Diamond Mind Baseball


Baseball Articles \| 2001 Team Reviews

2001 Team Reviews -- The Strike Zone

Last updated: February 19, 2002

One of the big stories of the 2001 season was the strike zone. Umpires were ordered to call the high strike, as described in the official rules, and were trained extensively on how to do so during the winter of 2000-2001.

At the time our projected player statistics were generated (late February), nobody knew whether this plan would be followed all season and, if so, what the statistical impact would be. We could have guessed, but instead we opted to generate our player projections under the assumption that the league averages would end up around the levels seen in the 2000 season. And that's how we arrived at the projections that were shipped to our customers in the first week of March.

Roughly two weeks later, we were ready to start running the fifty simulations that would produce our projected team standings. By then, it was clear from spring training games that the strike zone was having a major impact. Here's how we described it in our projected standings article last March:

Through the first fifteen days of spring training, covering 219 games from this season and 215 games from last year, the walk rate has dropped from 6.9 per game to 5.4, a decrease of 22%. The strikeout rate is up from 11.4 to 12.4, an increase of 9%. And scoring is down from 12.1 runs per game to 10.2, a drop of 16%. Those are big differences.

So we decided to make an adjustment, even though we weren't sure those results would hold up in the regular season. Actually, we hedged a little, and decided to make an adjustment that was a little smaller than these numbers indicate. And that adjustment was in effect when we ran the simulations, so it influenced the projected team totals for runs scored and runs allowed.

As it turned out, the umps did indeed change how they called the games, and the following changes in the league averages were seen during the regular season:

                       American       National
Runs per game          down 8%        down 6%
Walks per game         down 15%       down 12%
Strikeouts per game    up 4%          up 3%
Homeruns per game      down 7%        down 2%

As you can see, these changes were smaller than the ones we saw during the first half of spring training. And they were smaller than the adjustment we made when running our simulations.

Now that we are ready to start comparing actual results to projected results, it raises a question as to which projections to use for comparison. Would it serve you best if we:

used our original player projections, the ones we shipped to customers in early March, and the ones that contained no adjustment for the potential impact of the strike zone?
adjusted our original player projections based on the early spring training results so they would tie in with the projected runs for and against numbers in our projected standings article, even though that adjustment would overstate the actual effect of the new strike zone?
adjusted our original player projections based on the actual impact of the new strike zone?

If we don't make any adjustments, there would be systematic differences between the projected and actual stats that were based only on the new strike zone, rather than based on the player's performance and/or the quality of our projection system.

And if we do make an adjustment, we'd end up presenting projected stats that were different from those we put out last March, and that could confuse some people and make it seem as if we're trying to fudge the results just to make our projections look better.

In the end, we decided to go with what we published last spring -- the original unadjusted player projections and the overly-adjusted team totals for runs for and against.

So, as you peruse these team reviews, please keep in mind that the new strike zone caused actual player stats to be a little below what we projected, especially for batting average, walks, runs, and RBI. In other words, a hitter who matched his projected numbers performed above, not at, the expected level.

And keep in mind that teams generally scored and allowed more runs in real life than they did in our simulations. So a team that matched its projected runs total actually performed a little worse than expected.