Thursday, March 27, 2008

Shock news: Humans don't act as economists say they should! (update 2)

Austrian economists have known for a long time that people don't act the way mainstream economists say they do.  But mainstream economists and the New York Times are only just finding it out themselves [hat tip Brain Stab]. 

"Anomalies" is what one Nobel prize winner called behavioural departures from the rationalistic idea that people act as mainstream economists say they should, ie., with both 'perfect' knowledge and 'perfect' self-interest.  This wouldn't be important to the rest of us, except that when these 'anomalies' occur in ways that attract the attention of the business pages the more state-worshipping economists start to yell about "market failure" -- whining in other other words that people don't fit their "models," and calling for government agency to make them fit, ie., to"fix" a problem that doesn't actually exist.

But the way people behave is only anomalous to people who think other people should behave the way their equations say they should.  The fact is that economics is not a science of human choice, as mainstream economists insist it is; it is instead a science that studies the result of human action, however mistaken the graduates of business schools might think those actions to be.

Rather than devising and formulating abstract, arbitrary models predicting the choices people make, and then throwing up the hands in horror when people over-eat, over-spend and generally act in ways other than those predicted, Austrian economists base their system of thought on one very simple principle: Man acts.  Specifically, he acts in ways he think will move him from a less satisfactory to a more satisfactory state of affairs, whatever the hell that might actually be. He might be wrong in the way he acts to achieve his goals, and he might even be mistaken in the goals or choices themselves (and people might even act in ways they explicitly disown) but they act nonetheless, say Austrians, and in their actions they seek to further their own ends (whatever these might be) -- it is with this 'action axiom,' says economist Ludwig von Mises, that the science of economics begins.

A shame that mainstream economists and the New York Times's writers haven't read more Mises or observed more human activity before they burst into print calling what we do every day "anomalous."

UPDATE 1:  New Zealand's funniest bloke, John Clarke, has the testimony of an unusually truthful  "economics expert" appearing before a Royal Commission on What the Hell's Happening to the Australian Economy. The relevance to this post can be seen almost from the first exchange.  Here's a small sample:

BARRISTER: Mr Trouser, you've told us over a 3 month period you collect information and analyse it for trends...  So what's going to happen Mr Trouser? In the next few months?  What's a likely next development in the economy.
TROUSER: I've no idea.  I don't have the current information.
BARRISTER: I thought this [holds up telephone book sized sheaf] was the current information.
TROUSER: They are the current figures, but they're not based on the current information.
BARRISTER: The current figures are not based on current information?
TROUSER: No. The current information is coming in now.
BARRISTER: So what are the current figures based on?
TROUSER: They're based on the information available at the time the current figures were being prepared.
BARRISTER: Which was when?
TROUSER: 3 months ago.  The figures take 3 months to assemble.
BARRISTER: Is it possible for you to tell what's happening now?
TROUSER: Yes, of course it is.
BARRISTER: When could you do that?
TROUSER: In 3 months.

Read it all.  It's hilarious.  And then send the link to your nearest economist.

UPDATE 2: Clifford Thies draws parallels between the predictions of mathematical economics, and the equations and prediction of pro-warmist super computers.

The emerging discipline of climatology is an interesting one. It has no laboratory. Instead, various measurements are put into computer models to see the extent to which they are consistent with the hypothesis that human activity has contributed to the trend of global warming. Unable to conduct experiments, all climatologists can do is examine statistical correlations.

In my field of economics, we have generally dismissed inferences based on mere correlations. Such things as interest rates, inflation, the unemployment rate, and real GDP growth are highly trended. Almost necessarily, they are highly correlated over discrete periods of time. This correlation doesn't prove anything regarding cause and effect, and the still short history of econometrics is littered with theories such as the Phillips Curve that once enjoyed a consensus among economists, that are now understood to be little more than statistical illusions.

Today, the econometric standard for testing theories involves determining if changes in one variable tend to be followed by changes in another variable. Thus, the sequence of fluctuations in variables is seen as a key to discerning cause and effect. This new standard has been humbling in revealing just how complex are macroeconomic phenomena.

Commenters on this thread are recommended to read and digest.

Labels: , , , , ,

40 Comments:

Anonymous Falafulu Fisi said...

PC may I informed you that anomaly detection is what make algorithmic trading so successful. Anomaly detection originated in Pattern Recognition & Statistics, which are of course , highly mathematical modeling. And this anomaly was not based on Thaler from that article you linked to, but original idea from the statistics & pattern recognition community alone.

I noted your distaste of financial/economic modeling that you don't quote anything else but whatever Mises said, of which he disapproved.

