That or Micro four thirds/ Micro four thirds.
Two things to consider when trying to learn photography are 1) Interchangeable lenses and 2) Getting the largest possible image sensor you can get (Noise decreases with image sensor size, not with megapixels).
Micro four thirds (or similar formats from Sony/Samsung) have a larger sensor than a typical point-and-shoot. So they work better in low lights and generally have a higher Signal-to-Noise ratio.
Olympus EPL line is a pretty good and cheap micro 4/3 camera. Sony makes their NEX series which are the full blown APS-C (DSLR) sensors squeezed into a small camera. Olympus and Panasonic both make micro-4/3 cameras, so the lenses are easier to come by. Also they can use adapters for various other kind of lenses. On the flip side, the micro 4/3 sensor is only 60% or so in size compared to a DSLR sensor.
All the above also have pre-set modes to increase/decrease brightness, contrast etc. -- they are useful to start. Once there you can set the camera to full manual and learn the physics part of photography. I've been told that Samsung makes the easiest to use interface , but I have never used it.
(Microsoft OneNote running on a Samsung Series 7 with Windows 7) then they certainly won't exceed a regular laptop with a keyboard.
Not true in my experience.I used OneNote on a Lenovo X60 for about a year and somemore to take notes. I write both in cursive and regular block letters and OneNote had no problems recognizing the text. My cursive can be very ornate (and is not standard American) but still OneNote had no issues. I dual booted and tried using Linux for a while, but quickly went back to OneNote for the following reasons.
The advantages I found are :
1) Only one laptop to carry around instead of multiple books.
2) Searchable text - One Note lets you Ctrl-F through handwritten notes, which is very useful when looking for a specific class note. You just open up OneNote and search through all notebooks - haven't had a problem with "lack of integration".
3) Images once drawn can be moved around on the page - so you take notes, draw diagrams etc and then you can move them around to make logical sense.
4) Silent - no "clicky-clicky" sounds when you are taking notes- very useful when in a humanities class. Also the tablet lies flat on the desk instead of upright, so the professor/teacher can see your face, and the guys behind you can't see your screen if you are browsing instead of taking notes.
after all, if it were trustworthy, why would it need backing
Banks get deposits which can be withdrawn at any point, while their lending (loans) are long term. If customers demand money, the bank cannot then turn around and sell another customers house. Banks never have enough money to meet all their clients at once. So even prudent,trustworthy banks need government backing to ensure that they are protected against a run on the bank
If you take away govt. backing to banks, arguably, you would not have banks at all.You would have organizations which lend money (like pawnbrokers etc.) and other organizations that hold/transfer money for you (like paypal). It is not necessarily a bad thing, but it would not be a bank as we know it.
True, but I think Taleb's point was to cut both Bonuses and Salaries to make Bankers "Lifers". He explicitly points out that a lot of banking employees would leave, and hedge funds (which need not be bailed out) will pick up the risk that banks lay off.
I believe this is how it would work out according to Taleb.
1) Banks cut salaries/bonuses
2) Lot of workers leave
3) The banks can't do much else other than old fashioned banking
4) Hedge funds partnership based pick up all the "cool stuff"
5) Since partnerships and hedge funds do not have "Limited Liability" , they are much more careful
The Atlantic article is incorrect in using the Lehman example. Taleb's point is that a partnership with unlimited liabilities would never have gotten into the situation Lehman got into. He does not say that Lehmans corporate structure mattered in any way after Lehman got into the situation it did.
Unlimited liability + Bonus is a very different incentive compared to Limited liability + Salary which are both better than Limited liability + Bonus.
If you were a trader who could do a billion dollar trade that had 50% chance of winning that would fetch you a million dollar bonus and 0 liability (except loss of job which pays you 100K a year), you would likely take the trade if you thought of sticking to the job only for 5 years. But you would be less likely to take it if you had unlimited liability.
There is a bigger problem in the story
It says that Androids have a higher percentage of calls related to Hardware compared to Apple. They do not compare (hardware call numbers)/(Phones in existence).
