"As far as I'm concerned, making it 'hard enough' to where it's more cost effective to resort to old fashioned brute force is just as good as not putting it in."
In principle, I agree. This probably isn't as big an issue as some think. But when you read some of the comments in the article, it is clear that the people making the decisions don't even understand most of the basics of what's going on. For instance, from TFA...
"Other comments suggested that the passport data should be encrypted. The passport data on the chip does not require encryption in order to be secure and protected. It is the same data that is visually displayed on the passport data page."
"The same rules apply, say, for driving. If you want to drive other people around, commercially, you generally need some form of taxi driver's license."
Ironic you should use this example, as I was just thinking about taxi drivers.
Suppose I decide to sell an old painting at my local brick-and-mortar auction house. I grab my painting, and since I don't have a car, I get in a cab, and go to the auction house to drop it off for auctioning.
Question: Does my cab driver need an auction license?
What if I'm handicapped, and the cab driver has to carry the painting into the auction house for me?
What if I just hire him to take the painting to the auction house for me, so I can stay home?!
At which step here do we decide that this cab driver needs an auction license? Because that last analogy is almost exactly what the Ebay consignment guys are doing.
"I don't feel like doing the probability math, but I would safely assume that the odds very good that there is at least one unique second in the day where 165+ messages are simultaneously."
Surprisingly even just looking for a peak over 165 doesn't change the conclusions too much.
We can use the Poisson distribution to model random arrivals (like text message requests). For example, if the average number of messages sent per second is 138, but they arrive randomly over time, then there is approximately a 1% chance of getting over 165 messages in any particular second we choose to examine.
Or, if the average TM's per second is only 119, then the probability of going over 165 in any single second is 1/37,000. Since there are 28,800 seconds in the 8-hour working day, if the average is below 119 (which it surely is), it would be unlikely even to see a single disruption in an 8-hour period.
I admit I was surprised the numbers came out this way, but they did. It ain't gonna happen.
"The cost to dell may indeed be 30 dollars, but as such the selling price should come down more than 30....What happens with most manufacturers is that they price for retail at a desireable price above Bill of Materials."
Who ever said this maximizes profits? In fact it usually doesn't.
Suppose Dell is selling only two products: Computer W and Computer N, which differ only in that W has Windows, and N has no OS. Presumably, Dell would set prices (pX, pN) on those two computers to maximize profits. Anyone who has taken a sufficiently advanced economics (or pricing) course would know that the profit maxmizing prices could differ by virtually any amount. That is, pX-pN need not be $30 (the cost of Windows). In fact the difference could even be less than $30, depending on consumer preference!
(Here's an extreme example. Suppose we finally reach Slashdot Utopia: everyone uses Linux. In such a world, Dell could not sell Computer W at any price higher than Computer N's price! pX-pN would actually be negative in this extreme example. Now this is not a realistic example, but it illustrates why "cost-plus pricing" may not be profit maximizing.)
So why is Dell dropping the price by exactly $30? *Shrug* Probably to keep goodwill with the masses who think cost-plus pricing is somehow "fair". No math can overcome that factor.
This actually has nothing to do with the CLT. The CLT is a statement about the distribution of a sample average in relation to the population average.
OTOH, TFA is about information markets, which concerns comparisions between the average of people's forecasts of events, and the "true" (or natural) frequency of events. We are not comparing a sample average to a population average. (In other words, to apply the CLT here, we'd make a statement comparing the average forecast of 100 people to the average forecast that would be made by all the people in the population. We're absolutely not doing that.)
Instead, the concept to discuss here is what is sometimes called calibration. This is a concept used to measure the accuracy of forecasters (such as meterologists). It works like this: say we look at all the days a weatherman said "chance of rain is 30%". Then, we'd hope to see rain on around 30% of those days in order for that kind of prediction to be accurate. Calibration (roughly) asks: How accurate were all such statements that were made by the weatherman? (I.e., check "chance of rain 0%", 10%, 20%, etc... looking at each subsample of his forecasts.)