Here is some quotes from the article about Thaler:

Thaler set out to prove that it did not. His first effort, a 1985 paper with Werner De Bondt, his doctoral student, showed that stocks tend to revert to the mean -- that is, stocks that have outperformed for a sustained period are likely to lag in the future and vice versa.

But you must have read the article that I linked to from the other thread about mathematical economist that this is what algorithmic trading software is exploring, because opportunities don't hang around too long. They might be in there for about a few seconds, minutes, hours etc,... but the model needs to detect this anomaly and trade immediately before the opportunity disappears. This type of trading uses economic mathematical modeling. Note that Thaler set out to prove. Would you care to clarify what sort of proof that he would have done? Here is what I think. The guy must have used statistical hypothesis testing (SHT) to prove whatever he was trying to prove. Without using SHT, it means that whatever he observed was anecdotal. But using SHT is mathematical modeling which is the sort of thing you & Mises dismissed. Do you accept anecdotal evidence in this world or rigorous proof? I know you like rigorous proof, but of course you will disregard it when it suits you, simply because Mises didn't approve of those methods.



This was a finding that Chicago School types couldn't ignore -- according to their theory, no pattern can be sustained, since if it did, canny traders would try to profit from it, correcting prices until the pattern disappeared.

Again, this is algorithmic trading and it is highly mathematical. Every one knows that opportunity don't hang around for long in the financial market, but models are used to alert the patterns or signs of those opportunities when they arised.


Thaler, actually, is a director in a California money management firm, Fuller & Thaler Asset Management, which, according to figures it provided, has been beating the market handily since 1992. The firm tries to exploit various behavioral patterns, like "categorization": when Lucent Technologies was riding high, people categorized it as a "good stock" and mentally coded news about it in a favorable way. Lately, Lucent has become a "bad stock." But Thaler, who does not get involved in picking stocks, stops short of suggesting that investors versed in his research can beat the market. Mispricings that spring from anomalies are hard to spot, he says, particularly when the people looking for them are prone to their own behavioral quirks.

Again, that is why algorithmic trading (highly mathematical) is there in the first place, because computers react faster than a human to spot the opportunity.

If this sounds muted, it may be because Thaler is ready to declare victory and join the establishment. The neoclassical model, he admits, is a fine starting point; it's misleading only when regarded as a perfect or all-encompassing description. People aren't crazy, he adds, but their rationality is "bounded" by the tendencies that Kahneman, Tversky, himself and others have studied. What he hopes is that a future generation will resolve the schism by building behavioral tendencies into a new, more flexible model.

Well, you can rant all day what you like about mathematical modeling, but Thaler person (and Mises alike) cannot ignore the fact that this world, there is only anecdotal evidence (seem to agree with observations but no rigorous test applied at all) and one that works (in agreement with observation via rigorous test).

3/27/2008 02:45:00 pm  
Anonymous Anonymous said...

TA is used in all trading rooms - forex is not traded in any other way. So even if it seems irrational the chart becomes a self-fulfilling prophecy. You can't ignore the charts - even Frank Shostak said that.

You don't have to like it or agree with it philosophically to make money out of it. That's capitalism, and that's reality.

3/27/2008 03:19:00 pm  
Blogger Elijah Lineberry said...

Falafulu, do you know anyone engaging in the rather 'over complicated' mathematical modelling you mention who is rich?

Rather than engaging in a lot of nonsense trying to predict the future, the rich and successful traders just 'get on with it'.

I doubt a single speculator Billionaire such as Carl Icahn, Warren Buffet or George Soros uses any of that sort of thing...(or would know what it meant, anyway)..they just stick to what they know.

On the other hand, a lot of unsucessful, broke, know-it-all smart alecs in trading rooms around the World waste their time making relatively simple activities extremely complicated.

3/27/2008 03:42:00 pm  
Blogger Paul Walker said...

The behavioural economics stuff is certainly interesting but its not clear just far it does go in an attack on the standard models. Steven D. Levitt and John List have a short commentary in a recent issue of the Science journal on topic of behavioural economics. They see positives in the behavioural approach but they still remain somewhat skeptical. They write,

Perhaps the greatest challenge facing behavioral economics is demonstrating its applicability in the real world. In nearly every instance, the strongest empirical evidence in favor of behavioral anomalies emerges from the lab. Yet, there are many reasons to suspect that these laboratory findings might fail to generalize to real markets. We have recently discussed several factors, ranging from the properties of the situation — such as the nature and extent of scrutiny — to individual expectations and the type of actor involved. For example, the competitive nature of markets encourages individualistic behavior and selects for participants with those tendencies. Compared to lab behavior, therefore, the combination of market forces and experience might lessen the importance of these qualities in everyday markets.

So the case for behavioural econ is still unproven.