So it is equally possible that Android software has a higher quality compared to Apple. for eg: Android gets 100 calls,14 for h/w, 86 for software. Apple gets 50 calls, 3 of which are hardware related. If there were three times as many Androids in existence compared to Apple, Android is just better than Apple. If Androids were twice as popular as Apple, then Android has better software and Apple has better hardware. If Apple has equal market share to android, then Apple is the better phone
There are two unknown variables - Software/Hardware quality ratio and quality/phones sold ratio. The article picks just one of them and tries to interpret it.
I believe what they are proposing is to make up a currency backed by physical assets. Not sure how this would help considering that our problem was a very similar one to the real estate crisis
In the real estate crisis, people started using their houses as a sort of currency. No money? No problem, tap the equity on the house and buy a car. And once the house prices started to fall, we realized that we had too many houses for the people we have in the country, most of the extra being built as "investment properties". And all through the boom, at least from what I saw in Spain, the people who need to buy them could not afford them. After the crash, no one has a job to buy them.
I am afraid we will end up with something similar might happen with Terra. The currency involves stockpiling commodities at high (3%-4%) costs. All that will do is to cause people to pile up commodities, pay for losses/theft etc. and then tap it out as need be. I am not sure how this would be any different from using CDOs on houses as currency, which the banks were doing during the boom times.
I agree somewhat with what you say but disagree with the idea that sitting around in a park won't change anything. All you have to do is to look up recent south African or slightly older Indian history. Heck, even in America a woman sitting in a bus changed a lot of things.
The idea that you can only change things by voting or by violent overthrow is simplistic. It is also possible to influence how people vote. For example: Nelson Mandela started out with the idea of violent overthrow of govt. (since voting was not a real option), then he moved on to sitting in his jail cell and protesting until the ruling class gave up. And that caused real change -- S. Africa has not had a serious race riot or violence because their leader renounced it. Or in the case of Gandhi, he pretty much convinced his nation (which early on did not care) that the British were a bad idea by repeatedly getting beaten up or thrown in jail for non-violent protests.
This is the age old idea of marketing -- you have to convince the "user" that he needs your product even if the user would not have picked it out from a self-made list of "needs". I believe sitting around in a park may convince some of the electorate that some sort of change is required.
The stories at the time of arrest indicate that it was Equity Index linked securities that the trader was gambling on, not Swiss Franc like it was widely assumed.
That was also the time when European indices, emerging market stocks and to a lesser extent US stocks crashed. But otherwise you are right - apparently Adoboli had done hidden trades starting as far back as 2008 and they were generally profitable. http://www.guardian.co.uk/business/2011/sep/17/kweku-adoboli-ubs-fraud-charges
Nick Leeson did not work in IT according to his biography or according to Wiki.
He used an error account, which he realized was unaudited, but that is something you pick up from being a trader or an auditor- not necessarily IT. These things are common in investment banks/brokerages which have a lot of accounts and client trades and errors need to be isolated in an account that does not belong to a client. ie. if a client asked to buy 100 pork belly contracts and you bought him lean hogs instead, you need a place to dump the pork bellies you bought. It does not mean a "test account" in the IT sense.
Economic models show that in a competitive market, prices fall until the marginal producer is making zero profits. Cable TV is not very competitive (barriers to entry in the form of laying cables etc) but still it is not a pure monopoly either. Even in such a market, prices can be raised by bundling. Breaking the bundles are more likely than not to decrease prices for individual customers.
The paper specifically uses Cable TV as an information good and shows how bundling works http://ebusiness.mit.edu/erik/Bundling%20Competition685305.pdf
If that was all - here is their filing below. RESTATEMENT
The Company has restated its previously issued Consolidated Statements of Operations for the years ended December 31, 2008, 2009 and 2010 to correct for an error in its presentation of revenue.
Most significantly, the Company restated its reporting of revenues from Groupons to be net of the amounts related to merchant fees. Historically, the Company has reported the gross amounts billed to its subscribers as revenue. All prior periods have been restated to show the net amount the Company retains after paying the merchant fees. The effect of the correction resulted in a reduction of previously reported revenues and corresponding reductions in cost of revenue in those periods. The change in presentation had no effect on pre-tax loss, net loss or any per share amounts for any period presented.
