The posts about Maryland and visiting the link to the Terp's board made me think about the occasional "wonder" years that teams have. Maryland winning 10 games last year certainly was a great story. In the 5 prior years, Maryland won a grand total of 20 games, for an average of 4.0 per year. Improving on a team's 5-year average by 6 games is certainly cause for celebration by that team's fans and eyebrow raising by the rest of the world.

The immediate question, however, is whether it's likely to last. Is the team more likely to continue at the new, higher level or revert to its old form?

To investigate this, I looked on a college football historical site for "wonder" years by teams since 1980. These I defined as years in which a team won 2.5 or more games in excess of its average of the previous five years. So, for example, if a team won 7 games and in the previous 5 years had averaged only 4 wins, it made my list. There are always a few such teams every year; Maryland and Washington St. are two good examples from 2001. Neither one of Davie's 9-win teams qualified, because even in the depths of the Davie era N.D.'s 5-year win average never dipped below 7 per year.

Once I identified a wonder year, I then looked at the next year after the wonder year to see how the team fared. A team might, for instance, win the same or more games as the wonder year (the permanent improvement hypothesis) or it might win fewer games the next year.

As it turned out, wonder years weren't hard to find, so I quit after I'd identified 20 of them, which is a pretty decent sample.

In these 20 wonder years, the average "gain" in the wonder year was 4.5 games over the 5-year average. In the years following the wonder year, however, the gain from the 5-year average was only 2.3. In other words, on average a team "gave back" almost exactly half the gain it made in a wonder year. To illustrate, if a team averaged 4 wins per year for 5 years and then uncorked an 8-win season, on average in the next year it won about 6 games. So, while a wonder year usually indicates that a team has improved, it overstates the improvement somewhat. Instinctively, this makes sense. Often the things that fuel these magical seasons -- a big turnover margin, an easier schedule, winning lots of close games, avoiding injuries -- just don't repeat themselves every year.

Now, of course, there's a complication with Maryland, which is that its wonder year followed a coaching change. So, one might argue, wonder years with the same coach are probably due to a great deal of good fortune, but wonder years immediately following a coaching change should indicate permanent improvement. Thus, an incensed Terp fan might argue, winning 10 games with the Fridge as coach shows that the savior has arrived and Maryland will continue to average 10 wins per season as long as he coaches.

Fair enough; let's look at the data. Wonder years following coaching changes turned out to be harder to find. In the end, I identified 13 of them in relatively recent times. There probably are more of them, but I have no reason to think that this isn't a representative sample.

In the coaching-change wonder years, the average gain from the 5-year average was 4.0 games. In the next year, the average gain from the 5-year average was only about 1.4 games. In other words, the fact that the wonder year followed a coaching change actually means that a team, on average, gave back about two-thirds of its gain. To illustrate, if a team averaged 6 wins per season for the previous 5 years, made a coaching change and won 12 games the next season, on average the team won only 8 games the next season.

Again, the wonder year probably means that the team has indeed improved, but it usually overstates the amount of the improvement. The fact that the wonder year followed a coaching change does nothing to help the team hold onto the gains. In essence, in those years when a program makes two giant steps forward, it's likely to one step back the next year. It's probably headed in the right direction, but the path to improvement is hardly ever a straight line.

So what about Maryland next year? As I mentioned, Maryland had averaged 4 wins per year and won 10 last year. If you were playing the odds, you'd figure them for 6 or 7 wins next year. Again, instinctively, this makes a certain amount of sense. Maryland won't sneak up on anybody next year. The Terps' +16 turnover margin is unlikely to repeat itself (when N.D's +14 margin that fueled the 9-win season of 2000 fell to +3 in 2001, the win total fell too). The trick play that beat Georgia Tech in overtime probably won't work two years in a row. Opening on the road against N.D. (and it will be a road game for Maryland, no matter what the Maryland fans think), will be a little bit tougher test than opening at home against North Carolina, Eastern Michigan and Wake Forest, the stuff of last year's momentum.

So, is Maryland improved and is the Fridge a good coach? The answer to both of those questions is probably "yes." Will Maryland provide a serious test for N.D.? "Yes again" is my answer. But is Maryland more likely to go 10-2 or 7-5 next year? My money's on 7-5.