So funny that they should call it Waterloo University in an article on grammar.
On an ironic note, when I was at the University of Waterloo, the it was usually the non-native english speakers who did the best on this because of their extensive training in taking the TOEFL. Most of these guys couldn't talk their way out of a wet paper bag.
I agree, although I would also suggest using real numbers. In data-mining its a relatively annoying but straightforward process of eliminating outliers. The would be your 555 numbers the 123 Yosemitie Sam Lanes and SSNs like 123456789. If you really want to be a pain in the ass you have to make sure that all the information corroborates plausibly...this makes you a data confounder and not just an outlier.
>>But the "unreasonable effectiveness" of mathematics in explaining the world, as the physicist Eugene Wigner once put it, is a minor motivation at best for those immersed in the field. Most mathematicians say they are in it for the math itself, for the delirious quest for patterns, the thrill of the detective chase and the lure of beautiful answers.
>I sure hope this isn't really true. If mathematicans aren't really interested in helping understand the world, why should society fund them? I certainly know that a major motivation for my career in science is that understanding the world through science will help people, cure diseases, etc
Both wrong. As a mathematician I can assure you that we're in it for the hot chicks, cha-ching and bling bling.
The TIA is scary conceptually, but when coupled with human incompetence it becomes absolutely terrifying. The problem with TIA is not that they collect the data, but that they must at some point make some interpretation and inference on the data.
Those familiar with data mining know that there are very serious problems in turning data into information. This means that some human must make some pronouncement that some types of data mean certain things. If they collect all your emails, what exactly identifies you as a terroist? If they collect your credit card transactions, what do you have to buy to be a terrorist? Etc.
Aside from the technical problems associated with coding and interpreting the data there are other problems associated with detection. Basically, in data mining there are two paths that TIA can take. Supervised algorithms (here's what a terrorist looks like go find others) and unsupervised algorithms (let's clump all people who buy lots of fertilizer together). The problem with supervised algorithms, is that you need examples of what a terrorist looks like. Even, then you need a whole bunch of examples to teach algorithms to identify them. The problem with unsupervised techniques is that once you have grouped everyone together, do you have the right number of groups, and are the groups any good at identifying terroists?
I'm just grazing the technical problems...and there are many many more, but in hte end, the most problematic is that even if it is possible to create some algorithm, no method is 100% accurate. This means that (many) innocent people will get targeted and victimized by the TIA just because they happen to "fit the profile".
The best way I see to beat this is 1) create multiple personae 2) look like everyone else.
My 2 cents.
Re:Too bad Coke pricing isn't weather sensitive
on
Which Price is Right?
·
· Score: 1
Wrong. The reason you pay the outrageous prices is not because of limited supply, but something called a captive market. They are the only vendor/vendors and so have exclusive access to your $$$. This means that unless you are bringing your own goodies, which many establishments frown upon, you are going to pay however much the vendor wants. i.e. THey jack up prices because they can.
At first I read this as a troll. But I think the issue here is that the poster does not recognize that risk costs money.
Financial services to the poor have, all else equal, much higher default risk. And default costs swamp everything else. Consider that the margin over cost of funds for most consumer credit is 2-3%. A default rate of 1% destroys the profitability.
This is an interesting if not a naive point of view. There are certainly differences in the cost of lending due to default risk, but believe me, that the margins that you quote are extremely low. Margin that low would only be for mortgages and auto, but these tend to be securitied and thus have other risk dynamics, but for true consumer finance (i.e. sales finance, credit cards, personal loans, overdraft), the margins you are looking at are 15% - 25%. Sometimes more if you consider fees.
Credit card companies typically have a NIM (net interest margain) of about 15% on average. They expect and reserve for approximately 5% of that as losses. This means that you have a 10% margin to cover the costs and make money. This net interest margain has increased steadily over the last 5 years. Ever see your interest rate go down?
The funny thing is, here the pricing IS proportionate to the risk. Generally your interest rate increases with missing payments. THe funny part is, these people are the most profitable...meaning if you are just on the margin, you end up being the best customer. Rich people tend to pay off their balances in full so the companies make very little money off these people. So in the end it is the poor people getting fleeced.
Also a side point. As a general rule in credit: If more you need credit, the harder it will be to get it.
Yes, quite. Let me point out that like it or not, stereotypes persist because they are usually quite good generalizations.
Jocks as a group tend to be stupid (your words not mine)
Popular kids as a group tend to be jerks (to geeks)
Many people are just plain dicks.
Also recognize that many kids do not fit into a clear cut category, people fit into stereotypes because they have many traits of that stereotype, but sterotypes are not exhaustive.