The Company has also changed the presentation of certain other income statement expenses to be consistent with reporting revenue on a net basis. These changes include presenting loyalty programs as a component of marketing rather than as an offset to revenue. The Company believes that this classification is most appropriate as it is acting as an agent on behalf of the merchant in driving traffic to generate revenue. In addition, refunds made to subscribers under the Groupon Promise are presented as a component of cost of revenue, rather than as an offset to revenue, as these amounts are not paid directly to the merchants. Credit card and other processing expenses have been reclassified to cost of revenue from selling, general and administrative for all periods presented. The Company has concluded the amounts could alternatively be viewed as a cost of the service the Company is providing.
So not only did they change their revenue recognition policies, they also had an "error" which is separate. And then they had to restate revenues from their loyalty programs and credit card payments. What were they making up their accounts for two years? Crayons and Etch-a-Sketch?
Disclaimer: I don't work for UBS or a bank.
I don't think there is a need to read too much into the word "hedged"
The guy seems to have worked on ETF/delta one desk. The desks job is to sell clients securities and replicate the returns using equivalent assets . For eg: if an insurance company wanted a product that tracks the inflation of Eurozone, UBS would (presumably) sell them a swap or a debt instrument which would pay the weighted average of inflation across Eurozone. They would then "hedge" this risk by buying European inflation protected bonds. As long as they keep the $$ amounts balanced and adjust for currency movements , they are hedged. If German inflation increases, UBS has to pay the insurance company a higher amount for the month, but the higher payouts on the German inflation protected bonds balance this out. UBS's profit would be the fees they charge the insurance company for providing this service. UBS can also make additional profit by buying Italian bonds (which pay higher interest rates) instead of German bonds -- but this involves "risk". The risk here is that Italy goes bankrupt while Germany does not.
It is the job of the trader to ensure that (s)he stays within the risk limits prescribed and the risk department should monitor any excessive risk taking . ie. if the trader sold someone a bond pegged to South African risk, risk department should ensure that the trader did not hedge it with bonds involving Somalia.
The OP was surprised that the speculation seems to suggest that it was not a case of bad hedging, but the fact that a bunch of hedges were not in place at all!! It should be a case of simple accounting to see that there is an instrument on UBS's books which has a notional value of a few billion dollars which has no equivalent hedge.
In the case of Nick Leeson (Barings Bank), he used an "error account" - a pseudo account in the bank to hide his losses. In case of Societe Generale (Kerviel) he hacked into the system to hide his trades. It is yet unclear how this guy hid his trades.
The article does not specify either (in one place it HTC obtained and in another place, "bought").
It could be that HTC got an exclusive license on that patent from Google or the right to collect royalty on those patents while the patents are still held by Google. It is not uncommon,for ex:in case of Novell/SCO, Novell held the copyrights but SCO had rights to collect royalty (or so it was alleged).
It took them a bit to take action on my report written by a street in the first person begging to be found, but they did wnd my street now appears in google maps
Your street found a beggar on it and wrote a report to Google? Smart street!
That is not necessarily a bad thing or actually unheard of.
Try calling up Matlab helpdesk and you'll get someone with a masters degree in Boston or someplace. They don't often ask you to reboot the machine to see if the problem went away.
Coherent, No. Response, apparently.
From their post For every defendant in the anonymous "conspiracy" we are attacking two top Texas police chiefs, leaking 3GB of their private emails and attachments.
The full pdf paper is slightly more informative. While the summary would seem to indicate that chances of mortality decreases linearly with increasing exercise, from Figure 2 of the chart it seems like after 100 minutes a day the benefits taper off. So it seems like 15 minutes a day is good,150 is overdoing it.
http://www.natap.org/2011/HIV/PIIS0140673611607496.pdf
They also note that ex-smokers exercised more than the norm, so that might be contributing to the decrease in cancer rates (correlation, not causation). Probably very imp. considering the study was done in Taiwan.
