Short Story: In 1987, garish power numbers helped Blue Jays’ LF George Bell edge out Tigers SS Alan Trammell for the American League MVP. The bad news? Bell was – at best – the 5th most valuable player in the AL that season.
Longer Story
The 1987 American League MVP vote was one of the closer races in history. Toronto left fielder, George Bell won the award with a total of 332 points (16 first-place votes) over Detroit shortstop Alan Trammell, who received 311 points (12 first-place votes). Twins center fielder (and future Hall of Famer) Kirby Puckett came a more distant third (201 points) with Red Sox 1B/RF Dwight Evans and Brewers DH/3B (and future Hall of Famer) Paul Molitor rounding out the top five.
In 1987, the Tigers and Blue Jays were locked in a tight race for the AL East (and best record in MLB) for the entire second half of the season. From mid-July onward, neither team was more than 3.5 games out of first place. Going into the final week of the season, Toronto looked to be in control. They had their largest lead of the season, but they ended the year on a seven game losing streak (including a final series sweep in Detroit) to end the year two games back of the Tigers.
Like the Jays and Tigers (the two winning-est teams in MLB that season), Bell and Trammell were also close mirrors of one another in several ways. Each was the best player on his team – Bell arguably so (a good case can be made for SS Tony Fernandez), and Trammell was unquestionably so, and each spent the entire season hitting clean-up in his team’s lineup.
The 1987 American League MVP Vote
By the stats popular at the time, the original vote makes a decent amount of sense. Bell was first in the American league in RBI, second in both HR and slugging (Oakland’s Mark McGwire, then in his rookie year, was first in both), and second in runs scored (to Molitor). He also hit .308, a very respectable number for a power hitter in the 80s.
Trammell on the other hand didn’t lead the league in any category. He did finish 3rd in batting average (behind Boston’s Wade Boggs and Molitor), 5th in runs scored, 8th in slugging, and 10th in RBI. He finished well back of Bell in HR, hitting 28, good for 23rd in the league. Trammell also contributed 21 stolen bases while getting caught only twice.
Here are their full (traditional) stat lines:
1987 was a career year for each player, with Trammell hitting his career high marks for batting average (Bell has his second highest, behind his .309 in 1986), and both setting career highs in on-base percentage, slugging, home runs, and RBI.
Breaking Down the Original Vote
The traditional stat line alone suggests an issue with the original vote, even before we start looking at our advanced stats. On the surface, their offensive production seems relatively similar, with Bell having better power numbers and Trammell having more balanced production.
Of course, offense isn’t the entire picture. Trammell produced those numbers while playing above average defense at shortstop (saving about 8 runs above average)
To give you some idea of how rare production like Trammell’s was: in 1987 Trammell had a wRC+Weighted Runs Created Adjusted – a comprehensive stat measuring offensive production relative to the league average of 152, meaning he produced 52% more runs than the league average player. No other infielder in MLB (not just the AL) had a wRC+ of even 130. The next middle infielder on the list is Yankees 2B Willie Randolph at 127 and the next SS on the list is Cleveland’s Julio Franco at 120.
On the flip side of the coin, there were six left fielders (and 18 OF total) who produced a wRC+ of 125 or more. Just looking at outfielders, Bell is actually in the middle of the pack at #9 – eight outfielders in the majors created more runs than Bell did in 1987.
When we combine Trammell’s defensive contributions with his more offensive production (especially his significantly higher on-base percentage) it’s clear why his actual value was significantly higher than Bell’s.
Trammell leads Bell in all three of the weighted offensive measures (OPS+Adjusted On-Base plus Slugging, a measure of the player’s OPS relative to the league average, wOBAWeighted On Base Percentage: A comprehensive measure of offensive production put on a scale roughly equal to “normal” on base percentage. .320 is average, .340 above average, .370 great, and so on., and wRC+) and their Speed Adjusted SluggingSaSlg, slugging adjusted for stolen bases and caught stealing is more than 30 points closer. Trammell is also way ahead in WAR (with Baseball Reference being a bit more generous than FanGraphs), reflecting his far greater defensive contributions.
