A Thread on Chemistry

[now]

Here’s a thread with more about how the intangibles of team chemistry can be measured and what some results are. In another thread I outlined some research, but left out this most recent study: “In Search of the Holy Grail: Team Chemistry and Where to Find It” (Scott A. Brave, R. Andrew Buttersy, & Kevin A. Roberts, 2018). (click link for pdf)

It’s a technical research paper in statistics so I’ll only pass on here some salient points, including some Orioles notes...

“Roughly 40 percent of the unexplained variation in team performance by WAR can be explained by chemistry.”

“In the movie Miracle about the 1980 U.S. men's Olympic hockey team, there is a scene where Herb Brooks is talking about the team he has chosen to his assistant coach Craig Patrick where he says: “I'm not looking for the best players, Craig. I'm looking for the right ones." (Guggenheim, 2004). This statement sums up in a nutshell how we view team chemistry in this paper. What defines the "right fit" of players can often be very subjective, however. Here, we provide a formal definition based on player complementarities, or complementary skills, in the production of team wins via the interconnectedness of teammates' on-the-field interactions. This allows us to then objectively measure team chemistry in Major League Baseball (MLB).

“[It’s] possible to capture a team's historical ability to consistently turn individual player talents into extraordinary team outcomes, allowing for a relative ranking of MLB teams that can be used to measure organizations on the dimension of team chemistry, or what we refer to as organizational culture.”

“[At a team level] we demonstrate that it would have been possible to improve upon PECOTA pre-season projections for the 2008-2016 seasons by a statistically significant margin of roughly 1 win on average.”

“[One study] estimates that team chemistry can account for up to four wins in a regular season based on characteristics of roster composition like wage parity and demographic variation.”

“[Team Chemistry results]: For instance, the 2012 Orioles, 2008 Angels, 2007 Diamondbacks, 2006 Athletics, and 1998 Padres all show up as teams with large positive tcWAR values and the 1998 Mariners, 1999 Royals, and 2015 Reds all show up as teams with large negative tcWAR values.”

“[Cp. fWAR, bWAR]: Of the 2012 Orioles' nearly 15 team wins above fWAR's expectation, tcWAR attributes roughly 4 of these to good team chemistry. In contrast, of the 2012 Orioles' nearly 8 team wins above bWAR's expectation, tcWAR attributes roughly the same number to team chemistry.”

“our results are consistent with the notion that team chemistry may be accurately described as catching lightning in a bottle."

Orioles rank in Team Chemistry wins above expected, 1998-2016: #29

Last on list of Players chemistry rankings: Joey Rickard

Negative rankings for Players Intangibles: Bundy, A. Escobar, Schoop

=====

Okay, maybe that shiny “Holy Grail” is still elusive, but I’m glad to see some efforts made to quantify this long debatable and subjective side of MLB player value and team construction. Notably the realm of "clubhouse chemistry" was left out of this study, though reference was made to (Cubs) teammate accolades of intangibles poster boy, David Ross, named also in one of the earlier studies. 

 

 

 

 

[Beef Supreme]

Just moments before you posted this, I was shoveling snow and thinking about how the Orioles during the Showalter era were greater than the sum of their parts. (Except for the last year and one month, of course.)

[Il BuonO]

24 minutes ago, Beef Supreme said:

Just moments before you posted this, I was shoveling snow and thinking about how the Orioles during the Showalter era were greater than the sum of their parts. (Except for the last year and one month, of course.)

Also, ‘a lot of want to’.

[weams]

2 hours ago, Il BuonO said:

Also, ‘a lot of want to’.

This is chemistry.

 

[now]

56 minutes ago, weams said:

This is chemistry.

 

Wow! 24-0?

[now]

3 hours ago, Beef Supreme said:

Just moments before you posted this, I was shoveling snow and thinking about how the Orioles during the Showalter era were greater than the sum of their parts. (Except for the last year and one month, of course.)

Yeah one of the charts in the study shows the 2012 O's (Orioles Magic 2.0) as the extreme outlier in more than expected wins (15).

[scOtt]

4 hours ago, Beef Supreme said:

Just moments before you posted this, I was shoveling snow and thinking about how the Orioles during the Showalter era were greater than the sum of their parts. (Except for the last year and one month, of course.)

Sodium (didn't even post exploding beakers...)

Plus Chlorine

Equals Table Salt. Greater than the sum of it's parts.

In other forms...

Chemistry.

[OsFanSinceThe80s]

I was hoping this might be a discussion on Breaking Bad, but I guess that would be off topic.

Breaking Bad would be a perfect name for the 2018 Orioles team video.

 

[DrungoHazewood]

If I have some time I might read the paper.  But I'm highly, highly skeptical about any study that claims to assign a value in wins or runs to team chemistry.  