Felt it in Connecticut too. 3 out of 6 people in the room felt it, other did not.
Twitter already had news of this (reporting epicenter as Virginia) within the next 3 minutes, so we all went back to work. Mostly Manhattanites in skyscrapers tweeting. Bloomberg, WSJ (blog) and Washington Post had news updated in the next few minutes in that order too.
Maybe true, but that is only exactly true if there is no vertical load applied on that bar tied to its "waist".
Also that thing does not have any hands to provide the balancing force for the fact that one of its legs is ahead of the other, instead it uses a strange load on top of its "neck" to balance itself. That basically means that the "hips" of this robot does not twist like a human being walking. So the adjustment for forward balance actually prevents it from being laterally balanced.
(Human walking being right leg forward, left hand forward and the hip rotated slightly to the left providing the lateral stability).
Actually it is even worse than that.
In the absence of HFT's the alternative is to have specialist market-makers at exchanges. That would mean an oligopoly of people who can buy a seat of exchanges making money on bid ask spreads. They were kicked out of exchanges after a set of large investors sued them for front-running their trades http://www.businessweek.com/bwdaily/dnflash/dec2003/nf20031217_6559_db016.htm.
If you had neither specialists nor HFTs, a trade could only be done in case there was a buyer and a seller. It would be great in case of normal markets, but in case of market panics there would be far more sellers and buyers causing price swings to worsen. It is believed that this would be bad for liquidity and markets in general.
There are some other ways proposed to solve this problem - some have suggested an auction every second, with no trades done in between those seconds. IE there would be only 28,800 "trades" in an 8 hour day. Every order that comes in between is consolidated. It is not clear if that would still help because presumably some computer would figure out an algo. that could beat the auction.
Thanks for the link. It is pretty spot on.
At the risk of being off topic.
But that said the latest crisis was not one of unexpectedly large standard deviations alone, but one of badly modeled correlations. The article is correct that the variance of a portfolio is larger than it looks, but many in the equity markets had figured that out. What happened in this recent crash was that they used scant data from mortgage markets to try and estimate correlations. For eg: Biggest possible loss for Californian mortgages was say 10%. Adjusting for a large variance, you could say that the chance of loss is say 50%. You could account for that in some way, say.
But what if someone bought a bond that would only default (not pay) if 40% of bonds in 20 states defaulted? That's where the correlation came in. They assumed in a rather complicated way that the chances of bonds defaulting are related to the correlations of bonds in different states. After all, a default is just a very large negative move.
Correlations between bonds in different states were pretty low because each state had varying economic conditions etc. So plug that into the model and you would find that the newfangled bond that would only default if 40% of bonds in all 20 states defaulted was very safe. So everyone called it a AAA.
Except it turned out that all the states were on a property bubble fueled by low interest rates and greed and had a correlation of 1 in bad times. Hence the mess we are in.
Anyway, what I am trying to say is that Fractal model might be very good at estimating variances , but still would not solve the problem of estimating covariances.
The story pretty much says the same thing.
AT&T has said that it will cost them $39 Billion to buy TMobile. It has also claimed to investors that it will save them $10Billion.
The justification is that it will enlarge Wireless coverage in USA.
The leak now claims that the expanded network coverage will cost only $3.8 Billion http://www.wirelessweek.com/news/2011/08/unredacted-ATT-filing-shows-high-price-tag/
So AT&T Pays $39 Billion, saves $3.8 Bilion in network costs and $6.2 Billion in non-network costs (say closing and selling stores/laying off duplicated jobs). Assume that Sprint is worth the same as TMobile intrinscially - which is approx. $10B. The remaining $19 Billion premium paid must then be the cost of eliminating competition or the cost of keeping sprint from expanding.
Google seems to have fixed the error message on their code search
Also +"<FILEHARD>" seems to do better than simple quotes - The "+" forces google to look for single word Filehard.
Slashdot Mirror
That or Micro four thirds/ Micro four thirds.