3/27/2008 04:25:00 pm  
Anonymous Falafulu Fisi said...

Elijah said...
Falafulu, do you know anyone engaging in the rather 'over complicated' mathematical modeling you mention who is rich?

Yep Elijah, I know (not personally) of 2 billionaires who have been using complicated mathematical modeling to beat the market.

- Kenneth Griffin

- James Simon

Elijah said...
Rather than engaging in a lot of nonsense trying to predict the future, the rich and successful traders just 'get on with it'.

Nonsense according to what?

Elijah said...
I doubt a single speculator Billionaire such as Carl Icahn, Warren Buffet or George Soros uses any of that sort of thing...(or would know what it meant, anyway)..they just stick to what they know.

Man, don't you know that Warren Buffet's fund managers use quantitative modeling (mathematical modeling)?

Elijah said...
On the other hand, a lot of unsucessful, broke, know-it-all smart alecs in trading rooms around the World waste their time making relatively simple activities extremely complicated.

Nope, wrong. Jesus, do you think that market dynamics is simple? Man , just ask PC why he thinks that economic modeling is bollocks? Yep, PC will tell you that his reason is that the market is complex that the models are unlikely to capture its behaviour in realtime, and PC is correct in that. The market is complex, and if you think otherwise, then our current models can predict exactly what will happen tomorrow, next week, next year and so on. But our models can only give some indications of the future, one chooses to act on those useful indicators and perhaps benefit from it or perhaps ignore it entirely, thus using their own intuitions. There is nothing wrong with intuitions, but sometimes emotions creep into ones intuition which blinded him to the facts and rational thinking therefore leading him/her to act irrationally. Models don't have emotions at all and that's the advantage. Now I can quote you a comment from an article (How a Computer Knows What Many Managers Don't) that I linked to, on another thread. Here it is:


But perhaps the most important attribute of quant funds is their superior risk control. Most models keep sector and equity picks within the benchmark attributes they are built around. This process reduces the risk that may otherwise be introduced by active managers who decide to make a bet on a particular sector or stock.

And that leads to what may be the most attractive aspect of a quantitative fund.

"With quant funds there is no emotion in the investment process," said Jeff Mortimer, head of equity portfolio management at Charles Schwab.

In other words, the investor is not exposed to the whim of an active fund manager who travels to a company's offices and whose investment decision may be partially based on his very complex, but unquantifiable, human experience while there.


I will post an article later that was published in the Herald last year (2007) , about algorithmic trading and superiority of mathematical modeling.

3/27/2008 04:53:00 pm  
Anonymous Falafulu Fisi said...

Here is the original Herald article Elijah:

Humans made redundant as super-trader does the sums

NEW YORK - Louis Morgan, managing director of hedge-fund firm HG Trading, has never talked to his best trader. That's because his best trader is a machine.

Morgan's top earner is a computer with software than can monitor thousands of stocks simultaneously and respond in less than a blink of an eye when opportunities arise.

"Doing what we do by hand would be impossible," said Morgan, who focuses on statistical arbitrage - taking advantage of sudden and potentially profitable price anomalies between securities that usually trade in correlation.

Morgan uses a computer trading system based on algorithms, complex mathematical formulas that quickly weigh a huge number of possible trades and execute orders in milliseconds (a millisecond is one thousandth of a second).

He is part of a rapid and relentless trading revolution that has transformed financial markets.
The increasing adoption of algorithmic trading - "black box trading systems" - is changing the way Wall St works and is creating new royalty with billion-dollar-plus incomes.

Former mathematics professor James Simons, who was one of the best paid hedge fund managers in the world last year, made US$1.7 billion ($2.3 billion), says Alpha magazine.

But there are predictions that the number of traders will sink by as much as 90 per cent by 2015 and that most of those left won't need traditional Wall St skills but will be whizzes in math, statistics, computer science, astrophysics and linguistics.
The presence of so much computer power has reduced short-term market volatility, but it also is giving regulators sleepless nights.
Many wonder what would happen if a rogue trader with rocket-scientist skills had control of a Wall St firm's black boxes, or if the machines turned a market slide into a meltdown before anyone could trigger a halt.

The change is even affecting where big banks and funds do business as they seek to get closer to exchanges and news services, so they don't lose precious seconds through delays in transmission of information.

And financial news organisations are increasingly focusing on machine-readable, as well as human-readable news.

About a third of US equities trading is being done using algorithmic trading, and Brad Bailey, a senior analyst at the Boston researcher Aite Group expects that figure to soar to more than 50 per cent by 2010.

"I'm even afraid I'm underestimating that number," Bailey said.

The London Stock Exchange says about 40 per cent of its trading is algorithmic.

