, though for your first go around, the 2.5 can be better put elsewhere.

I will say this: I had a very poor defensive team for several seasons and, after finally focusing on key defensive positions, it's as if I'd had a religious awakening. I just stick Gold Glove players into the SS and 2B spots and make sure I have good guys at CF and 3B and I have easily the best team ERA in the league (improving my pitching also helped, but it seems like several of my pitchers clearly overachieve due to my stellar defense. All 5 of my starters have ERAs under 4 playing in a neutral park right now).

Like people said above, everyone has preferences when running a team, but now I couldn't imagine living without my defense.