Two things to consider when trying to learn photography are 1) Interchangeable lenses and 2) Getting the largest possible image sensor you can get (Noise decreases with image sensor size, not with megapixels).
Micro four thirds (or similar formats from Sony/Samsung) have a larger sensor than a typical point-and-shoot. So they work better in low lights and generally have a higher Signal-to-Noise ratio.
Olympus EPL line is a pretty good and cheap micro 4/3 camera. Sony makes their NEX series which are the full blown APS-C (DSLR) sensors squeezed into a small camera. Olympus and Panasonic both make micro-4/3 cameras, so the lenses are easier to come by. Also they can use adapters for various other kind of lenses. On the flip side, the micro 4/3 sensor is only 60% or so in size compared to a DSLR sensor.
All the above also have pre-set modes to increase/decrease brightness, contrast etc. -- they are useful to start. Once there you can set the camera to full manual and learn the physics part of photography. I've been told that Samsung makes the easiest to use interface , but I have never used it.
(Microsoft OneNote running on a Samsung Series 7 with Windows 7) then they certainly won't exceed a regular laptop with a keyboard.
Not true in my experience.I used OneNote on a Lenovo X60 for about a year and somemore to take notes. I write both in cursive and regular block letters and OneNote had no problems recognizing the text. My cursive can be very ornate (and is not standard American) but still OneNote had no issues. I dual booted and tried using Linux for a while, but quickly went back to OneNote for the following reasons.
The advantages I found are :
1) Only one laptop to carry around instead of multiple books.
2) Searchable text - One Note lets you Ctrl-F through handwritten notes, which is very useful when looking for a specific class note. You just open up OneNote and search through all notebooks - haven't had a problem with "lack of integration".
3) Images once drawn can be moved around on the page - so you take notes, draw diagrams etc and then you can move them around to make logical sense.
4) Silent - no "clicky-clicky" sounds when you are taking notes- very useful when in a humanities class. Also the tablet lies flat on the desk instead of upright, so the professor/teacher can see your face, and the guys behind you can't see your screen if you are browsing instead of taking notes.
after all, if it were trustworthy, why would it need backing
Banks get deposits which can be withdrawn at any point, while their lending (loans) are long term. If customers demand money, the bank cannot then turn around and sell another customers house. Banks never have enough money to meet all their clients at once. So even prudent,trustworthy banks need government backing to ensure that they are protected against a run on the bank
If you take away govt. backing to banks, arguably, you would not have banks at all.You would have organizations which lend money (like pawnbrokers etc.) and other organizations that hold/transfer money for you (like paypal). It is not necessarily a bad thing, but it would not be a bank as we know it.
True, but I think Taleb's point was to cut both Bonuses and Salaries to make Bankers "Lifers". He explicitly points out that a lot of banking employees would leave, and hedge funds (which need not be bailed out) will pick up the risk that banks lay off.
I believe this is how it would work out according to Taleb.
1) Banks cut salaries/bonuses
2) Lot of workers leave
3) The banks can't do much else other than old fashioned banking
4) Hedge funds partnership based pick up all the "cool stuff"
5) Since partnerships and hedge funds do not have "Limited Liability" , they are much more careful
The Atlantic article is incorrect in using the Lehman example. Taleb's point is that a partnership with unlimited liabilities would never have gotten into the situation Lehman got into. He does not say that Lehmans corporate structure mattered in any way after Lehman got into the situation it did.
Unlimited liability + Bonus is a very different incentive compared to Limited liability + Salary which are both better than Limited liability + Bonus.
If you were a trader who could do a billion dollar trade that had 50% chance of winning that would fetch you a million dollar bonus and 0 liability (except loss of job which pays you 100K a year), you would likely take the trade if you thought of sticking to the job only for 5 years. But you would be less likely to take it if you had unlimited liability.
There is a bigger problem in the story /Hardware quality ratio and quality/phones sold ratio. The article picks just one of them and tries to interpret it.
It says that Androids have a higher percentage of calls related to Hardware compared to Apple. They do not compare (hardware call numbers)/(Phones in existence).