"It's becoming much more mainstream," said Guy Cirillo, manager of global sales channels for Credit Suisse's Advanced Execution Services unit, an algorithmic trading platform that serves major hedge funds and other buy-side clients.

"The traditional firms that took longer to adopt have come in strong in the past year to two years," he said. "If you are not using this type of technology, you are at a serious disadvantage."

Algorithmics can be used to execute almost any strategy.
Some aim to capture fleeting price anomalies that generate thousands of buy and sell orders every second.

Others slice up a large trade into smaller trades to mask intentions and prevent rivals squeezing the trader on price.

That ability to trade quietly and anonymously solves one of a large firms' biggest headaches about using brokers - information leakage.

"You don't want anybody to know that you are trying to buy, say, 10 million shares of some small-cap stock," said Adam Sussman, senior analyst at hedge fund research firm the Tabb Group. "Otherwise, the price is going to shoot up."
Algorithmic trading has reduced volatility in equities and foreign exchange markets, as many of the programs are based on mean reversion, so that as things go astray, the programs pull the market back to the norm.
But algorithmic trading has had the opposite effect on financial and corporate news.

The appetite for information on everything and anything on Wall St has never been greater as data-hungry computer programs can absorb and respond to new and old input.

Software tools can analyse years worth of news stories to see how certain headlines affect certain stocks and asset classes and use those patterns to programme computers to trade on the latest news developments.

News and financial organisations such as Reuters, Dow Jones and Thomson Financial, which is acquiring Reuters, have jumped into the fray, making sure they present stories in ways computers can understand.

Louis Morgan said that five years ago, creating an algorithmic trading system was expensive as everything had to be done in-house. Today, he works with two people and a software program to build his algorithms.

The effort has paid off. "We are profitable on a little more than 60 per cent of our trades," he said.
While such a win rate may not look great on a traditional trading desk, it is very profitable when computers are trading much greater volumes.

Such success in what is being called "the algo wars" may mean that it is not only the exchange floor traders whose jobs are threatened but many of those sitting in front of screens, too.

Last year, an IBM Institute for Business Value study found that of every 40 traders active today, only four will be left by 2015 because of electronification of markets.
"The four traders will be the stars that assume risk, achieve true client insight and, of course, consistently beat the market," says the study.

And those survivors are set to do very well.

Simons, the 69-year-old founder of Renaissance Technologies in Long Island, New York, is not alone in the new billionaire class.

Ken Griffin, the 38-year-old founder of Citadel Investment Group in Chicago , took home US$1.4 billion last year, up from US$240 million in 2004, according to Alpha.

Both employ rapid-fire quantitative techniques designed to exploit minute price differences through algorithms.

But behind the proprietary mathematical models are the developers - mathematicians, astrophysicists and computer scientists - who make algorithmic trading successful.

"We haven't hired out of Wall St at all," Simons told the International Association of Financial Engineers annual conference last month.
Of course, the high-powered programs are also available to traders who might use the same technology and aggressive tactics to steamroll other market players.
There are fears, too, about those whose algorithms might fail.

"They can pump in more orders per millisecond than a person can, so there's always the concern that they can melt a firm and blow it up in a matter of seconds," said Edward Dasso, who directs market surveillance for the National Futures Association.

Market regulators worry that the kind of trading Nick Leeson did over several years to cause US$1.4 billion of losses and the collapse of Barings Bank in 1995 could be done today in milliseconds - with similar results.

The algo wars have become so fierce that location has become a crucial element in the quest to shave milliseconds off execution times.
Exchanges have been providing locations for companies that want to place their servers in proximity to the exchanges' systems.

The length of time it takes bits of data to travel from one city to another can make a difference.
"The trading standard now is sub-one millisecond," said Philadelphia Stock Exchange CEO Sandy Frucher at the Reuters Exchanges and Trading Summit last month. "If you get faster than sub-one millisecond, you are trading ahead."

The growth in black box investing goes beyond stocks.

In three years, 25 per cent of all foreign-exchange trading will be conducted with algorithmic trading, up from 12 per cent now, says the Aite Group.

By 2010, algorithmic trading in fixed-income securities will have grown to 12 per cent, up from about 5 per cent, and algorithmic trading in options is projected to increase to 20 per cent from 2 per cent.
About US$480 million is being spent today on information technology directly attributable to supporting algorithmic trading of US equities, says the TowerGroup, an industry research and advisory services firm in Needham, Massachusetts.

IT spending for algorithmic spending is projected to grow at a 15 per cent compound annual growth rate, which puts 2010 spending at roughly US$750 million, TowerGroup added.