I don’t think it’s a stretch to say that in 1987, voters saw two roughly equal stat lines and voted for the one with the shinier power numbers. Afterall, everyone digs the long ball, right? Examining the situation even a bit closer, Trammell was the far superior player.
In addition to his stats being better, he was also on the team that won the most games in MLB, leading his team past Bell’s team for the division title. While I personally don’t consider that very relevant to evaluating a player, the MVP voters consistently do (though in this case, the fact that the Tigers then lost to the Twins in the ALCS likely mitigated some of that benefit).
An Interesting Wrinkle
Oddly, I have heard more than one person put forth the idea that the Blue Jay’s end of season losing streak (and Bell’s poor performance in those games) actually gained Bell votes. The thinking here is that Bell’s value was demonstrated by the fact that when his performance dipped, his team reeled off seven straight losses because they needed him so much.
On a most basic level, this narrative is true: the Jays did drop seven straight to end the season and Bell’s performance in those games was indeed abysmal. He went 3 for 27 with four walks, no extra base hits, no runs scored, and one RBI. That said, there are several issues with this explanation of things.
First of all, I can’t find a single article from the time putting forth this idea when discussing the MVP vote. All of the articles I can find (examples here, here, and here) all discuss Bell’s slump as a negative that hurt his MVP prospects.
Further, one of the beautiful things about baseball is that one man does not a team make. It takes a lot of players playing badly to lose seven straight, and Toronto was bad from top to bottom. Collectively they batted .188 with an OPS of .548 (with only 16 runs over those 7 games). Those numbers are, not surprisingly, well below the team’s season marks of .269 and .782.
Finally, Bell’s late season swoon didn’t affect his season-long averages as much as one might think. Prior to their seven-game losing streak, the Jays had put together a seven-game winning streak. Like teams, players also go through hot and cold stretches and during his team’s winning streak, Bell was on fire, hitting .483/.500/.759. His slash line prior to the winning streak? .309/.350/.621. His slash line after both the hot streak and cold streak? .308/.352/.605. His cold streak was a bit colder, but not by much.
The Re-Vote
Hopefully you’ve figured out by now, I’m suggesting that Alan Trammell should have won the American League MVP in ’87. He was the best player on the best team and had stats that place him within the top few players in the American League. If the vote were held with today’s stats and the voter climate now, it likely looks something like this (original vote rank in parentheses):
Note: that’s not how I would personally vote, but rather how I think voters would vote with the way they look at stats today. Puckett was certainly not the 3rd best player in the AL in 1987, but he was the face of the eventual World Series champs and he’d still get a boost with voters because of that (just as he did in ’87). And, just as today, playing for a sub-.500 team hurts a player, so Boggs would have a hard time passing Trammell, even though his stats were better.
A couple of interesting things stand out when looking at the 1987 American League MVP vote.
One, Roger Clemens had an incredible year, following up his incredible 1986 MVP season with a performance that was likely even better (both rWAR and fWAR say he was about 1 WAR better in 1987 than 1986). There was zero chance Clemens was going to garner support a second time though, especially not with the years position players put up.
Two, Wade Boggs (Clemens’ teammate on the 1st-to-5th Red Sox) had a monster year. Like Trammell and Bell, this was likely his career year (1985 is the other candidate). He had his 3rd-highest season batting average, his 2nd-highest on-base percentage, and his career high in RBI and slugging (by a wide margin) due to his 24 home runs – more than twice as many as any other season in his career and only one of two seasons where he hit 9+. He hit more than 20% of his career total in 1987 alone.
The final player whose performance is notable is Paul Molitor.
Prorating Molitor’s stats for 20% more games (to bring him to 147) he’d have 20 HR, 131 RBI, 142 runs scored, 51 doubles, and 56 stolen bases. Given ’87 was the year that Molitor had his 39-game hitting streak, if he’d have stayed healthy, he might have taken the MVP over both Trammell and Bell.
Alas, that didn’t happen (Molitor was rather fragile during the first half of his career before becoming a stalwart into his 40s). Molitor didn’t win, nor did the best player (Boggs or Clemens), nor did the player who “should” have won (Trammell). Instead, the sparkly appeal of big power numbers held the voters’ attention long enough to elect a player who – if we’re being brutally honest – barely sneaks into the top ten in the league for that year.