So it is equally possible that Android software has a higher quality compared to Apple. for eg: Android gets 100 calls,14 for h/w, 86 for software. Apple gets 50 calls, 3 of which are hardware related. If there were three times as many Androids in existence compared to Apple, Android is just better than Apple. If Androids were twice as popular as Apple, then Android has better software and Apple has better hardware. If Apple has equal market share to android, then Apple is the better phone
There are two unknown variables - Software
I believe what they are proposing is to make up a currency backed by physical assets. Not sure how this would help considering that our problem was a very similar one to the real estate crisis
In the real estate crisis, people started using their houses as a sort of currency. No money? No problem, tap the equity on the house and buy a car. And once the house prices started to fall, we realized that we had too many houses for the people we have in the country, most of the extra being built as "investment properties". And all through the boom, at least from what I saw in Spain, the people who need to buy them could not afford them. After the crash, no one has a job to buy them.
I am afraid we will end up with something similar might happen with Terra. The currency involves stockpiling commodities at high (3%-4%) costs. All that will do is to cause people to pile up commodities, pay for losses/theft etc. and then tap it out as need be. I am not sure how this would be any different from using CDOs on houses as currency, which the banks were doing during the boom times.
I agree somewhat with what you say but disagree with the idea that sitting around in a park won't change anything. All you have to do is to look up recent south African or slightly older Indian history. Heck, even in America a woman sitting in a bus changed a lot of things.
The idea that you can only change things by voting or by violent overthrow is simplistic. It is also possible to influence how people vote. For example: Nelson Mandela started out with the idea of violent overthrow of govt. (since voting was not a real option), then he moved on to sitting in his jail cell and protesting until the ruling class gave up. And that caused real change -- S. Africa has not had a serious race riot or violence because their leader renounced it. Or in the case of Gandhi, he pretty much convinced his nation (which early on did not care) that the British were a bad idea by repeatedly getting beaten up or thrown in jail for non-violent protests.
This is the age old idea of marketing -- you have to convince the "user" that he needs your product even if the user would not have picked it out from a self-made list of "needs". I believe sitting around in a park may convince some of the electorate that some sort of change is required.
The stories at the time of arrest indicate that it was Equity Index linked securities that the trader was gambling on, not Swiss Franc like it was widely assumed.
That was also the time when European indices, emerging market stocks and to a lesser extent US stocks crashed. But otherwise you are right - apparently Adoboli had done hidden trades starting as far back as 2008 and they were generally profitable. http://www.guardian.co.uk/business/2011/sep/17/kweku-adoboli-ubs-fraud-charges
Nick Leeson did not work in IT according to his biography or according to Wiki.
He used an error account, which he realized was unaudited, but that is something you pick up from being a trader or an auditor- not necessarily IT. These things are common in investment banks/brokerages which have a lot of accounts and client trades and errors need to be isolated in an account that does not belong to a client. ie. if a client asked to buy 100 pork belly contracts and you bought him lean hogs instead, you need a place to dump the pork bellies you bought. It does not mean a "test account" in the IT sense.
Economic models show that in a competitive market, prices fall until the marginal producer is making zero profits. Cable TV is not very competitive (barriers to entry in the form of laying cables etc) but still it is not a pure monopoly either. Even in such a market, prices can be raised by bundling. Breaking the bundles are more likely than not to decrease prices for individual customers.
The paper specifically uses Cable TV as an information good and shows how bundling works http://ebusiness.mit.edu/erik/Bundling%20Competition685305.pdf
Intel used to ship Itanium and other servers with UEFI.
And I know I have installed RedHat enterprise versions on it with no problems.
If that was all - here is their filing below.
RESTATEMENT The Company has restated its previously issued Consolidated Statements of Operations for the years ended December 31, 2008, 2009 and 2010 to correct for an error in its presentation of revenue.
Most significantly, the Company restated its reporting of revenues from Groupons to be net of the amounts related to merchant fees. Historically, the Company has reported the gross amounts billed to its subscribers as revenue. All prior periods have been restated to show the net amount the Company retains after paying the merchant fees. The effect of the correction resulted in a reduction of previously reported revenues and corresponding reductions in cost of revenue in those periods. The change in presentation had no effect on pre-tax loss, net loss or any per share amounts for any period presented.