Credit Suisse's Cirillo said the firm's stock algorithms had become so popular that clients have requested that they be used in other asset classes.
"We found that clients wanted to take positions not just in stocks, but also in foreign exchange, options and futures," he said.
Fast and smart

* Algorithmic trading uses computer software based on complex mathematical formulas that quickly weigh a huge number of possible trades and execute orders in milliseconds.

* About a third of US equities trading is done using algorithmic trading.

* Only four of every 40 traders active today will remain by 2015 because of the electronification of markets, says an IBM Institute for Business Value study.
- REUTERS

3/27/2008 05:03:00 pm  
Anonymous Falafulu Fisi said...

Here is a very good blog about mathematical & quantitative modeling (advanced analysis & algorithmic trading).

What is 'quant' or 'algorithmic' trading

There were times when the old guards dismissed anything that is new that perhaps it can't be right because it is too complex. The old guards hang on to their previous way of doing things (or course ineffective) and watch how new innovative methods just bypass them. The old guards now realized that sticking to old methods is not going to give them any competitive advantage, so at the end, they jumped the bandwagon and joined the new fads, because they can see advantage. And my point here, is that by about 2015 (from the article that I posted in my previous message), stock brokers will be out of the game, since trading becomes automated.

3/27/2008 05:12:00 pm  
Anonymous Anonymous said...

Falafulu, do you know anyone engaging in the rather 'over complicated' mathematical modelling you mention who is rich?

Rather than engaging in a lot of nonsense trying to predict the future, the rich and successful traders just 'get on with it'


And how do you think they 'get on with it"? By abacus, horoscope?

I knew you were blowing smoke about being a gun forex trader - you may be able to fool your homies at SOLO but you can't fool the REAL traders - who have never heard of you BTW.

FX is traded by technical analysis - nothing else. Your question just shows you up. If dealers and corporates could not make billions out of TA they would would not be using it.

A genuine dealer would know this.

3/27/2008 06:28:00 pm  
Blogger Richard said...

You defame economists PC. That Clarke piece is more relevant to statisticians than economists!

3/27/2008 06:40:00 pm  
Blogger Elijah Lineberry said...

Anonymous, I am not a Forex trader...(where did that idea come from?!?)...I trade Index Futures, Contract for Difference shares, and (occasionally) Gold, along with writing call options on shares.

With regards to how they 'get on with it' here is a real life example...

Today in Australia the Sharemarket was down, as such, I made money selling the share index futures short.

When the market had 'bottomed out' I bought them back again, before selling on the close for a fairly successful day.

In my experience most market trading is based on irrational behaviour and speculation, rather than on fundamentals or long term investing.

The reason the market was down in Australia was for no other reason than the US sharemarkets were down 'overnight'...(hence my observation of 'irrational' behaviour)

No whizz bang rinky dink computer programme can overcome the unfounded and irrational fact that the market was "down because the US markets were down" *shrug* which everyone was engaging in.

With regards to no one having heard of me in Forex circles, I am delighted to hear it!

No one has made money trading Forex in New Zealand ever, and for a very good reason...it is a very risky activity and the risks people take invariably fail.

The sensible traders are busy making money out of shares and other commodities...(whilst ringing me when they get bored for a chat)

3/27/2008 06:58:00 pm  
Anonymous Anonymous said...

Corporates use CFD's to manage their risk - traders use TA re CFD's.

I recall a conversation on SOLO where you said you traded forex and admired others who said they did. Don't make me look it up.

So you are saying you simply use your intuition on calls and puts. This is the equivalent of horoscopes.

No one has made money trading Forex in New Zealand ever, and for a very good reason...it is a very risky activity and the risks people take invariably fail.

Really? Stop embarrassing yourself. Forex is a zero sum game. For every loser - and I presume you are one - there is a winner. And the govt is not taking a gaming tax cut.

And gold and the gold standard is just plain irrational - I agree with you on that one.

3/27/2008 07:19:00 pm  
Anonymous Anonymous said...

Oh and I think you may be getting confused by a trader and proprietary trader.

They are quite different in respect of the thought process behind them. So I may have been too harsh on you, although most proprietary positions don't last interday.

3/27/2008 07:21:00 pm  
Anonymous LGM said...

FF

What you are writing about is not economics. What you are discussing is automated trading, operated according to arbitrary algorithms processed and executed by machines. That is a point you'd do well to understand.

---

Some asides.

Interestingly, the article you posted mentions the great risks of the utter melt down of a company as the result of automatic transactions undertaken in micro or milliseconds. No human will know it is happening until far too late to intervene. Then....disaster. Too bad about Ma and Pa Kettle's superannuation; it was vaporised faster than 60cc of rasping wet fart!

