I'd rather look at games back instead of a pre-season estmate of what'll be needed. As the season progresses, this'll give a better indication of what kind of a hole you have to climb out of. It's perfectly reasonable to pick up a game a month - even over a fairly protracted amount of time. Many teams have done this. Over short periods, it's possible to do significantly better.
Your approach is to use season-long projections and apply them over ever shorter periods of time. The shorter the time period, the more expected spread you'll find and the less unreasonable it is to exceed any projections.
The projection of .570 for the Red Sox means that there's approximately a 50% chance that they'll do better. I forget the expected spread for a season, but even after just one month the spread would go up by about 10%. After two months, it'd go up 22%. Etc.
That's why a team that's four games back the first of September is distinctly "in competition". It's not unheard of for teams to pick up four games in a week. One of the Pirates teams I mentioned earlier was something like 14 games out in mid-June. They won the division by a large margin. They averaged a pickup of something over 5 games a month for the entire rest of the season.
While I'd consider that a pretty historic comeback and the Pirates team had more talent (I think) than the Rox, IMHO it's far too early to rule out a more modest performance. I think I'd be pretty jazzed if the Rox were still 6 games back at 1 September. That'd mean that they'd held their own over the heart of the season. It'd also mean that they'd be close enough that the unthinkable could happen!
|