The Company has also changed the presentation of certain other income statement expenses to be consistent with reporting revenue on a net basis. These changes include presenting loyalty programs as a component of marketing rather than as an offset to revenue. The Company believes that this classification is most appropriate as it is acting as an agent on behalf of the merchant in driving traffic to generate revenue. In addition, refunds made to subscribers under the Groupon Promise are presented as a component of cost of revenue, rather than as an offset to revenue, as these amounts are not paid directly to the merchants.
Credit card and other processing expenses have been reclassified to cost of revenue from selling, general and administrative for all periods presented. The Company has concluded the amounts could alternatively be viewed as a cost of the service the Company is providing.
So not only did they change their revenue recognition policies, they also had an "error" which is separate. And then they had to restate revenues from their loyalty programs and credit card payments. What were they making up their accounts for two years? Crayons and Etch-a-Sketch?
Disclaimer: I don't work for UBS or a bank.
I don't think there is a need to read too much into the word "hedged"
The guy seems to have worked on ETF/delta one desk. The desks job is to sell clients securities and replicate the returns using equivalent assets . For eg: if an insurance company wanted a product that tracks the inflation of Eurozone, UBS would (presumably) sell them a swap or a debt instrument which would pay the weighted average of inflation across Eurozone. They would then "hedge" this risk by buying European inflation protected bonds. As long as they keep the $$ amounts balanced and adjust for currency movements , they are hedged. If German inflation increases, UBS has to pay the insurance company a higher amount for the month, but the higher payouts on the German inflation protected bonds balance this out. UBS's profit would be the fees they charge the insurance company for providing this service. UBS can also make additional profit by buying Italian bonds (which pay higher interest rates) instead of German bonds -- but this involves "risk". The risk here is that Italy goes bankrupt while Germany does not.
It is the job of the trader to ensure that (s)he stays within the risk limits prescribed and the risk department should monitor any excessive risk taking . ie. if the trader sold someone a bond pegged to South African risk, risk department should ensure that the trader did not hedge it with bonds involving Somalia.
The OP was surprised that the speculation seems to suggest that it was not a case of bad hedging, but the fact that a bunch of hedges were not in place at all!! It should be a case of simple accounting to see that there is an instrument on UBS's books which has a notional value of a few billion dollars which has no equivalent hedge.
In the case of Nick Leeson (Barings Bank), he used an "error account" - a pseudo account in the bank to hide his losses. In case of Societe Generale (Kerviel) he hacked into the system to hide his trades. It is yet unclear how this guy hid his trades.
The article does not specify either (in one place it HTC obtained and in another place, "bought").
It could be that HTC got an exclusive license on that patent from Google or the right to collect royalty on those patents while the patents are still held by Google. It is not uncommon,for ex:in case of Novell/SCO, Novell held the copyrights but SCO had rights to collect royalty (or so it was alleged).
It took them a bit to take action on my report written by a street in the first person begging to be found, but they did wnd my street now appears in google maps
Your street found a beggar on it and wrote a report to Google? Smart street!
That is not necessarily a bad thing or actually unheard of.
Try calling up Matlab helpdesk and you'll get someone with a masters degree in Boston or someplace. They don't often ask you to reboot the machine to see if the problem went away.
Coherent, No. Response, apparently.
From their post For every defendant in the anonymous "conspiracy" we are attacking two top Texas police chiefs, leaking 3GB of their private emails and attachments.
The full pdf paper is slightly more informative. While the summary would seem to indicate that chances of mortality decreases linearly with increasing exercise, from Figure 2 of the chart it seems like after 100 minutes a day the benefits taper off. So it seems like 15 minutes a day is good,150 is overdoing it. http://www.natap.org/2011/HIV/PIIS0140673611607496.pdf
They also note that ex-smokers exercised more than the norm, so that might be contributing to the decrease in cancer rates (correlation, not causation). Probably very imp. considering the study was done in Taiwan.