Here (pdf) is a randomly chosen article on this kind of calibration.
For those that read TFA, did you see the graph? Calibration asks for those two lines to overlap. Whether they do or not has nothing to do with the Central Limit Theorem.
"If FedEx and UPS choose to give away free boxes for their own reasons, that is between them and the people that take advantage of it."
Fine.
"It doesn't effect other customers one bit."
Err, yes it does. At least theoretically. Now practically speaking, I doubt there are many people stea^H^H^H^H borrowing boxes like this. So, practically speaking, there's no effect to speak of.
But in general, variable costs affect prices. Eager for an econ homework exercise? I'm always happy to oblige. Read
a discussion
that finds the Cournot equilibrium in a 2-firm model. Then add this exercise: First, using the results on that page, find the equilibrium profit of each firm. Second, take its derivative with respect to the variable cost "c" (read: the cost of boxes). Verify that as "c" increases, price increases.
Don't like Cournot? Check out his counterpart Bertrand.
After reading that, verify that as marginal cost (MC) increases, so does price!
Don't like any French economists? Then look to the old general equilibrium models of economics (while pretending Debreu doesn't exist). The applicable ideas from that field are simple: if your costs go up, but you don't raise prices, you are no longer earning a decent rate (read: risk-adjusted market rate) of return on capital. You either need to charge more, or move your capital somewhere else. Conclusion? If costs go up, so do prices.
Is it unclear that the repetition was probably intentional? Judging by parent's post and its children, I am guessing so.
Generally "real" ROVs cost anywhere from $8,000 up to Millions of dollars. [On the other hand] We had two weeks and a $100 budget.
Informative....Usually they have high-resolution camera systems and high powered thrusters to maneuver. [On the other hand] We had two weeks and a $100 budget.
Someone who can stick a price on human life, or argument for improving the economy by killing people deserves no respect from me.
It must be awfully comfortable being in a position where one doesn't have to make decisions involving human life.
You think the speed limit should be 5 mph? After all, by raising it to, say, 55mph, we are trading the risk of human death for economic convenience. Well, someone has to make that decision.
Maybe all commercial flight must be stopped immediately. I know it's "economically convenient" for business people to be able to make trips, create jobs, etc. But one of these days, a plane will crash into someone's house and kill the occupants. No amount of economic gains is worth that "price", right? Of course not!
The list of examples is endless.
Did you ever do something in which you risked your life? (For example, drive a car!) Guess what. When you did that, you implicitly put an upper bound on the worth of your life; you said that it was worth taking the risk, in exchange for some convenience.
Blanket statements that "you can't put a price on life" sound nice, but are simply wrong. The proof is in everyday living.
Thus, out of a sample of 29 people and with p=0.21 the standard deviation is 2.2. Thus, 13 is 3.2 standard deviations away from 6. There is only a 0.07% chance that these are from the same distribution.
You are ignoring the variance of the "6". When you account for that variance, these outcomes are around 2 standard deviations away from each other.
"the people who were 'overcharged' could have chosen not to purchase the product, if they happened to think that it wasn't cost-effective. There are always alternatives.
"You may want to check with your state, but most states require everyone over the age of 18 to carry a state ID/DL/Passport/etc. If you do not and a cop stops you, he can cite you (possibly arrest you) for not carrying ID. Yes not many people know this,
I am one of the many people who don't know this. Can you cite just one state's law which backs up this claim?
I wonder what happens if someone is broke, and cannot afford to pay for the ID. These states would have an easy, legal way of tossing most homeless people in jail whenever they feel like it.
(In as troll-less a way as possible, I'm trying to say I don't believe the claim.)
Slashdot Mirror
"As far as I'm concerned, making it 'hard enough' to where it's more cost effective to resort to old fashioned brute force is just as good as not putting it in."
In principle, I agree. This probably isn't as big an issue as some think. But when you read some of the comments in the article, it is clear that the people making the decisions don't even understand most of the basics of what's going on. For instance, from TFA...
