Here's what people don't always understand about sample sizes. At this point in his career, we are not observing a random small sample of Trevor Hoffman's pitching performances. We are observing a trend, and the trend is telling us that he is done.
Put it this way. Baseball Reference tells us that Trevor's appeared in 890 games. If we were to randomly take seven of those appearances and try to piece together what sort of pitcher Trevor Hoffman was, then we're looking at much too small a sample size. If we want to make a decision now about what sort of pitcher Trevor Hoffman will be in the future, we actually run the risk of looking at too large a sample.
We cannot go back to 1998 to determine what kind of pitcher Trevor Hoffman will be in 2008.
As people who are statistically inclined, we must always be looking at what the statistics are actually telling us and what they aren't telling us. A context needs to continually be applied or else we'll run the risk of being a bunch of Joe Morgans... spouting random "facts".
Let's put it another way. If Trevor Hoffman were some sort of save factory and suddenly started spitting out a defective product, our engineers wouldn't look at how well the factory was running 2 years ago to determine if something needs to be fixed. They'd look at the most recent production and make a determination based on that.
We've all seen it. It's now plain as day. Trevor Hoffman is done.
Boilerplate: When Trevor is elected into the Hall of Fame I will be cheering as loud as anyone. I will probably cry as I think of all of the terrific memories that Trevor provided and I will do my best to explain to Elliott exactly how exciting and magical the song Hell's Bells is despite my not owning any other recordings by AC/DC. I'll talk about Trevor's changeup and how Trevor learned to adjust after the velocity on his fastball declined. I'll use it as an example of the fact that limitations and obstacles can be overcome. Trevor Hoffman is a great person and the greatest closer of all time.