Felt it in Connecticut too. 3 out of 6 people in the room felt it, other did not.
Twitter already had news of this (reporting epicenter as Virginia) within the next 3 minutes, so we all went back to work. Mostly Manhattanites in skyscrapers tweeting.
Bloomberg, WSJ (blog) and Washington Post had news updated in the next few minutes in that order too.
No it worked for Joshua Moulinie in this riot.
They didn't even charge him, just asked him to write an apology letter to the targeted business.
Maybe true, but that is only exactly true if there is no vertical load applied on that bar tied to its "waist" .
Also that thing does not have any hands to provide the balancing force for the fact that one of its legs is ahead of the other, instead it uses a strange load on top of its "neck" to balance itself. That basically means that the "hips" of this robot does not twist like a human being walking. So the adjustment for forward balance actually prevents it from being laterally balanced.
(Human walking being right leg forward, left hand forward and the hip rotated slightly to the left providing the lateral stability).
Actually it is even worse than that.
In the absence of HFT's the alternative is to have specialist market-makers at exchanges. That would mean an oligopoly of people who can buy a seat of exchanges making money on bid ask spreads. They were kicked out of exchanges after a set of large investors sued them for front-running their trades http://www.businessweek.com/bwdaily/dnflash/dec2003/nf20031217_6559_db016.htm.
If you had neither specialists nor HFTs, a trade could only be done in case there was a buyer and a seller. It would be great in case of normal markets, but in case of market panics there would be far more sellers and buyers causing price swings to worsen. It is believed that this would be bad for liquidity and markets in general.
There are some other ways proposed to solve this problem - some have suggested an auction every second, with no trades done in between those seconds. IE there would be only 28,800 "trades" in an 8 hour day. Every order that comes in between is consolidated. It is not clear if that would still help because presumably some computer would figure out an algo. that could beat the auction.
Thanks for the link. It is pretty spot on.
At the risk of being off topic.
But that said the latest crisis was not one of unexpectedly large standard deviations alone, but one of badly modeled correlations. The article is correct that the variance of a portfolio is larger than it looks, but many in the equity markets had figured that out. What happened in this recent crash was that they used scant data from mortgage markets to try and estimate correlations. For eg: Biggest possible loss for Californian mortgages was say 10%. Adjusting for a large variance, you could say that the chance of loss is say 50%. You could account for that in some way, say.
But what if someone bought a bond that would only default (not pay) if 40% of bonds in 20 states defaulted? That's where the correlation came in. They assumed in a rather complicated way that the chances of bonds defaulting are related to the correlations of bonds in different states. After all, a default is just a very large negative move.
Correlations between bonds in different states were pretty low because each state had varying economic conditions etc. So plug that into the model and you would find that the newfangled bond that would only default if 40% of bonds in all 20 states defaulted was very safe. So everyone called it a AAA.
Except it turned out that all the states were on a property bubble fueled by low interest rates and greed and had a correlation of 1 in bad times. Hence the mess we are in.
Anyway, what I am trying to say is that Fractal model might be very good at estimating variances , but still would not solve the problem of estimating covariances.
The story pretty much says the same thing.
AT&T has said that it will cost them $39 Billion to buy TMobile. It has also claimed to investors that it will save them $10Billion.
The justification is that it will enlarge Wireless coverage in USA.
The leak now claims that the expanded network coverage will cost only $3.8 Billion http://www.wirelessweek.com/news/2011/08/unredacted-ATT-filing-shows-high-price-tag/
So AT&T Pays $39 Billion, saves $3.8 Bilion in network costs and $6.2 Billion in non-network costs (say closing and selling stores/laying off duplicated jobs). Assume that Sprint is worth the same as TMobile intrinscially - which is approx. $10B. The remaining $19 Billion premium paid must then be the cost of eliminating competition or the cost of keeping sprint from expanding.
Google seems to have fixed the error message on their code search
Also +"<FILEHARD>" seems to do better than simple quotes - The "+" forces google to look for single word Filehard.