"Other comments suggested that the passport data should be
encrypted. The passport data on the chip does not require encryption in
order to be secure and protected. It is the same data that is visually
displayed on the passport data page."
Um, wasn't skimming the issue? Sigh.
"The same rules apply, say, for driving. If you want to drive other people around, commercially, you generally need some form of taxi driver's license."
Ironic you should use this example, as I was just thinking about taxi drivers.
Suppose I decide to sell an old painting at my local brick-and-mortar auction house. I grab my painting, and since I don't have a car, I get in a cab, and go to the auction house to drop it off for auctioning.
Question: Does my cab driver need an auction license?
What if I'm handicapped, and the cab driver has to carry the painting into the auction house for me?
What if I just hire him to take the painting to the auction house for me, so I can stay home?!
At which step here do we decide that this cab driver needs an auction license? Because that last analogy is almost exactly what the Ebay consignment guys are doing.
[.../ap_on_re_eu/bombed_smurfs;_...]
Surprisingly even just looking for a peak over 165 doesn't change the conclusions too much.
We can use the Poisson distribution to model random arrivals (like text message requests). For example, if the average number of messages sent per second is 138, but they arrive randomly over time, then there is approximately a 1% chance of getting over 165 messages in any particular second we choose to examine.
Or, if the average TM's per second is only 119, then the probability of going over 165 in any single second is 1/37,000. Since there are 28,800 seconds in the 8-hour working day, if the average is below 119 (which it surely is), it would be unlikely even to see a single disruption in an 8-hour period. I admit I was surprised the numbers came out this way, but they did. It ain't gonna happen.
Ah yes, the ubiquitous cost-plus pricing.
Who ever said this maximizes profits? In fact it usually doesn't.
Suppose Dell is selling only two products: Computer W and Computer N, which differ only in that W has Windows, and N has no OS. Presumably, Dell would set prices (pX, pN) on those two computers to maximize profits. Anyone who has taken a sufficiently advanced economics (or pricing) course would know that the profit maxmizing prices could differ by virtually any amount. That is, pX-pN need not be $30 (the cost of Windows). In fact the difference could even be less than $30, depending on consumer preference!
(Here's an extreme example. Suppose we finally reach Slashdot Utopia: everyone uses Linux. In such a world, Dell could not sell Computer W at any price higher than Computer N's price! pX-pN would actually be negative in this extreme example. Now this is not a realistic example, but it illustrates why "cost-plus pricing" may not be profit maximizing.)
So why is Dell dropping the price by exactly $30? *Shrug* Probably to keep goodwill with the masses who think cost-plus pricing is somehow "fair". No math can overcome that factor.
This actually has nothing to do with the CLT. The CLT is a statement about the distribution of a sample average in relation to the population average.
OTOH, TFA is about information markets, which concerns comparisions between the average of people's forecasts of events, and the "true" (or natural) frequency of events. We are not comparing a sample average to a population average. (In other words, to apply the CLT here, we'd make a statement comparing the average forecast of 100 people to the average forecast that would be made by all the people in the population. We're absolutely not doing that.)
Instead, the concept to discuss here is what is sometimes called calibration. This is a concept used to measure the accuracy of forecasters (such as meterologists). It works like this: say we look at all the days a weatherman said "chance of rain is 30%". Then, we'd hope to see rain on around 30% of those days in order for that kind of prediction to be accurate. Calibration (roughly) asks: How accurate were all such statements that were made by the weatherman? (I.e., check "chance of rain 0%", 10%, 20%, etc... looking at each subsample of his forecasts.)
Here (pdf) is a randomly chosen article on this kind of calibration.
For those that read TFA, did you see the graph? Calibration asks for those two lines to overlap. Whether they do or not has nothing to do with the Central Limit Theorem.
"Corporate America of course won't be happy until you pay a per-word usage fee for reading a library book."
As opposed to the good ole days, when books arrived in libraries for free?
Tax-dollar road map:
citizens' pockets --> gov't coffers --> local library budgets --> book publishers [aka "Corporate America"] --> authors
There's a bit of skimming at each step, of course, but please don't try to make it sound like we aren't already paying to read library books.