In effect you have created auto-Leeson (copyright claimed by me for inventing the Leeson as a metric for trading losses, eg milli-Leeson, centi-Leeson, kilo-Leeson, mega-Leeson, giga-Leeson etc). Imagine, a mindless rouge in a box doing all your trading. There is no way to effectively stop loss such a device unless you intend to slow the trades right down and limit what can be accomplished automatically. That requires manual supervision by humans... gasp!

The machines do not really KNOW what is going on just as surely as the geeks who wrote the software don't have a clue about the future either. Get this clearly in mind. You do not know exactly what results future automated use of your algorithms will yield. That's because you do not know all the situations and contexts in which they will be employed and what the circumstances of the time will conspire to produce, for or against.

In the end these machines mindlessly apply formulae, look up tables and output an action. Hopefully it is the correct gamble to take, but you do not know that.

--

A serious concern with this type of activity is that all we are experiencing here is a form of rent seeking, diverting portions of existing trading streams to different pockets simply because it can be accomplished (some of the time). In short, a classic case of ticket clipping.

Note how all this auto-trading action is going on at the margin. None of it is going on in the boiler-room of productive enterprise.

--

What I expect to occur is the casino mentality of market participants over the last few years is going to significantly cool now. There is going to much more old fashioned regulation and state intervention in the markets. While I'm no fan of this type of thing, it is no secret intervention is being encouraged by the rise of "mathematical" fortune telling.

Astrophysics majors and mathematics puzzlers will continue to proclaim their cleverness. They get paid to work both sides of the street; trading and regulating. All it'll need to bring in a much more restricted finance and investment climate will be one or two spectacular trading collapses. Then Ma and Pa Kettle will shriek for the regulators to "DO SOMETHING" to get their super back. Other Kettles will bail out to safer climes. Watch and see!



Meanwhile, if it is economics you seek to understand, then it is important to avoid confusing the subject with trading stocks, bonds, comodities, derivatives or whathaveyou. Not the same.


LGM

3/27/2008 07:24:00 pm  
Blogger Elijah Lineberry said...

Yes, anonymous, I would suggest you looked it up and produced the quotation where I claimed to be a Forex trader.

I strongly suspect (due to the 'anonymous' part of your post) you are that silly Natalie Harrington who has such a sad, empty life she follows me all over the internet trolling and posting silly messages.

(You have been spotted)

I do not 'trade' puts and calls (that game is for prize bunnies)...I said I write call options, quite a safe activity (and rather lucrative)

With regards to Forex trading, I think anyone who holds a position for longer than a few hours is a mug.
Long term positions are for idiots...(just ask anyone who was foolish enough to hold one)

3/27/2008 07:30:00 pm  
Blogger Oswald Bastable said...

People- leave the long, overlong posts to the left!

If you can't sell the idea in a couple of hundred words- we have lost!

example:

Q- Why don't we understand economists?

A- We live in another frickin' world- the real one!

3/27/2008 07:59:00 pm  
Anonymous Anonymous said...

Lineberry you said it after your meeting at the Birdcage or something.

Remember the NZ trading community is very small. You have shown your ass and lack of understanding of trading on this thread and I hope Libz notice it.

Oh and no one phones brokers anymore. They process deals online.

3/27/2008 08:11:00 pm  
Anonymous Falafulu Fisi said...

LGM said...
Meanwhile, if it is economics you seek to understand, then it is important to avoid confusing the subject with trading stocks, bonds, comodities, derivatives or whathaveyou.

What is your definition of economics that is different from this economics definition?

Is it the definition of Lord Almighty Mises that you adopt? Doesn't it occur to you that economics is a huge field? What I have been describing here on this thread is economics.

It seems that you just can't get your head around it. Some disciplines of economics is quantitative and you can spin all you like but that is fact.

You're doing wordsmithing again. I think that you have a problem with trying to apply logical deduction & induction too much, that everything that doesn't suit your preconceived ideas must be dismissed since it doesn't fit in with Heiress AynRand or Almighty Mises.

you said...
All it'll need to bring in a much more restricted finance and investment climate will be one or two spectacular trading collapses.

No, this has happened already in 2000 with Long-Term Capital Management. But as a Physicist from Oxford who quoted that the demise of LTCM doesn't mean that the models are useless, because 99% of the time it is correct, however its the 1% chance that it could fail just happened to occur. The point is, would you use horoscopes to guide your portfolio management, perhaps just stick to your best judgment or use models?

you said...
The machines do not really KNOW what is going on just as surely as the geeks who wrote the software don't have a clue about the future either.