When they talk about 2012 Orioles team chemistry, we know that most of the delta between their Pythag and their real W/L record is sequencing and leverage of the bullpen, which was +13 WPA or thereabouts.  I'd like to know how the author disentangled chemistry from WPA.  Or if he even tried.  Or how he defines chemistry seperate from luck and leverage and other effects.

Very skeptical.  It's like the people who teased out that Leo Mazzone was worth +8 wins or +14 wins or something and should be in the Hall.  Until he came to the Orioles and appeared to be worth nothing at all.

Please don't take this to mean I don't think the players and manager and coaches jelling and pulling together have no effect.  I'm just very skeptical of the ability to pull that out of the data and assign values.

[UpstateNYfan]

Not too long ago a number of OH'ers felt team chemistry was basically unimportant. Although certainly not the most important aspect in putting a team together and not easy to measure, I think it's worth a few W's each year. Likewise poor chemistry is worth a few loses. Many here felt last years team would be in the running for a WC spot after the addition of Cobb and Cashner, predicting somewhere near 90 wins. Was this team really capable of winning only 47 games, or did other factors kick. 1) Bad start. 2) Last year for Manny, Jones (Did the team go into mourning after the slow start?) 3) Last year of contract for DD and BS. 4) Manny at SS?? Really?? This team had the perfect storm hit them, and everyone, including ownership had a role in its lack of success. 47 wins is not only Baltimore bad, it was historically bad. This team, this organization just went through the motions. You can call it bad atmosphere, chemistry, bad analytics, whatever. But, IMO, the team's output was considerably less than the sum of its parts.

Baseball is unique in many ways..One or two great players on a team does not make a champion caliber team 162 games, plus ST is a brutal schedule. Long road trips, shifting times zones, time away from family, etc. The team had better become family or better have some chemistry in order to survive. The 47 win team(organization) was impossible to watch, it needed a colonoscopy, and got it. I hope the '19 team enjoys every W and at the minimum display some sort of growth and a "chemistry" that make them watchable.

[weams]

1 hour ago, OsFanSinceThe80s said:

I was hoping this might be a discussion on Breaking Bad, but I guess that would be off topic.

Breaking Bad would be a perfect name for the 2018 Orioles team video.

 

Oh, they broke bad. 

[now]

1 hour ago, DrungoHazewood said:

If I have some time I might read the paper.  But I'm highly, highly skeptical about any study that claims to assign a value in wins or runs to team chemistry.  

When they talk about 2012 Orioles team chemistry, we know that most of the delta between their Pythag and their real W/L record is sequencing and leverage of the bullpen, which was +13 WPA or thereabouts.  I'd like to know how the author disentangled chemistry from WPA.  Or if he even tried.  Or how he defines chemistry seperate from luck and leverage and other effects.

Very skeptical.  It's like the people who teased out that Leo Mazzone was worth +8 wins or +14 wins or something and should be in the Hall.  Until he came to the Orioles and appeared to be worth nothing at all.

Please don't take this to mean I don't think the players and manager and coaches jelling and pulling together have no effect.  I'm just very skeptical of the ability to pull that out of the data and assign values.

I must admit the statistical methodology was over my head. But the authors did make an effort to "disentangle" various factors as you suggest. It seems the main concept they were focusing on had to do with how players complement (and to a lesser extent, compliment!) each other on the field.

[jsbearr]

Just a note...

This research is descriptive and not predictive.  The models may do very well describing what happened, but they may be constructed in a way that variables are actually a proxy for underlying mechanisms that may not be related to the variables that fit into in this model.  This method could also be impacted by linear events.  A string of success is more likely to lead to further success.  In other words, if you face a really good pitcher then you tend to be facing a really good pitcher.  But, if the three guys ahead of you are demolishing that overall really good pitcher then you probably are facing a worse pitcher than he normally is.  It looks like those aspects could be covered within the chemistry notion in this study.  The also simply things, such as ignoring any in-season manager changes or only using the most frequent position.

That said, there are a lot of interesting things they are working on here.  To be clear, the data is fairly terrible and this model seems to go out of its way to ignore a lot of data that we have that impacts corporate workforces.  Things like country or state of origin can play a big role.  Or minor league play.  Other things.  I tend to give greater weight to studies that try to characterize the individual.  I don't think this one really does that.

Anyway, it reads more like an undergrad capstone project than as anything that would be ready to implement.

[jsbearr]

And I would suggest that the Mazzone Effect study by...I forget his name...was done in a very simplistic way.  This model is going much further beyond what the guy from Kenesaw State did.  Apples and Oranges.

[theocean]

I think it's pretty difficult to assign a value to "chemistry" because you're assuming that all "chemistry" is the same. It's difficult to assign a numerical value to something qualitative.

That being said, it definitely plays a role. It is a game played by humans. To think that there isn't a human element to the game would be silly.