Prospectus: Spring training vs. regular season (2024)

They're already grumbling in Philadelphia. The Phillies, after astrong offseason that has elevated them to the status of NL Eastfavorites, have started their exhibition season by losing four oftheir first five games. Phillies fans, not exactly the most patientand supportive group around, are beginning to gripe about the poorstart on talk radio and the internet.

But I'm here to put Philadelphians' minds at ease, at least until nextmonth. Even if the Phillies play this way for the rest of March, itwon't have any bearing on whether they'll be playing in October.Aside from dugout chatter, power ballads, and Yogi Berra quotes, there is nothing less meaningful in the baseball world than spring training records.

That's not an especially controversial statement. No one bases theirpreseason expectations entirely or even mostly on the spring trainingstandings. And plenty of analysts have looked for a strong linkbetween spring success and regular season success over the years, onlyto come up empty. But fans and writers still give in to thetemptation to read something into spring records, especially when areputedly good team has a bad record or vice versa. So it's worthtaking a look at some numbers to see just how little March means forthe months that follow.

To get a sense of how well spring training records predict regularseason success, we'll look at all spring training records back to 1996(stopping there before we got to the severely abbreviated springtraining of 1995). Let's look first at the very best spring teams.

Does a great spring mean a great regular season? Here are the 15top spring records since 1996, along with records in the subsequentregular season:

 Spring Training Reg SeasonTeam Year W L Pct W L PctFlorida 1997 26 5 .839 92 70 .568Boston 1998 20 8 .714 92 70 .568Kansas City 1999 22 9 .710 64 97 .398Los Angeles 1999 21 9 .700 77 85 .475Baltimore 2002 20 9 .690 67 95 .414Oakland 2001 22 10 .688 102 60 .630Arizona 1999 22 10 .688 100 62 .617Houston 2002 19 9 .679 84 78 .519Texas 1998 21 10 .677 88 74 .543California 1996 21 10 .677 70 91 .435Baltimore 1998 18 9 .667 79 83 .488Detroit 1996 20 10 .667 53 109 .327Arizona 2002 23 12 .657 98 64 .605St. Louis 1997 21 11 .656 73 89 .451San Diego 1998 19 10 .655 98 64 .605TOTALS 315 141 .691 1237 1191 .509

There isn't much reason to believe in spring training records here.Seven of these 15 March Monsters finished below .500 in theregular season. There are some outstanding teams on this list, butthere are some truly awful teams too. In particular, notice the greatspring put together by the 1996 Detroit Tigers. After they leftFlorida, they turned out to be the very worst major league team sincethe '60s.

(The '96 Tigers' mirror image -- the 2001 Seattle Mariners, with a116-46 regular season record -- showed no signs of their upcomingdominance during spring training. They finished March at 13-19.)

Even if having a great spring record doesn't mean anything, maybehaving a spring meltdown is a sign of bad things to come. What aboutthe very worst spring performers? Here are the bottom 15 springperformances of the past seven years:

 Spring Training Reg SeasonTeam Year W L Pct W L PctMontreal 1998 8 23 .258 65 97 .401Toronto 1996 9 22 .290 74 88 .457Cincinnati 2002 9 22 .290 78 84 .481Florida 1996 9 22 .290 80 82 .494Anaheim 1997 9 21 .300 84 78 .519San Francisco 1998 9 21 .300 89 74 .546San Francisco 2001 9 21 .300 90 72 .556New York AL 2001 9 20 .310 95 65 .594Milwaukee 1998 10 22 .313 74 88 .457Montreal 1996 9 18 .333 88 74 .543Chicago AL 2002 11 21 .344 81 81 .500Milwaukee 1999 11 20 .355 74 87 .460 Milwaukee 1996 11 20 .355 80 82 .494 Florida 2002 10 18 .357 79 83 .488 Atlanta 1997 10 18 .357 101 61 .623TOTALS 143 309 .316 1232 1196 .507

This list tells pretty much the same story as the previous one. Theworst four spring training teams finished below .500 in the regularseason, but the next four, and seven of the 15 overall, finished.500 or above. The most surprising thing here is that the very worstspring training teams of recent years cumulatively playedbetter than .500 ball during the regular season, effectivelythe same as the very best teams shown in the previous table.

Of course, if we really want to know about the predictive power ofspring training, it's better to look at all the teams rather thanjust the ones on the fringes. We can do that by figuring thecorrelation coefficient between spring and summer records. Acorrelation of 1 would mean spring performance predicts regular seasonperformance perfectly; a correlation of 0 would mean there was norelationship at all. Since 1996, the correlation between springrecords and regular season records is 0.15. So there is arelationship, but it's very weak. For comparison, the correlationbetween last year's regular season record and this year's regularseason record during the same period is a much stronger 0.52.

So a team's record in the spring means next-to-nothing about theirrecord in the regular season, but maybe that's just because they playso few games in the spring -- not enough for teams' true abilities toemerge. If we look at spring training records over a longer period oftime maybe the best teams will show themselves.

No such luck. Here are the top five overall spring records since1996, along with regular season records during the same period:

 Spring Training Reg SeasonTeam W L Pct W L PctArizona 101 64 .612 440 370 .543San Diego 129 86 .600 560 574 .494 Texas 120 87 .580 566 568 .499 Baltimore 114 83 .579 547 586 .483 Oakland 125 97 .563 600 533 .530 

Three of the top five spring performers of recent years are sub-.500teams in the regular season, even when you look at seven years' worthof games. Meanwhile, the best teams in the league, the Yankees andBraves, were each below .500 in the spring. The correlation betweenspring records and regular season records over the seven years is 0.28-- better than we saw in single seasons, but still not very strong.

So there's very little relationship between spring training recordsand regular season records, in the short term or the long term,whether you look at all teams or just the extreme ones. And that'sactually pretty surprising when you think about it. After all,they're largely the same teams playing just months apart. Why can'twe tell anything about a team from how it plays in March?

You can list the reasons as well as I can. It's partly the parade ofminor leaguers and NRIs in the spring who will never see playing timeduring the regular season. It's partly the severely unbalancedschedules that teams play in the spring. It's partly the lack ofmotivation for veterans who are just marking time until Opening Day.

But more important than any of those is that the game itself isfundamentally different in the spring than it is in the summer.Players are yanked from the game with no regard to the score.Pitchers will go an entire appearance without using their best pitchin an effort to work on the No. 2 and No. 3 pitches intheir arsenal. Managers will send up a lefty pinch-hitter against alefty reliever just to see how well he can go against his platoonsplit. In short, the game is not played to win.

And that's fine. There isn't a baseball fan around who doesn't lovespring training, where the stakes are lower, the rosters are bigger,and the players are carefree. Enjoy the games. Just don't foolyourself into thinking that a win for your team in March willtranslate into five in June.

You can check out more work from the team of writers of the Baseball Prospectus at baseballprospectus.com. Baseball Prospectus is a registered trademark of Prospectus Entertainment Ventures, LLC.