Wednesday, May 05, 2010

Provoking Thoughts: Expected Values

Recently, we've seen an event draw the ire of some folks. And naturally, there are perfectly logical arguments made by people on both sides. And naturally, I think that this whole thing is hilarious.

I'll just put it like this. If you DON'T ask Dez Bryant if his mother was a prostitute, then the terrorists win, and we shame the memory of those who died in 9/11. Bear with me... as I present some things for you to consider.

The first is the story behind the Jeff Ireland - Dez Bryant's Mom's thing. Where basically, as the Miami Dolphins are doing a pre-draft interview, and there's some exchange, where apparently (and this is not clear), Dez Bryant (the player being interviewed) tells Jeff Ireland (the Dolphins' GM) that his dad was a pimp, and that his mom worked for his dad. So, Ireland decides to ask Dez if his mom's a prostitute. Seems logical enough. Sadly, there were plenty of other questions that weren't asked. The question that I'm curious about is, "Is your dad friends with Lawrence Taylor?"

So there are all sorts of questions flying around. Should Ireland have asked this? Should Bryant be offended? Who should apologize to whom? And embedded in all of this is the underlying question of should this even matter?

I can see Dez Bryant being upset, and that Ireland maybe apologizing for asking an inflammatory question. After all, it's a small league, and you never know when the Dolphins may end up bringing in Bryant for another workout in a few years. And if I'm Ireland, I probably send a little letter to Dez Bryant's mom to apologize, as well. But what I think is hilarious is Dez Bryant's mom getting all righteous on us and demanding an apology. Now, this is a woman who has had a few run-ins with the law - she was arrested 12 years ago for selling crack, was arrested last April for selling crack, and then arrested last August for possession. And along the way, she decided that she was going to be a lesbian, much to the chagrin of her son. The lesbian thing isn't really pertinent to my point, but just makes me enjoy the story more. In any case the problems with this is that any time you have a story that makes headlines, the idiots come out of the woodwork.

Now, the idiots on the fringes, I don't worry much about. For example, you have this moron taking 1 or 2 sentences out of context, and vilifying someone who dislikes someone's choices. But happily, idiots don't make it on to my radar. But then you get other idiots who have a bigger stage. When you have folks like Rick Reilly (who have access to a huge, global stage) writing an article that has roots in something between intellectual laziness and downright stupidity, then you have idiocy that I'm worried about. Because of the way the article is written - pretty smooth and apparently unassuming, Reilly's argument becomes attractive because it's so simple and so appealing. And listening to stuff like that, without thinking further, is generally a recipe for disaster.

The reality is that we are very much influenced by the folks around us, and the presence (or absence) of parental influences is crucial to proper development (intellectual, emotional, whatever). And we are very much ensnared by the sins of the father, much like we are bolstered by the successes of the father. The point that I hope that Reilly was trying to make was that the failures of a parent are not guaranteed to lead to the downfall of the next generation, and that genius can arise from meager beginnings. We should intuitively know that, and can see it all around us. However, the more interesting question is how much should we temper hope with prudence. And in this question, not everyone is created equal.

Suspend disbelief for a second, and make the assumption that decision making in real life can be described with fantasy sports analogies. I've written the past that winning fantasy is all about making sure that you don't screw up the first few rounds, and then get lucky in the late rounds or in free agency. Basically, you need to make sure your sure things are sure things. This reflects this 80-20 rule that you hear people throw around all the time (and many folks argue that it's more like the 95-5 rule). Basically, it boils down to something like this: 20% of the things you encounter have 80% of the impact. 20% of the people in an organization do 80% of the work. 20% of the events of your life drive 80% of the outcome. 20% of the people control 80% of the wealth (or commit 80% of the crime). You get the idea.

The next step is the thought that not everything should be equal, because the gravity of each thing is different. Certain decisions require much more intense scrutiny. Again, this should be intuitive. It makes more sense that for the decisions that are in the 20% (which lead to the 80% of utility), you want to dig deeper. But I'll argue that the fundamental question that you should ask is different.

In science, we talk about Type I and Type II errors. Basically, you can screw up in one of two ways. The first, a Type I error, is a false positive. You have an alarm system, the alarm sounds, but there was no intruder. The second, a Type II error, is a false negative. Here, you have an alarm system, the alarm doesn't sound, but there was an intruder. In this instance, a Type I error leads to some annoyance, but a Type II error may make you dead. Evolution has taught us to be a little skittish at times, mainly so we avoid fatal Type II errors. Think about it in terms of hunting. Basically, thinking that a deer is there, when it isn't (a Type I error) makes you go hungry. Not noticing that a tiger is behind you when it actually is (a Type II error) makes you tiger food.

As we think about the case of Dez Bryant, we're looking at a player that was going to be a first round pick. The investment would have been somewhere around $8-$15 million dollars in guaranteed money. You don't want to brick your first round picks. Teams that get production out of their draft picks, do well, can manage the salary cap, and build teams that are successful year to year. Teams that brick their draft picks end up having to get free agents, wrecking their salary cap, and need to get lucky to win. The impact of first round picks are even more amplified because of the guaranteed money.

At this stage of the game, you assume that if things go well, the players you're looking to get are of roughly the same talent and will contribute roughly the same. However, if something goes poorly, you could submarine yourself. Think about outcomes on a 0 to 10 scale. First round players, could tank (and score you a zero), or they could play reasonably well, and perform somewhere between a 7 and a 9. A Type I error, may cause you to miss out on a guy that will perform at a 9, but if you avoid a Type II error, the worst you'll do is a 7, and you could just as well end up with a 8.5 or even another 9. On the other hand, making a Type II error, missing out on the fact that a guy may end up in jail, or whatever, could result in a 0. In this case, it's more important you you avoid the 0, than it is to make the pick that gives you the optimal expected value. This is a case where variance is more important than the mean.

In the choices that we make on a day-to-day basis, more of them are ones where the downside is minimal. These are the situations where we can be a little risk-seeking, take some chances and try to get a little margin. This is trying a new beer, ordering pheasant at the restaurant, or seeing some foreign film. But there are situations where we need to see the entire picture a little more clearly. These situations may be somewhat exotic, like eating fugu (the poisonous blowfish that kills roughly 5 or 6 people per year in Japan) or street racing (where about 100 or so people per year die in the U.S. each year). But they can be much more mundane, as well.

Think about something like screening x-rays at the airport. If you make a Type I error, you slow down the queue, and maybe bring some poor schmuck in for a strip search and/or waterboarding. If you make a Type II error, the terrorists win - literally. Now, statistically inclined folks generally make the case that people tend to spend too much time worrying about the tiger, and that they would gain more utility by trying to find more deer. Or, in this case moving the lines at the airport along a bit quicker. And that's definitely a valid point, for certain situations, but finding that sweet spot is a bit beyond this discussion. And there's something intuitively wrong about worrying too much about the productivity or personal convenience lost when you have to wait in line a little longer, when that tiny bump up in productivity or personal liberty results in being less likely to catch that 1 in a million event.

So, at the end of it all, what I'm saying is this. If you DON'T ask Dez Bryant if his mother was a prostitute, then the terrorists win, and we shame the memory of those who died in 9/11.

-Chairman

No comments: