Oog Robot: July 2009

2009-07-23

A Test for Rental Scams

I'm looking for an apartment, and I seem to get an awful lot of e-mails from people that have beautiful apartments for low, low prices but for whatever reason they won't show me the places. Or they say they'll meet me and when I show up a confused woman answers the door and says, "No, there is no apartment for rent here."

Today I'm going to describe to you how you can check if you are being scammed. Go into your email client and turn on all of the headers (in Thunderbird, this is View->Headers->All). You'll see a giant chain of "received" headers that look like:

At one end of these headers will be your e-mail provider, while at the other end will be the IP address of the computer that sent the email address:


Received: from SRV502.tudelft.net ([131.180.4.18]) by ...
...
...
...
Received: from [41.205.182.125] by web111904.mail.gq1.yahoo.com via HTTP;
          Wed, 22 Jul 2009 13:09:45 PDT

Sometimes the address of the originating computer is listed a bit differently:


Received: from BLU116-W26 ([65.55.116.72]) by ...
X-Originating-IP: [62.173.35.6]

The person might have masked the IP of their computer, but this is complicated and most scammers are idiots. Anyways, you can take this IP to a handy geolocation look-up service and find out the originating location.


IP: 41.205.182.125
Host: dial-pool07.ab.starcomms.net
Country: Nigeria

Every now and then the country won't be listed, but the host name you get back has information you can use to determine the country. For example, 62.173.35.6 resolves to 35-6.rv.ipnxtelecoms.com but the geolocation service doesn't give a country for this IP. ipNX Telecom has its headquarters in...Lagos, Nigeria.

I suppose in the future I'll have to screen any email about apartments. Sometimes I really hate people.

EDIT: Your email client actually uses the same method I describe above to label some emails as being probable scams. Usually IP blocks assigned to certain ISPs and sometimes even entire countries are labeled as suspect.

Sentences that Make Me Laugh

If we ignore (thing one), (thing two), (thing three), (thing four) and assume (thing five), then the problem becomes tractable using only elementary results from field X.

Sadly I wrote the sentence...though it was followed by:

Subsequent sections will discuss the effects of (thing one), (thing two), (thing three) and (thing five), cumulating with results from a simulation including all these effects. The authors plan to revisit (thing four) in later work.

Planning to revisit something is a total cop-out.

2009-07-22

CEOs and Pay

While I was home the issue of executive compensation came up several times. I'm a person who thinks that a CEO could easily deserve tens of millions a year or more. The main reason behind this is two-fold.

First, I've worked at a company that has undergone restructuring at a management level. From personal experience, small changes in strategic direction have a massive impact on people at the bottom of the "job-food" chain. It is easy to lose months, probably even years of work because someone above you made the wrong call. Since a CEO helps determine the strategic direction of the company, it seems reasonable to me that the pay of the CEO should scale with the sum pay of every other member of the company.

Second, there are a few studies which show a CEO's personal circumstances can have a percentage-point effect on a company's performance. A great example is a study of the effect on profit by a death in the immediate family of a CEO. Excerpt:

Sorting by the number of children we find the biggest effects on firm profitability in cases where the CEO only has one child. Specifically, one-child death shocks correlate with a 5 percentage point decline in firm profitability irrespective of the age of the child.

The study goes on to show that the deaths of a spouse or child are significant events for a firm. I think the study underscores the importance of choosing the correct CEO. Though the study doesn't show the variation in profit during standard circumstances, to me it seems like a reasonable conclusion that variations in a CEO's ability to carry themselves through tough times will have a large effect on firm profit. For a company with hundreds of billions in profits per year, like an oil company, the CEO's personal circumstances could have an effect in the billion dollar range, meaning if the CEO themselves only received a fraction of this pay it would still be in the tens of millions range.

My math is a bit fuzzy here, but as always you're welcome to disagree in the comments and point out any mistakes I've made.

2009-07-21

It's All About the Teachers

I've been on vacation for the past, well, month, and I had a lot of interesting conversations. Aside from the very personal ones, the most engaging ones were on the topic of teachers. I had lunch with an old class-mate who is now a teacher at a high school. We talked a little bit about evaluating teachers. There isn't really a coherent point to this post, but if there was it would be that there is enormous room for improvement in the way we allocate resources to educate and invest in our next generation.

There is a lot of interesting quantitative data in how we invest in students. My class-mate and I talked about how too many smart people opt out of teaching. As the feminist movement has matured, a lot of really smart women no longer become teachers and the quality of teachers has suffered. I think the take-away is not that the feminist movement is bad, but rather we as a society don't have a great understanding of the importance of good teachers.

Another subject we talked about was the politics of teaching. If you haven't heard it, I suggest the TED speech of Bill Gates, the education portion begins at about 8 minutes. He talks, amongst other things, about how it is illegal in New York to use performance-based data to evaluate a teacher's performance. Coincidentally my old class-mate was for qualitative data like committees but against quantitative performance-based data. She claimed the quantitative data would have too much variance. I think I dissuaded her from this stance, but I do agree that quantitative data is not perfect. Any one metric can be taken advantage of. As an engineer, however, I have more faith in a well-designed system with numbers than a well-designed system with committees.

Finally, we went over the cost of education at private schools. There is a huge debate about whether private schools outperform public schools. The debate is quite contentious, especially when trying to account for socioeconomic and ethnic diversity factors. She said that public schools are critical to integrating immigrants into society. I do not have a great link to summarize this stance, as I haven't seen anything like this online - please comment and send along a link if you have one. I said that I thought public schools were wasteful due to the political pressure and lack of transparency into the teachers.

Anyways, I think the high level take-away is that evaluating teacher performance is a complex problem that has to touch politics and statistics, though a lot of information suggests there are inefficiencies. It would be interesting to attempt to evaluate schools, but creating a non-partisan report that accurately represents the facts seems to be near impossible. The only thing I know for sure is that I don't have a clue how parents choose schools for their children.

2009-07-05

Big Pimpin'

Sometimes it helps to remember your fashion roots.

Only a select few can go pantsless with combat boots and diapers. I am one of these few.

Sometimes, though, you've got to break out those whale pants.

When I was big in the 80's, I wore jump suits for a while. I think my life-long relationship with red started around this time.

I also tried the grunge scene in the 80's. Or the potato chip scene. They were kinda the same thing.

After the grunge scene I classed it up a little.

Speaking of classy, nothing says classy like gray shoes and a bow tie.

Sometimes it was hard being with other people who didn't get it. Going as a mouse to Halloween was so last year. Being a pre-schooler was where it was at.

Oog Robot