They doesn't matter.
Not really anyway. For example :
Best hitter on the Nats last Spring? Matt Skole. Also very good in the Spring. Tony Gwynn Jr, Wilson Ramos.
Best pitchers on the Nats last Spring? Matt Grace, Blake Treinen.
Yes, I cherry picked a little - you could toss in Clint Robinson, Bryce and Max in there. But that's doesn't kill the point. How well you do in a limited time frame against a large variation of competition produces results that are pretty much random. That's even before you factor in the "working on something" aspect of Spring Training.
Oh the stats guys will try to pull together something or other from
it because that's what they are being paid to do but try not to pay too
much attention. At best these are "maybe something is there" or "confirming what I already thought" types of stats. For the former you need more data, more REAL data. For the latter you already have the data, this is just the final shovel of dirt on the coffin.
Ok fine. What are these things? Well for me I look at three things
(1) If someone hits for substantially more power than they usually do. That's falls into the "maybe something is there" category.
(2) If someone does terrible and you thought they might be terrible. That falls into the "confirming" category. But here when I say "terrible" I mean "terrible". ERAs over 6.00. BAs under .100. That sort of thing.
(3) I'll also allow a "I don't really have a preference between these two players so I'll let Spring Training decide". Something has to right? The Nats shouldn't be in this position. They have decent reasons to make their decisions. But if you are choosing between say Jedd Gyorko and Johnny Giavotella for your starting 2B, first I'm sorry, and 2nd, sure, what the hell. Use ST stats to make the decision.
Anyway friendly reminder out of the way. When Solano or Difo hits .380 for the Spring I hope you remember this when I call your cries of "Start him!" stupid.