Wednesday, December 14, 2005

What is wrong with all these sub-optimal teams! Let's fix it with some experts!

Our friend 'The Amateur' has posted on the Olympic Trials for Curling as well as the general subject of Olympic Team Selection. It is a valuable and interesting contribution to our ongoing debate. You can read his post as linked above, but as he imputes some notions to me, I am putting some effort into responding.
He describes the issue as follows:
The British approach for 2006 is to have a "bunch of administrators," as Alan puts it, assemble the "best" foursome from the members of the best rinks in the UK (read Scotland in this case).
I am not sure what the scare quotes are for. It IS a bunch of administrators making the decision in the UK, not any of the players who will be on the ice. The administrators are clearly picking the team they will pick because they think it is the best one to pick - else why have adminstrators around to pick that best rink? My concerns about his framing of the discussion are key to my views on this.

He then, rather oddly distancing himself from his own position, says:
Without defending the specific process being used in Scotland, I have argued that creating an all-star foursome might be, in some circumstances at least, the best way to optimize the performance of your team.
OK let me agree that if we were assembling the Jamaican curling team I would feel we should look around to try to find four reasonable shotmakers in the country. But let us recall this is the UK which won the gold medal in the last Olympics, with a rink I do not believe was assembled by those who know better, but rather by the players.

He goes through a number of points about how rules allowing composition of a team to change can get into a battle with our understanding of what the team is, points all handled earlier in this post (and btw if you go look you will see Amateur scooped us all back in May - that long ago Gushue had named Howard as his alternate).

Imputing some more views to me he then writes:
Now this is clearly an example of the type of "free market" solution that Alan advocates; Gushue recognized that his team needed to be stronger, looked around for a top-quality curler who was not otherwise committed and would be a good fit with his team, and brought him on as his fifth. After experimenting with different arrangements, he optimized his team's performance by bringing Howard to the trials as second, and now they're all off to the Olympics.
Well I think it is good that Gushue could reconstitute his team to a degree. But hey, I would not have minded rules saying he had to play the team he qualified with. So I am not sure about this free market stuff, especially again with those arch scare quotes. I would object to rules whereby some appointees of Sport Canada tell Gushue he has to replace his second with Russ Howard.

And then what seems to me an odd point:
Of course, I would have to counter that Gushue's rink might have been even better had he switched out even more of his players, and if he had been able to select players from other qualified teams instead of just picking up "leftovers" like Howard!
Hmm - I actually think that Gushue's rink might have dominated even more without Howard! This is a science experiment we don't get to do; but it is clear there were moments during the week where they suffered a bit from not being a fully baked team. These speculations are speculations. The team Brad Gushue brought to the rink won.

Further speculating about what might have happened had there been a free-for-all with rules allowing teams to get anyone they wanted, he goes on:
I guess Alan might argue that competitive pressure would be enough to lead to the formation of an "optimal" Canadian team. I think that the players have many sentimental and irrational reasons not to be performance-optimizing selectors; I think that outside experts (national team coaches, for example) are more likely to be unbiased evaluators.
Hmm well no. There are industrial organization issues here. One might argue (don't worry about facts here) that GM's financing organization, Mercedes' engineering, Ford's quality assurance, and Toyota's manufacturing are the best in class, so why are we wasting our time allowing all these sub-optimal organizations to continue to confuse us - let us get a team together to unite these best performers. My guess at the result? You don't even have to ask.

The mere notion that The Amateur describes of the 'optimal team' flies totally in the face of how I think the world should work. His blatantly managerialist view is that there are experts who can sort out how things should be done - if only they were in charge we would have the best solutions. We would have the 'optimal' team. I will bet that in some context he is one of the experts.

No current car manufacturer is optimal in his sense. But I am very certain that any expert attempt to concoct one optimal solution will collapse in the face of all that cannot be evaluated by the experts - emotions, history, human relations, and so much more. I am much happier knowing there are several companies out there fighting as I think it gives us (those who buy cars) a better solution than any directed one would. It is not 'optimal' in the sense of the experts, but is almost certainly, to my mind, better than what the experts would produce in real life. And in my day job it is no different - I am on a crack team that is one of a handful doing what it does and competing with the others; I know full well that the process of us constantly trying to top one another is producing much better products and giving our customers something no expert-assembled team would even feel motivated to produce.

What we are now sending to the Olympics is a team that showed that on the day (well, the week) it could prevail against the best the rest of the country could throw at it.

Let us also not forget that the margin that separated Gushue and Stoughton was minimal (really!!). Sports, like business success, can be pretty arbitrary in outcome. Only a managerialist thinks there is a best solution that can be nicely determined by experts.

And there is no reason to assume the UK team will not win the Gold again this year, despite their selection process. :-)


At 12/19/2005 11:58 PM, Blogger Amateur said...

Well Alan, my education continues. The end of your post has clarified your position, and it is somewhat different than I assumed.

(I will try to write this whole comment without quotes and see if I can manage it. It will take discipline.)

I am not comfortable defending the specific process being used in Britain because it looks -- from here, seen through the press -- that they are making a cock-up of it. I guess I believe that there is a right way and a wrong way to do this kind of thing, and it looks like they are taking the wrong way. (Is this blatantly managerialist of me?) Having said that, it is a tricky problem, and I do not have any real concrete suggestions for improving the process. Perhaps if I was a curler of any description I would have some good ideas.

I am also not entirely comfortable with the economics analogies (as you can tell) but I am learning something as I go. I have a half-baked idea here that may or may not provoke some further conversation. You draw an analogy to your own work, where your team is one of several competitive entities. What about within your organization? Do the sub-teams or individuals within your team compete against each other, unmanaged? If there is a task that needs doing, does every team member just produce their own deliverable, and then you pick the one that best meets your requirements? Or does the boss try to manage the work, assigning the task to the person or team that (s)he thinks will do the best job?

Coming back to curling, if the objective of the CCA is to crown a national champion, then of course there is no need to identify the optimal team in any absolute sense, and in fact that would not really be to the benefit of the customer (the fans). In this sense, the curling teams are analogous to your work team and your competitors.

When selecting a team for the Olympics, though, the objective is different. In this case the CCA has a responsibility to try to send the best possible curling team that they can. OK, perhaps some would dispute this statement, but if you asked them they would say that it is their goal to maximize their (Canada's) chances of winning the gold medal. At this level, the CCA and the other national curling associations are analogous to your work team and your competitors; the various rinks are then members of an organization that is in open (and essentially unmanaged) competition with other similar organizations. So within the CCA, wouldn't there be some merit to managing the work done by the members of the organization? Couldn't a good management team organize the team members in a way that is better than the way that the members might choose to organize themselves?

As I said, this idea is not yet very well-formed but perhaps your response will help me solidify it.


Post a Comment

<< Home