Experience isn’t an advantage in pennant races

It’€™s going to be a barnburner of a race in the AL Central. The Kansas City Royals and the Detroit Tigers are running neck-and-neck to be champions of Middle America. 

Obviously, the Tigers have the advantage because they’€™ve "€œbeen there before." The Tigers have been to the ALCS the past three years, and have been through their share of pennant races. The team is filled with seasoned veterans who know how to manage a pennant race because of their vast experience. In a stunning feat of doublethink, the Kansas City Royals also have an advantage. The Royals’ last trip to the playoffs was when "1984"€ meant "€œlast year."€ Come to think of it, the Royals’ playoff drought is older than most of the Royals’€™ regulars. But because they have a team of young guys, they are hungry. Starving really. They’€™re just a fun bunch of young guys who don’€™t know they should be scared. They play loose. Their lack of experience is actually their greatest asset.

The actual 1984 pennant races weren’€™t all that interesting. Oddly enough, the two division winners in the American League that year were Kansas City and Detroit. The Royals were a team of grizzled veterans who were in their ninth consecutive year (excluding the strike-shortened 1981 season) of finishing either first or second in their division. The Tigers, on the other hand, had finished second in their division in 1983, but had not been to the postseason since 1972. Funny how history has a nerdy sense of humor.

But, yes, the capacity of humans to hold two completely contradictory statements in their heads as "€œfact,"€ something that George Orwell, author of the book "1984," dubbed "doublethink," never ceases to amaze me. So, let’€™s have at the question. Who’€™s right? Does experience in pennant races give a team an advantage or should you back the neophytes?

Warning! Gory Mathematical Details Ahead!

Let’€™s get an operational definition of a "€œpennant race."€ Similar to an article I wrote last year, I’™m defining being "€œin a pennant race"€ as a regular season game in September (or early October) in which a team is either within three games of a playoff spot or holds the playoff spot by three games or fewer over its closest competitor but has not yet clinched.

I counted the number of "€œpennant-race"€ games that a player had actually played in (either by coming to bat or appearing on the mound) on a time-series basis. If a player appeared in a pennant race game on Sept. 16th, he’€™d get credit for that game the next day on the 17th, and he retains credit for that game until he retires.

Then, for the years 2009 to 2013, I found all cases in which there was a game where at least one of the teams was in a pennant race. (Don’€™t worry, the model knows about the pennant race games played prior to 2009.) I used my trusty standby, the log-odds ratio model, to get a probability that each plate appearance in that game would end in one of several outcomes (e.g., a strikeout), based on the seasonal stats of the batter and pitcher (min 250 plate appearances that season, each). We don’€™t want to over-count the fact that games involving pennant chasers usually involve better players.

There’€™s often an imbalance in September, with a lot of games where one team is fighting for its life while the other is fighting the urge to leave early because it’s meaningless. To correct for that, I looked only at plate appearances in which the batting team was in the middle of a pennant race and entered in how many previous "œpennant-race"€ games that the batter had been in. I did similarly for plate appearances involving a pitching team playing for a playoff spot.

I then ran a series of logistic regressions predicting the outcome of each plate appearance with the anticipated probability from above (as a control) and the number of "œpennant-race"€ games that the batter (or pitcher) had been in. Later, I looked for games in which both teams were in the thick of a race. It didn’€™t change the results.

If previous pennant-race experience makes a difference, then we should see hitters with more previous experience getting more hits than we would expect based on the batter-pitcher matchup and pitchers fanning more hitters. The coefficient in the regression equation for that variable would be a significant predictor of success.

Except that it wasn’€™t. At all. For anything or anyone. There was no impact of previous pennant-race experience on what happened on the field. Hitters hit pretty much like we would expect given their overall talent, and pitchers pitched like it. As un-Hollywood as it sounds, the team was merely the sum of its parts. That sum was good enough to be in the hunt for a playoff spot in September, so it was probably a pretty good sum. But there is no magical edge conferred by previous experience (or lack of experience) in winning "€œgames that count."€ Neither the young Royals nor the battle-tested Tigers have an edge going into this year’€™s AL Central race. Sorry, guys. You’€™ll actually have to win it on the basis of silly things like talent.

The Ministry of Truth

We’€™re quickly approaching baseball’€™s version of Groundhog Day, when baseball writers all collectively see their shadows and write nothing but cliches for the next six weeks. Coincidentally, all 15 or so legitimate contenders for one of the 10 playoff spots available this year have some special quality about them, whether it’€™s youth or experience or both that will undoubtedly give them "€œthe edge"€ in getting into and succeeding in the playoffs. Or so will say their local beat writers. Once someone finally wins the World Series, that team will be lauded for the amazing strength of character that we all should have noticed before Game 7 was over. It’€™s as if someone will come in after the fact and rewrite the narrative of what actually happened to some party line. What kind of world would that be?

We’€™ll hear that it really matters how a team finishes down the stretch because that gives it momentum headed into the playoffs. (Nope.) That teams that have postseason experience are more likely to win October. (No.) That teams that have to fight all the way down to the last day have an advantage over those who coast in. (That one’€™s actually sorta true.) That big home runs "€œturn the tide"€ of momentum. (No.) That one team is the overwhelming favorite to win the World Series. (Nope.) That teams that can’€™t play "€œsmall ball"€ will struggle in the playoffs. (No. A run scored in a way aesthetically pleasing to the hipsters and 70-year-old men out there counts exactly the same as an entirely predictable home run.)

There needs to be a Snopes for bad baseball cliches. I guess that’€™s us.

Today we learn that having been in a pennant race before does not grant you superpowers to help you cope with this year’€™s pennant race. Here’€™s the twist: There probably are guys who learn from last year’€™s experiences in dealing with the stress of this year’€™s pennant race. In that way, their own personal experience was a great tool for learning how to handle pressure, and that’€™s great. The problem is that people will often make gigantic leaps of logic from one case study to the rest of MLB players.

The truth is that directly experiencing something is not the only way to learn from it. I have no doubt that being in the middle of a pennant race calls for some good coping strategies. Some players have already developed pretty good coping strategies by experiencing other stressful events. Some just have well-developed strategies in advance of needing them. On the flip side, going through something doesn’€™t mean that you learn how to handle it the first time. Know anyone who’€™s been through more than one bad breakup?

If nothing else, let this be a good reminder that ignorance is not strength, and that you shouldn’€™t believe everything you hear over the next few weeks. There will be plenty of outright garbage peddled as truth. Maybe some of it is true. We should always maintain a skeptical (in the proper sense of the word) attitude toward these things. Before you believe, ask for the proof.