Sterotypes....big deal.
Slashdot Mirror
So funny that they should call it Waterloo University in an article on grammar. On an ironic note, when I was at the University of Waterloo, the it was usually the non-native english speakers who did the best on this because of their extensive training in taking the TOEFL. Most of these guys couldn't talk their way out of a wet paper bag.
All of a sudden I see chainmail coming back into fashion!
I agree, although I would also suggest using real numbers. In data-mining its a relatively annoying but straightforward process of eliminating outliers. The would be your 555 numbers the 123 Yosemitie Sam Lanes and SSNs like 123456789. If you really want to be a pain in the ass you have to make sure that all the information corroborates plausibly...this makes you a data confounder and not just an outlier.
NYT doesn't spam. And the percentage of net.morons who register using cartoon names is remarkably low.
>I sure hope this isn't really true. If mathematicans aren't really interested in helping understand the world, why should society fund them? I certainly know that a major motivation for my career in science is that understanding the world through science will help people, cure diseases, etc
Both wrong. As a mathematician I can assure you that we're in it for the hot chicks, cha-ching and bling bling.
The TIA is scary conceptually, but when coupled with human incompetence it becomes absolutely terrifying. The problem with TIA is not that they collect the data, but that they must at some point make some interpretation and inference on the data.
Those familiar with data mining know that there are very serious problems in turning data into information. This means that some human must make some pronouncement that some types of data mean certain things. If they collect all your emails, what exactly identifies you as a terroist? If they collect your credit card transactions, what do you have to buy to be a terrorist? Etc.
Aside from the technical problems associated with coding and interpreting the data there are other problems associated with detection. Basically, in data mining there are two paths that TIA can take. Supervised algorithms (here's what a terrorist looks like go find others) and unsupervised algorithms (let's clump all people who buy lots of fertilizer together). The problem with supervised algorithms, is that you need examples of what a terrorist looks like. Even, then you need a whole bunch of examples to teach algorithms to identify them. The problem with unsupervised techniques is that once you have grouped everyone together, do you have the right number of groups, and are the groups any good at identifying terroists?
I'm just grazing the technical problems...and there are many many more, but in hte end, the most problematic is that even if it is possible to create some algorithm, no method is 100% accurate. This means that (many) innocent people will get targeted and victimized by the TIA just because they happen to "fit the profile".
The best way I see to beat this is 1) create multiple personae 2) look like everyone else.
My 2 cents.
Wrong. The reason you pay the outrageous prices is not because of limited supply, but something called a captive market. They are the only vendor/vendors and so have exclusive access to your $$$. This means that unless you are bringing your own goodies, which many establishments frown upon, you are going to pay however much the vendor wants. i.e. THey jack up prices because they can.
At first I read this as a troll. But I think the issue here is that the poster does not recognize that risk costs money. Financial services to the poor have, all else equal, much higher default risk. And default costs swamp everything else. Consider that the margin over cost of funds for most consumer credit is 2-3%. A default rate of 1% destroys the profitability. This is an interesting if not a naive point of view. There are certainly differences in the cost of lending due to default risk, but believe me, that the margins that you quote are extremely low. Margin that low would only be for mortgages and auto, but these tend to be securitied and thus have other risk dynamics, but for true consumer finance (i.e. sales finance, credit cards, personal loans, overdraft), the margins you are looking at are 15% - 25%. Sometimes more if you consider fees. Credit card companies typically have a NIM (net interest margain) of about 15% on average. They expect and reserve for approximately 5% of that as losses. This means that you have a 10% margin to cover the costs and make money. This net interest margain has increased steadily over the last 5 years. Ever see your interest rate go down? The funny thing is, here the pricing IS proportionate to the risk. Generally your interest rate increases with missing payments. THe funny part is, these people are the most profitable...meaning if you are just on the margin, you end up being the best customer. Rich people tend to pay off their balances in full so the companies make very little money off these people. So in the end it is the poor people getting fleeced. Also a side point. As a general rule in credit: If more you need credit, the harder it will be to get it.
Yes, quite. Let me point out that like it or not, stereotypes persist because they are usually quite good generalizations. Jocks as a group tend to be stupid (your words not mine) Popular kids as a group tend to be jerks (to geeks) Many people are just plain dicks. Also recognize that many kids do not fit into a clear cut category, people fit into stereotypes because they have many traits of that stereotype, but sterotypes are not exhaustive. Sterotypes....big deal.
If animal lovers are more likely to vegetarians are plant lovers less likely to be vegetarians?