I think that your understanding is fucked. No, one is going to know what's gonna happen tomorrow. Unless you can show me, where in the publications related to market modeling that says that. However a model could only give a rough guide to what will happen. This is not psychic or fortune teller. You're so hooked up in semantics of such words as prediction, forecasting, future projections that you equate them to an ability to see what is going to happen tomorrow, which is completely wrong. In your confusion on word semantics, you might as well tell your accountant at your work to eliminate terms such as sales forecast from his/her monthly financial statements as this might imply that your company foresees the future sales figures? Do you see how fucked is your reasoning is? But I know that you would have no problem with the term forecast being used by your firm's accountant, but you will dismiss it if it is used by financial market quantitative traders.

you said...
Astrophysics majors and mathematics puzzlers will continue to proclaim their cleverness.

Do you have a problem with that? Obviously you do, or perhaps you want them to proclaim that they're dumb?

3/27/2008 08:38:00 pm  
Anonymous LGM said...

FF

How about you calm down? No need for the histronics.

For the last two PC posts regarding economics you've raved ad nauseum about "mathematical" modelling for the purpose of auto-trading. There is one big problem; the elephant in the room. In all the raving you fail to seriously discuss economics or demonstrate even a basic understanding of that subject. Sure, you've linked, cut and pasted some interesting articles written by OTHERS. What you have not understood is what they are reporting about. Also, didn't you notice that the warnings regarding limitations and the dangers of the activity you are fixated upon? No? Oh? Guess you missed those bits then.

Had you considered YOUR approach to "mathmatical" modelling is really based on a blind faith? It is merely your personal religion and just as misleading as the more conventional variety.

Check your premise. Mathematics does not define, control or predict what occurs in reality, just as surely as tea leaves don't. You need to understand the attributes of the entities you are dealing with in order to attempt understanding or forecasting. In this case, you need to sit down and learn about what economics actually is and what it means. All the magic numbers and computers will not help you achieve that.

A few simple points:

1/. You need to understand at least some economics to undertake the modelling of an economic system. Clearly you do not have this type of fundamental insight. Skills in maths will not compensate for that lack.

2/. A model is of limited capability in forecasting what is going to happen in the future when dealing with a social system. Economics is about a society of individual humans. It is a result of applied politics...

3/. What your activity is really directed at seeking, is a ticket clipping of existing revenue streams. This is what is known as a "scheme". In this case it is thought worthwhile to seek profits from multiple trading transactions, automatically executed in fractions of a second and intended to exploit small time lags and small price differentials. That is a different proposition from understanding economics or even of economic forecasting.

4/. The rise of the auto-trading activity can be expected to see the profits from that activity decline. The reasons range from regulation to attraction of new competition and eventually some spectacular melt-downs (some biggies yet to come).

5/. The best definition of what economics actually is was given by Prof George Reisman in the book, Capitalism. You'd do well to read the whole of it in order to strainghten yourself out.

6/. You need to learn to read and comprehend what is being pointed out to you. In your hurry to defend your self-worth (as a practioner of mathematics and software development) you are failing to consider what PC and others, including me, are telling you about. So far you have not addressed ANY of the fundamental points raised.

7/. Unfortunately you have already been caught in a lie. Von Mises did not alter his views regarding the casting of numbers as a substitute to possession of a real understanding of the subject of economics. Wishing he did does not alter the fact that he didn't. Claiming he did does not alter the fact that he didn't. Building a mathematical model to pretend that he did will not alter the facts either. Think on that.



There is more, but this'll do for now.

LGM


PS Yes, I have scorn for hucksters promoting their self-perceived cleverness, especially when they posses limited understanding of what they are actually doing.

3/28/2008 07:29:00 am  
Blogger Elijah Lineberry said...

LGM makes an excellent point that profits would start to decline if everything moved to automated trading.

The entire premise of automated trading is illogical, as if you can somehow 'turn on the washing machine' and come back later and find you have made millions.

Most profits are made through instinct, experience, emotion and timing...a computer programme cannot do that.

I often feel sorry for new chums who are told how much money they can make by clicking a couple of buttons on a computer, rather than the more logical training method of 'trial and error', and getting a feel for market trends, and experience.

3/28/2008 08:13:00 am  
Blogger PC said...

Oswald, you said, "People- leave the long, overlong posts to the left!

If you can't sell the idea in a couple of hundred words- we have lost!"


Well, ridicule takes only a sentence or two, whereas persuasion usually takes a sentence or two longer.

I fear however that length is not synonymous with persuasion. As evidence to this end I offer our friends Falufulu Fisi and Mr Wikiriwhi, whose persuasive powers lie in inverse proportion to the length of their posts.

3/28/2008 09:41:00 am  
Blogger PC said...

Falufulu Fisi, vous accussez "everything that doesn't suit your preconceived ideas must be dismissed since it doesn't fit in with Heiress Ayn Rand or Almighty Mises."

Not true. In the field of economics I have many more heroes than just those two luminaries and Professor Reisman.

