07 March 2011
After the first 10 days of Spring Training, many Diamondbacks' fans are already worried about the team's poor performance in the Cactus League. Heading into Monday's games, the team has a 3-8 record, and the offense has been particularly disappointing. It is well known that teams are not playing their regular lineups in much of spring training games, and the starting pitchers usually only work a few innings. So how meaningful are the results from Spring Training games?
Let's start with the Diamondbacks' Spring Training results over their thirteen year history. The second column (ST) is the Diamondbacks' record in Spring Training Games. The third column (Projected ST) would be the Spring Training record projected over 162 games. The fourth column (MLB Games) is the team's actual record in the regular season, and the last column is the difference between the Projected Spring Training record and the actual record.
|Year||ST||Projected ST||MLB Games||Difference|
In general, the Diamondbacks have won a lot of games in the Cactus League, having a .500 record or better in 9 out of 13 years, while they finished over .500 in the regular season only 6 out of 13 times. One thing that jumps out is that the teams that did well (1999, 2001, 2002, 2007) all played very well in Spring Training. So is Spring Training really a useful indicator for regular season success? Probably not. A look at a larger sample size shows that the "trend" shown by the Diamondbacks doesn't really match most other teams.
The graph below (from Cork Gaines at BusinessInsider.com) shows a plot of Regular Season winning percentage vs. Spring Training winning percentage over the past eight seasons. If you look at the teams at the far right side of the graph, who ended up winning more than 2/3 of their regular season games, more actually had a losing Spring Training record than a winning one. Similarly, if you look at teams with a Spring Training winning percentage below .400, almost half had a regular season record over .500. Over 2/3 of the teams had a difference of over .100 percentage points between their spring training and regular season records, and 18% differed by over .200 percentage points (200 points is the difference between 98-64 and 65-97). Plotting a trendline through the data shows a slight correlation between the two, but the line is nearly flat rather sloping sharply diagonally upward.
(Graph courtesy of Cork Gaines at http://www.businessinsider.com/chart-of-the-day-spring-training-record-regular-season-2011-2).
Another interesting study was performed by Michael Wolverton a Baseball Prospectus and ESPN in 2002. He looked at the Top 15 Spring Training records from 1996-2002, and those teams (including Arizona in 1999 and 2002) combined for a 315-141 record (.691 Winning Percentage) in Spring Training. Yet the combined record for those 15 teams in the regular season was only 1237-1191, a .509 winning percentage. Similarly, the 15 worst Spring Training records during that period were a combined 143-309 (.316). But those teams ended up 1232-1196 (.507) in the regular season, virtually the same as the teams with the very best Spring Training records.
In conclusion, while the Diamondbacks' best teams have historically also performed well in Spring Training, that is not a trend that seems to apply to Major League Baseball in general. Looking at the results of all 30 teams, it seems that an excellent Spring Training record (greater than .667) is just as likely to lead to a losing season as a winning one, and a very poor Spring Training record (< .400) still results in a winning regular season team over half of the time.