"If FedEx and UPS choose to give away free boxes for their own reasons, that is between them and the people that take advantage of it."
Fine.
"It doesn't effect other customers one bit."
Err, yes it does. At least theoretically. Now practically speaking, I doubt there are many people stea^H^H^H^H borrowing boxes like this. So, practically speaking, there's no effect to speak of.
But in general, variable costs affect prices. Eager for an econ homework exercise? I'm always happy to oblige. Read a discussion that finds the Cournot equilibrium in a 2-firm model. Then add this exercise: First, using the results on that page, find the equilibrium profit of each firm. Second, take its derivative with respect to the variable cost "c" (read: the cost of boxes). Verify that as "c" increases, price increases.
Don't like Cournot? Check out his counterpart Bertrand. After reading that, verify that as marginal cost (MC) increases, so does price!
Don't like any French economists? Then look to the old general equilibrium models of economics (while pretending Debreu doesn't exist). The applicable ideas from that field are simple: if your costs go up, but you don't raise prices, you are no longer earning a decent rate (read: risk-adjusted market rate) of return on capital. You either need to charge more, or move your capital somewhere else. Conclusion? If costs go up, so do prices.
And how are we supposed to ward off a cylon attack without space weapons?
I must be old. When I was a kid, I would have made this joke by referring to cylons.
Is it unclear that the repetition was probably intentional? Judging by parent's post and its children, I am guessing so.
...Usually they have high-resolution camera systems and high powered thrusters to maneuver. [On the other hand] We had two weeks and a $100 budget.
Generally "real" ROVs cost anywhere from $8,000 up to Millions of dollars. [On the other hand] We had two weeks and a $100 budget.
Informative.
Witty.
MG
Someone who can stick a price on human life, or argument for improving the economy by killing people deserves no respect from me.
It must be awfully comfortable being in a position where one doesn't have to make decisions involving human life.
You think the speed limit should be 5 mph? After all, by raising it to, say, 55mph, we are trading the risk of human death for economic convenience. Well, someone has to make that decision.
Maybe all commercial flight must be stopped immediately. I know it's "economically convenient" for business people to be able to make trips, create jobs, etc. But one of these days, a plane will crash into someone's house and kill the occupants. No amount of economic gains is worth that "price", right? Of course not!
The list of examples is endless.
Did you ever do something in which you risked your life? (For example, drive a car!) Guess what. When you did that, you implicitly put an upper bound on the worth of your life; you said that it was worth taking the risk, in exchange for some convenience.
Blanket statements that "you can't put a price on life" sound nice, but are simply wrong. The proof is in everyday living.
M.G.
Thus, out of a sample of 29 people and with p=0.21 the standard deviation is 2.2. Thus, 13 is 3.2 standard deviations away from 6. There is only a 0.07% chance that these are from the same distribution.
You are ignoring the variance of the "6". When you account for that variance, these outcomes are around 2 standard deviations away from each other.
When testing the difference between proportions you need to account for the variabilities of both populations.
"the people who were 'overcharged' could have chosen not to purchase the product, if they happened to think that it wasn't cost-effective. There are always alternatives.
Just a thought about free markets."
Not if there weren't enough alternatives.
Just a thought about monopolies.
> Ctrl+PgUp
> Ctrl+PgDwn
> rather, mor usefull
Ctrl+Tab
Ctrl+Shift+Tab
Just as useful.
"You may want to check with your state, but most states require everyone over the age of 18 to carry a state ID/DL/Passport/etc. If you do not and a cop stops you, he can cite you (possibly arrest you) for not carrying ID. Yes not many people know this,
I am one of the many people who don't know this. Can you cite just one state's law which backs up this claim?
I wonder what happens if someone is broke, and cannot afford to pay for the ID. These states would have an easy, legal way of tossing most homeless people in jail whenever they feel like it.
(In as troll-less a way as possible, I'm trying to say I don't believe the claim.)