www.bestboy.org

**options** · 10-10-2002, 01:35 AM

Originally posted by jberryhill
You are failing to normalize your integral, son.

Bad attitude Ph.D.

**GeorgeK** · 10-10-2002, 02:32 AM

Huygen's principle ignores risk.

To modify your approach using economics/finance principles, I'd take the expected payoffs, but then discount them based on the riskiness of the payoffs (i.e. the required rate of return).

e.g. if you have an expected payoff in one scenario of $100, but it's in 5 years, and the relevant risk-adjusted interest rate is 40% per annum, then the present value is $100/(1+0.40)^5 = $18.59.

Coming up with the risk-adjusted interest rate is non-trivial, though. You'd want to find what comparable cash flows are being discounted in the marketplace, if you want to use a market-based approach. There aren't too many "pure plays" for domain name values out there that are traded on an exchange (Communicate.com might be one of them, though -- see:

http://finance.yahoo.com/q?s=cmnn.ob&d=t

Probably a high rate of return is required, given how risky these cash flows are, and one's inability to hedge them very well.

**options** · 10-10-2002, 05:21 AM

Originally posted by GeorgeK
Huygen's principle ignores risk.

Right. But in original simplified example there was no risk.
Everybody would buy sex.com for $100, thus 100% probability.

[B]
There aren't too many "pure plays" for domain name values out there that are traded on an exchange (Communicate.com might be one of them, though -- see:

http://finance.yahoo.com/q?s=cmnn.ob&d=t

Exactly 433 companies on finance.yahoo.
See:

http://finance.yahoo.com/l?m=US&s=.com&t=

**jberryhill** · 10-10-2002, 12:16 PM

There are a whole bunch of potential prices between, say, 10 dollars and 100 dollars. Do you understand that your method is totally dependent on the number of points on the demand curve that you choose to use in your calculation?

Repeating the explanation doesn't make it a relevant method. Since you didn't seem to follow my example, then let me use your own example.

You say:

-------------------------
Prizes.....Probability of winning the prize
----------------------------------------------------
$100.......................0.7
$200.......................0.2
$1000.....................0.1

Note that .7 + .2 + .1 = 1.

According to Huygens, a fair price would be:

fair price = (.7 * 100) + (.2 * 200) + (.1 * 1000) = $210

Note that by "prize", I mean selling the domain at the given prize.
--------------

Now, let's take a look at your chart. Why do you live in a universe of only three possible prices? If the probability at $100 is .7, and the probability at $200 is .2, then we can roughly interpolate that the probability at $150 is .45, right?

Now adding a data point of $150.........0.45 to your chart should NOT change the value of the domain name, since that data point falls on the same curve as your other data points. Nor should it change the probability of selling the name at any of the other prices on the chart. You can say ".7 + .2 + .1 = 1" for your three data points, but are you also saying there is a zero probability of selling it *at any other price*? Are you with me so far?

But, by simply adding another data point, your "fair price" now goes up by (.45 * 150).

Heck, what is the probability at $200.01?

At $200.02?

At $200.03?

Why don't you add contributions from those prices as well? Are they not possible prices?

You are getting an answer which merely depends on the number of points you pick. Huygen's (other) Principle relates to a situation where you play a game with a finite number of distinct outcomes, and where you know that you are going to get something.

**morel** · 10-10-2002, 05:28 PM

First, I need to clarify my definition of the word "prize". I am defining "prize" to be the maximum dollar amount that you can obtain for your domain. So if the "prize" of bestboy.org is $300, it means that you are able to sell this name for $300, but are unable to sell this name for $301.

As options pointed out, the probability of finding a buyer willing to purchase sex.com for $100 is 1. However, according to my definition, I would assign a probability of 0 to a $100 prize for sex.com because I can guarantee you that someone will pay more than $100 for the domain. Likewise, a $1000 prize for sex.com also has probability 0.

I should also clarify that I am not using any sort of demand curve in my analysis, I am merely treating a domain the same way you would treat a lottery ticket -- as a game to be played.

Having said these things, I agree with some of your points.

Originally posted by jberryhill
Why do you live in a universe of only three possible prices?
You can say ".7 + .2 + .1 = 1" for your three data points, but are you also saying there is a zero probability of selling it *at any other price*?

Yes, I am saying that; it is true that I am simplifying â€œthe gameâ€ immensely by only using three possible prizes. $150 could also be a possible prize (recall how I am defining the word), and so is $250. $0 could also be a possible prize. However, I think that using only three possible prizes is sensible for the point I am trying to make.

Very often on this board we hear that domain xyz.com is worth $1000. The appraiser will justify this price by saying that there is probably someone out there that is willing to pay this amount. Unfortunately, the domain name market is far from perfect, and that potential buyer may never find out that this domain he is willing to buy is for sale.

The point that I am trying to make can be summarized by this exaggerated example:

* If there is one person in the universe that is willing to pay $1000 for domain xyz.com, and there is only a 1% chance that you to will meet each other;
* If everyone else in the universe will not pay anything for domain xyz.com,

then I would argue that the worth of xyz.com is not $1000, but rather (0.99 * 0) + (0.01 * 1000) = $10.

-------------------------------

I would also like to clarify two other things:

But, by simply adding another data point, your "fair price" now goes up by (.45 * 150).

If you add another possible prize, then you have to make sure that the total probability equals 1. (Notice that .7 + .45 + .2 + .1 = 1.45 > 1).

Huygen's (other) Principle relates to a situation where you play a game with a finite number of distinct outcomes, and where you know that you are going to get something.

This theorem of Huygens can also be applied to a game with an infinite number of prizes, as long as there is only a countable infinity of prizes. Surely though it cannot be applied to an uncountable infinity of prizes. Also, you donâ€™t have to get a prize in order for Huygens to apply. For example, buying a lottery ticket is a game, and you can calculate the fair price for that, but the prize with the highest probability is $0.

**Ed30** · 10-10-2002, 05:55 PM

Ed's Scientific Theorem - bestboy.org = 0/0 = no possible uses = zilch.

**Fearless** · 10-10-2002, 06:04 PM

Just think, this all started with a crappy domain name.

Now put the same analysis to work on Motherboards.com.

**NamePopper.com** · 10-10-2002, 06:11 PM

Originally posted by gregr
Just think, this all started with a crappy domain name.

I was thinking the exact same thing. Fascinating.

Originally posted by gregr
Now put the same analysis to work on Motherboards.com.

Oh god.....please.....NOOOOOOoooooooo!!!!!

President Gringo · 10-10-2002, 06:17 PM

I think you guys are just showing off now :razz:

**jberryhill** · 10-10-2002, 10:35 PM

"If you add another possible prize, then you have to make sure that the total probability equals 1. (Notice that .7 + .45 + .2 + .1 = 1.45 > 1)."

That's my point. How did recognizing that there are 100 pennies in a dollar manage to LOWER the probability at the other points? If the P($200)=.7, then why would it go down from .7 just because you added another point?

You *are* trying to do calculus, but what you are doing is an integral curve fit with large rectangular approximations. If you want to find an expectation value, then yes, you integrate a *cumulative* probability curve, and not match probabilities with discrete points on the curve.

The BEST thing I can suggest that you take a look at is a book entitled "Cognition and Rationality In Negotiation" Neale & Bazerman (1991). It's a thorough but approachable treatment, and even a nodding skim through it is extremely helpful for anyone who has to negotiate.