The Decision Strategist

Often when making a decision we focus on ranking our options and then choosing the best one as in pareto analysis. We start at the top and proceed to the next best option and so on.

Sometimes it is better to start at the bottom.

Why is this helpful?

We tend to consider options in two stages. In the first we eliminate options that are deal-breakers. These are options with factors that disqualify them even if their positive aspects are great. It might be a great job located in a city you hate, or a great car with an automatic transmission when you only drive manual.

In the second stage we look at the remaining options and roughly weigh their positive and negative aspects, but tending to focus more strongly on the positives.

It works well if there are only one or two options that are easily discernible as the best choice, but if several options could be great or are roughly equal, we can have a really hard time with this second stage.

By ranking options from worst to best, we are forced to explicitly consider the negative factors, which can help bring to light issues with some of the options which we previously thought were all equal.

Popularity: 18%

Vote for this post on: These icons link to social bookmarking sites where readers can share and discover new web pages.

I received my rejection notice from Ycombinator last night. I’m not too disappointed, mainly because I didn’t have high hopes for being accepted in the first place.

The Ycombinator application process is a great example of principal-agent relationships (touched on in Signal to Noise Ratio in the Consulting Industry). Like all situations in which assets are transferable but not distributed equally, the principal (Ycombinator) has the difficult job of determining who gets the asset, while the agent (the applicant) does their best to appear desirable.

The agent has a big incentive to lie as long as they won’t get caught, and the principal has to deal with not knowing the qualities of the agent with any degree of accuracy.

At Ycombinator their application process is different from most. Their main tactic for eliciting a true view of an applicant is to ask interesting and difficult questions. It is very difficult for an agent to fake worthwhile answers. It’s similar to the hiring methods of Google and the like, focusing on the results of the thought processes of the applicant.

Which is why I don’t understand the questions that my girlfriend has to answer for her med school applications. The questions are typically something like: “Describe how our mission statement fits with your life goals and motivations”.

Everyone hates trying to answer questions like this because they are so open ended. You can’t help that everyone is using exactly the same lines. How many ways can you say that practicing medicine and helping people are very important to you?

Worse, I suspect that the determining factor on these essays is how well they resonate with the admissions officer reading them, which can’t be strongly correlated with the abilities of the applicant.

I suspect that the reason universities ask these questions isn’t so much to determine good applicants, but to weed out the ones who can’t write coherently. It’s an elimination technique instead of a selection technique.

Maybe it also has to do with a difference in the number of applicants they are accepting also. Ycombinator only accepts maybe 10 groups, so they have the luxury of selecting the very best. Universities have to select a much larger group, and so perhaps it is more efficient to simply eliminate the lower 75% instead of searching for the top 2%.

Anyway, this started off being more about principal agent models, but I guess I got sidetracked into the idea of how to evaluate information when there is a poor signal to noise ratio.

There are a lot of other situations that fit the same principal-agent model. Biologically, consider males (agents) competing for the right to mate with a female (principal). Economically, people (agents) compete to be able to work at a company (principal) and on the flip side companies (agents) compete to have their product bought by consumers (principal). Politically, candidates (agents) compete for the favor of the public (principal). In education, students (agents) compete to be accepted to schools (principal).

It’s not surprising that we should find this relationship in so many places, at the heart of the interaction is the need to determine the qualifications of an entity and the entity’s desire to be selected.

Enjoy the weekend.

-zot

Popularity: 14%

Vote for this post on: These icons link to social bookmarking sites where readers can share and discover new web pages.

By now you must have seen the news discussing Radiohead’s new album and their decision to shun record companies and release their record by themselves. In fact, they’ve decided to release it online and let people decide how much money to give them. Here’s the link.

Trent Reznor’s doing the same thing.

And then there’s things like Songslide, which lets bands add their music and users pay whatever price they like.

It’s more evidence of changes in the music industry and the failing tactics of the RIAA.

But what I find really interesting are the ways in which bands releasing their music for free are making use of referencing and priming to encourage people to spend money. Radiohead puts an empty price box in which you can put how many GBP’s you want to pay for the album and very little else. Not a lot of an attempt to make people pay more, but other artists do things differently.

There is one artist’s website I’ve seen that lets you pay a price you choose, but explicitly puts the ‘average’ price as one of your options, and tells you what that average is. It’s a great idea for two reasons.

First, having the average that people have given changes the reference point from $0 to a much higher value (I think it was usually around $10). Whatever you were considering paying gets referenced to this average, with the accompanying feelings of guilt (or getting a good deal) when you pay less and the possible pride in paying more. I could see the average backfiring by causing people who would pay more to just opt for the average, but my suspicion is that people not paying much of anything is more of a problem.

Second, it explicitly engages moral self judgment. If you pay significantly less than average, you are more likely to feel cheap or selfish. This self-watching, like pictures of watching eyes, has real effects on peoples behavior.

There isn’t any data on how much people pay for albums in this kind of setting yet, but it’d be very interesting to look at differences in payment based on how the interface for downloading is set up.

I’ve been trying to remember which artist does this, so if any of you know please tell me.

-zot

ps - I really like In Rainbows so far.

Popularity: 9%

Vote for this post on: These icons link to social bookmarking sites where readers can share and discover new web pages.

Some loose research by Griffin and Tversky (Overcoming Bias) suggest that we can predict what choice we will make with about 96% accuracy. It’s based on people facing a decision about a job opportunity and predicting what they will choose.

There are some problems with this. Mainly that you may be fairly sure of your choice before you make the decision, and so of course you are accurate when asked to predict what you will choose. For choices for which you are truly conflicted, I think the rate of success may be lower.

Yet it still highlights that we make decisions unconsciously very quickly, and the rest of the rationalization process is unlikely to cause us to change our choice.

If decision making techniques don’t really work and we are predictable enough that we can predict our own decision, perhaps the best decision making technique of all is to imagine our future after the decision and try to predict which decision we will make (I know there are problems here, bear with me).

Sounds easy right?

In some cases it is. I’m trying to restrict my chai drinking to one time a week, but will I get some tomorrow? I’m betting yes. Will I sleep in instead of getting up early? Almost certainly.

But will I choose to go to the Peace Corps next summer? That one is much harder. If I had to bet…I’d say I will go, but if you asked me again in an hour, that might change.

Still…there is something to be said for decisive action based on first impressions. Great military leaders are almost always described as being very decisive.

For those of us in less immediate situations, a greater reliance on our first choice may still be a better process than being mired in doubt.

Something like Inkling Markets could be used to set up a prediction market for your actions…it’s not to far off from what I’ve been trying to create myself.

-zot

Popularity: 8%

Vote for this post on: These icons link to social bookmarking sites where readers can share and discover new web pages.