Saturday, May 04, 2002

Spam Tax?


I agree in principle with the idea of charging for email. In fact, my twist on it is to give recipients the right to waive the charge. I would waive charges on all emails that I wanted to receive, and refuse to waive it for spammers.

However, I think all such ideas tend to run afoul of the principle of keeping the Internet simple. I am persuaded by the Internet's hippie fringe that a simple Internet is best.

What I don't understand is why email filters aren't better. I would like a filter that learns while I use it. If I had an email program that watched what I deleted without reading vs. what I read, and then applied elementary statistical modeling, I would soon have an effective spam filter.

Deep Linking Restrictions?


Speaking from my experience running the Homefair site, I cannot believe that you would send lawyers out to do this.

I have more to say here


Friday, May 03, 2002

Thursday, May 02, 2002

Wednesday, May 01, 2002

A new post on Live From the WTC


In which I predict a dim future for the stock market. Should be viewed if only for the fun of watching my ham-fisted attempts to explain a DCF in one paragraph.

A Conservative is a Liberal Who Has Been Mugged...


...by the nonprofit sector. If only this would happen to Doc Searls.

Better Environmental Policy


Two recent N.Y. Times op-ed pieces point out that the Bush Administration's "Clean Skies Initiative" represents an improvement in environmental policy. Paul Krugman says so grudgingly, but then catches himself and recycles the usual innuendo that the Bush administration is incompetent and beholden to evil corporate interests. A. Danny Ellerman and Paul L. Joskow just stick to the core issue of whether the new proposal is better than the policy it replaces.

Tuesday, April 30, 2002

Private Sector Encroachment


The private sector is encroaching on what rightfully belongs to the government, according to this new book, which was recommended by Doc Searls.
Government R&D laid the groundwork for some of the most significant innovations in computing - the original Internet architecture and protocols, e-mail, the Mosaic software that gave rise to the Netscape browser, among others -- but these investments have essentially been privatized and recast as the singular product of entrepreneurial vision.

And we know what privatization leads to, don't we: PROFITS (boo, hiss!); PERSONAL WEALTH (horrors!); ECONOMIC GROWTH (how threatening!). What we need to do is shrink the private sector so that THE PEOPLE can be rich(hooray!).


Basic economics is not all that difficult, but people would prefer to wallow in juvenile preconceptions. Hence, a book like this likely will be a hit.

More on the IPO Game


I got three emails related to the IPO game.
  1. Steve Kuhn agrees with my conjecture.
  2. Tom Maguire says more:


    ...you are just baiting me. I know that it is painfully obvious to you that this is a variation of the old martingale /
    anti-martingale trading strategies. And I know that you realize that an anti-martingale approach applied to a game with a positive expected value results in a huge probability of a small win.

    er, maybe I should know that, and I'm sure that some of my former classmates know it, but the truth is that Tom kinda one-upped me there.


  3. Finally, the Dreckmeister himself mentioned a Harvard Business Review article (summarized here) about the distribution of income. The article purports to explain why that distribution is so skewed--lots of poor folks and a few very tich folks.

    I don't have the patience to follow the math in the original research paper (maybe Tom can handle it), but it looks suspiciously to me like the Paulos IPO game. That is, if people's wealth comes from a sequence of bets (investments), and they make bigger bets as they get wealthier, then you can wind up with a Paulos-game style result. The lucky bettors wind up very rich, and most bettors end up with nothing. There are hard-core stock speculators and serial entrepreneurs who fit that model. The story line "he made and then lost a fortune, then made it back, then lost it again" sounds to me like the Paulos game.

Monday, April 29, 2002

Good Krugman column about why the 5.8% first quarter growth numbers might not be that impressive. As he writes, the stock market reacted badly to the growth figures. This means either, as Krugman points out, that the stock market doesn't think that the high growth will continue or that the stock market predicted the high growth and had already factored it into its valuations. If you are seriously interested in economics, you should read the column to learn about the importance of inventories.

Comments on MHD's simulation of the Paulos IPO game


In the game, you make a sequence of bets. At each turn, you bet everything. Each turn like a coin flip, where heads means that you win 80 percent and tails means that you lose 60 percent.


Winning 80 percent multiplies your stake by 1.8, and losing 60 percent multiplies your stake by 0.4. So, one win and won loss together gives you 1.8 times 0.4, or .72. If you play the game 2n times, and you get an equal number of wins and losses, your stake will be (.72)^n, which approaches zero as n gets large. In other words, the modal outcome is that you lose your shirt. In fact the probability of not losing your shirt gets close to zero as n gets large.

But the expected value of the game is still positive! That's because you win so much when you have spectacularly good luck. For example, a sequence of 1000 turns where you manage to win 625 times gives you over $11 million.


In the multi-turn game, a player can make the distribution of outcomes more attractive by having a rule that reduces the size of your bet whenever you're ahead. This reduces your winnings from a hot streak, but makes you less likely to go bust. My conjecture is that with the right rule (which might involve betting much less than $10,000 if you start with a $10,000 stake), there is a way to play the game that gives you a very high probability of coming out ahead.



Caroline Baum cites data suggesting the productivity boom is continuing:
the profit outlook is brightening, according to economists at Salomon Smith Barney. Using first- quarter GDP data and previously reported statistics on hours worked, they calculate that productivity, or output per hour worked, rose more than 7 percent in the first quarter. The projected 5 percent decline in unit labor costs augurs ``an expansion in margins, allowing sales to flow directly to the profit column,'' the economists wrote in Comments on Credit. ``Economic profits rose by an estimated 30 percent for the quarter,'' they said. (The Commerce Department's measure of economic profits measures income from current production with an adjustment for the value of inventories.)
Two years ago this would have sparked a huge rally.