Team Chemistry in Major League Baseball
How close were our preseason predictions?
Posted Oct 09, 2014
Its playoff time in major league baseball, and, as teams researchers, its also time to ‘face the music’ and see how accurate our pre-season predictions were about the which teams had the best ‘team chemistry.’ Back in March, ESPN the Magazine asked us to evaluate all 30 major league teams’ ‘team chemistry’ for the baseball preseason issue and assess what was the bottom line- how much would chemistry itself (or, how well a team functions and works together) make a difference in extra games won during the season for each team. Contrary to some popular media, we don’t consider team chemistry to be about how much players ‘like each other’- there are too many examples of teams that reportedly didn’t get along off the field yet won championships.
So, how did we go about doing this? Research on groups and conflict has shown that considering demographic rifts or splits within a team can affect working relationships and functioning, or team chemistry. So, to make our evaluations, we considered three factors. First, we used a formula, called the faultline algorithm, for assessing the amount of demographic and other ‘splits’ or rifts with the team- the more splits, the worse for chemistry. Splits that resulted in one person being very isolated from the rest of the team counted even more against chemistry (people that have little in common with anyone else on the team). Finally, we calculated splits based on salary levels (does the team have a few players paid way more than anyone else, or is the pay distribution relatively even). Based on this faultline model, we estimated that on average teams could gain around three wins a season based on chemistry alone (in other words, the composition of the team itself, beyond the talent of individual players, is worth about three wins on average).
Now on to our predictions, and how they turned out, division by division in baseball:
American League East
Two of our best teams “chemistry-wise” were Tampa Bay and Toronto. We predicted they would each gain about two games on the basis of good chemistry. While Toronto finished close to where we predicted, the Tampa Rays noticeably underperformed. But the problems there seemed to be more about injuries than any chemistry issues. What about the New York Yankees and Boston? We predicted both to lose games due to poor chemistry, and both teams were well off the pace set by…Baltimore, (the surprise winner of the division) which we predicted would have positive chemistry despite many sports publications predicting a last place finish for them, so we are feeling a little better about that one.
American League Central
Most people thought Detroit would win this division (despite losing a game to chemistry issues) given all their talent, and they did. Yet, the other surprise team in the American League, Kansas City, was, like Baltimore a team we predicted would have exceptional team chemistry. In Kansas City’s case, about three extra wins in our model were attributed to them having no ‘isolated’ players demographically- which groups research would predict to be a critical factor in team performance. So, it could well be that positive team chemistry helped KC nearly win the division from Detriot and reach the playoffs for the first time in decades. Cleveland, the Chicago White Sox and Minnesota all finished about where we thought they would, performance and chemistry-wise.
American League West
Ugh. Texas, the pick in the ESPN article to win the division (and a beneficiary of pretty good chemistry) finished dead last. Like with Tampa Bay, however, injuries (and possibly issues with their coach, who resigned late in the season) had no doubt much to do with their demise. This left the division to primarily Anaheim, Oakland, and Seattle; we predicted that Oakland would have a lower overall chemistry score than either Anaheim or Seattle, who they tied with at the end after being division leaders much of the year. Interestingly, we assessed Houston’s team chemistry as above average, while at the same time being short on star players; while Houston won only 70 games, the overall consensus was they did much better than expected given their talent alone (and managed to finish ahead of Texas in the standings at least). And Anaheim did much better than in the 2013 when we assessed their chemistry as relatively poor. Their changes in composition of players this year seemed to pay off.
National League East
As the predictions in ESPN said, Washington won the division; though the difference over Atlanta was expected to be only around one game, which was Washington’s advantage over the Braves in our team chemistry calculations. In our preseason assessment, we noted concerns about a lack of age diversity- sometimes known as veteran leadership, with respect to the Braves’ roster. That turned out to be exactly the buzz in much of Atlanta media late in the season when the team went 7-18 in enduring a September collapse and well behind Washington. The New York Mets, Florida and Philadelphia followed close behind, with both the Mets and Phillies (in last place, as expected) having among the worst chemistry scores we calculated for all teams.
National League Central
St. Louis, as the ESPN article again predicted, won this division with Pittsburg only two games behind. All of the teams in this division had negative scores for chemistry with the exception of Cincinnati, so differentiating on the basis of chemistry for this group may mean little. Yet, Cincinnati’s expected one win due to favorable chemistry only got them a fourth place finish behind Milwaukee, with Chicago Cubs, as expected, in last.
National League West
One again as expected, this division was a race between Los Angles and San Francisco, with the best chemistry score of all the teams in this group closing the anticipated talent gaps between these teams. Of the Colorado, San Diego and Arizona teams making up the rear of the division, San Diego had the best overall chemistry score which was worth an estimated one game in the standings and a third place finish.
So, what can we conclude from comparing our preseason chemistry calculations with what actually happened? As with all sports predictions, we missed some but were on the money with others. Injuries, mid-season trades and other factors can’t really be predicted, so that’s a caveat to the whole prediction game. But it has helped us put our model to the test of what actually happens in teams, and encourages us to account for, in the future, factors such as coaching that can play a role in team chemistry as well. Stay tuned!
Written by Chester Spell and Katerina Bezrukova