By the Numbers

December 31, 2007
A thought by James Berardinelli

In the wake of my recent statement that IMO 2007 was the worst year for movies since I started reviewing, I received a reader's challenge asking if I could come up with numbers to support that assertion. So I spent a few hours this past weekend crunching numbers. The results, while not earthshaking, do provide some interesting insights into what my star-rating system deems to be the best and worst years of movies in the period 1994-2007.

One of the first things I discovered was that the ratings for 1993 deviated so significantly from the norms that they couldn't be grouped with the rest of the years. The reasons for this aren't hard to understand. I started reviewing in 1992 but the reviews I wrote that year were short, incomplete, and not always accompanied by any sort of rating. (Any 1992 reviews that have migrated to this site have undergone an overhaul.) 1993 was the first year in which I began employing a systematic rating system. Those who remember that far back may recall that it was a numerical 0-10 scale, not a four-star one. In 1994, I changed to the four-star scale (although I kept a "dual system" for a short period). However, everything for 1993 was eventually converted to the four-star system and the conversion method was imperfect. This doesn't invalidate the 1993 review numbers; those that are "off" are within 0.5 stars, with the most typical divergence being a movie that today would get three stars being accorded two-and-one-half stars. I have no plans to go back and attempt to re-examine the entire year's worth of reviews for what would like amount to about 40 changes.

For the purposes of this study, I have divided ratings into three categories: recommended (3, 3.5, and 4 stars), borderline (2 and 2.5 stars), and not recommended (<2 stars). The 14-year average is 49% recommended, 38% borderline, and 13% not recommended. Almost every year during the span comes close to these averages; it's rare that any year is more than 5% off in any category.

So, based on the numbers, what were the best years? Those would be 1995, 1999, and 2002. Looking at the years purely on the basis of the recommended/borderline/not recommended categories, 1999 is the easy winner. An impressive 54.4% of the year's films were recommended, 34% fell in the borderline category, and 11.6% were not recommended. 2002 was in second place with 52.7%, 32.7%, and 13.6%, respectively. 1995 deserves special attention. The "recommended" category was only 46%, considerably below average. "Borderline" was 44.2%. "Not recommended" was a skinny 9.9%. The devil's in the details, however. 1995 had a whopping 9 four-star films (by far the largest number of any year) and an equally gaudy 43 three-and-one-half-star films. Those are big numbers. Add those percentages together and you get 22.4%. No other year is even close. Second place is 1994 with 19.2%. Putting all this together, I would rank 1995 as the best year for films since I have been reviewing, followed by 1999 then 2002.

Now, what about the worst years... Once the numbers have been crunched, there are really only two candidates. The first is 2000. The combined 4/3.5 score that year was 8.0%. The "recommended" number was 44.5% The "borderline" number was 40.8% and the "not recommended" number was 14.7%. The other candidate is 2007. "Recommended" = 45.4%; "Borderline" = 40.6%; "Not recommended" = 14.0%. The combined 4/3.5 score was 10.6%, although there were no four star movies. Nevertheless, based on the numbers, 2000 was worse than 2007. So even though my gut tells me 2007 was disappointing, maybe the passage of seven years has dimmed the memory of what 2000 was. A total of only 17 four and three-and-one-half-star movies combined! Yikes!

Bottom line: Even the best and worst years haven't deviated dramatically from the average, but it's possible to pick out the best years (1995, 1999, and 2002) and the worst (2000 and 2007). Here's hoping 2008 leans more toward the former category than the latter one.

This is the kind of column that results when I start playing around with numbers. No more like this for a while. Tomorrow, a few predictions for 2008, then back to a "regular" schedule starting Thursday. The lack of video columns for last week and this week is easily explainable by the lack of significant releases. The video column will return next Tuesday and continue every Tuesday for the foreseeable future. Also, reviews of older movies will begin this Saturday. Check the main page and the main video page for titles and links to the reviews.