I'm happy to take advice and wisdom from Adam Smith and David Ricardo, from James and John Stuart Mill, from Carl Menger, Eugen von Bohm-Bawerk, and Henry Hazlitt; from Israel Kirzner, Bernard Siegan and Mark Skousen; fronm Jean-Baptiste Say and Frederic Bastiat and Nassau Senior; from Larry Sechrest and Kevin Dowd and Frank Shostak; from John Ridpath, Yaron Brook, Richard Salsman, Walter Williams and Thomas Sowell.

I'll take economic advice from all of these and more. How 'bout you?

I'll even take advice from Hayek and Milton Friedman and James Buchanan when I have too - and even from Rothbard when he's rational.

How 'bout you?

I'm also aware of causality -- on which the science of the production of wealth rests -- and also of the multiple complexities of markets and human choice, and especially I'm aware of the free will of human beings, something no mathematical model can ever account for.

Mathematics is silent on causality, just as mathematical economics is silent on free will. It's this last that is most violated by the claims of mathematical economics to be "scientific" -- after all, without experimental proof, where is the science? Without control groups and experimentation and the like, where is the derivation of all the functions that mathematical economists dream up and then insist (without experimental proof) that human beings follow?

Free will indicates nothing more than that man acts, and generally acts to move from a less satisfactory situation to a more satisfactory situation.

We can derive this 'action principle' simply by observation and induction, but it's more than enough from which to found the science that studies the creation of human wealth (in other words, economics).

The chief problem with mathematical economics (and I'm talking here not of short-term aids to technical analysis and trend-spotting, but of all the many formulae used purportedly to describe and predict human action) is not just that the mathematics needed to describe accurately the behaviour of the thousands of interlocking markets and millions of thinking, choosing acting human beings is even more complicated than the close-coupled equations needed to describe the mechanics of global warming -- it's more problematic even than that.

You see, while molecules and mathematics are deterministic, human behaviour is not.

Mathematical economics seeks at root to either to force human beings and human choices into the Procrustean bed of their formulae, and (by citing "market failure" if human beings remain recalcitrantly unforced) seek the government's help to do so.

I'm against that. How 'bout you?

3/28/2008 11:04:00 am  
Blogger PC said...

Just to add to that list of economists from whom I'm happy to take advice, there are some local luminaries I neglected to mention -- but who if they tell me to jump economically, you'll generally find me asking how high. The list includes Paul Walker, Greg Balle, Tony Endres, Ian Hunter, Eric Crampton, Julian Darby & Ashley Chan.

Heroes all, in my book.

3/28/2008 11:36:00 am  
Anonymous Craig said...

So have we worked out who has the biggest ham yet?

3/28/2008 12:37:00 pm  
Anonymous Anonymous said...

PC, I noted that you haven't responded to my message on that thread : failing policies of present, where I pointed out your contradicting argument about mathematical economics. Is there a reason why? Perhaps LukeH is correct when he quoted this from the other thread:

Ah I see - refer to Wiki articles when you like and agree with the subject, but make it "off-limits" when the content of the article doesn't suit your purpose.

As long as we've got that clear ...


You dismissed mathematical economics when it suits you (because Almighty Mises said that mathematical economics should be dismissed), but happily quoted Economic Multi-factor Productivity , which obvious that it is quantitative , which is against what Almighty Mises had preached.

3/28/2008 03:44:00 pm  
Anonymous Falafulu Fisi said...

The previous message was mine, forgot to click the choice name selection radio button.

3/28/2008 04:22:00 pm  
Anonymous Falafulu Fisi said...

PC said...
I'll even take advice from Hayek and Milton Friedman and James Buchanan when I have too - and even from Rothbard when he's rational.

How 'bout you?


Economics has many sub-disciplines similar to Engineering/Physics/Psychology/Mathematics, etc, and you've quoted some , which they do belong to one category.

Here are some that I followed their work, and again they're from a different discipline in economics. I bet that you think that economics is only a one unified discipline?

- William Sharpe

- Harry Markowitz

- Merton Miller

- Franco Modigliani

- Myron Scholes

- Eugene Fama

- Robert Engle

- Robert A. Jarrow

- Emanuel Derman

and many others.

3/28/2008 05:21:00 pm  
Anonymous Roger Rogerson said...

Falafulu Fisi you is taking the piss man. You're a bit of a dude eh? Having a laugh because you can't really be serious. A bit of a joke eh?.

You reckon that some physicis wallah (Professor Lord Doctor Witch-Doctor Mullah Ayatollah Seriouslydunno was it?) from Oxford said that the demise of LTCM didn't mean that the models were useless, "because 99% of the time it is correct, however its the 1% chance that it could fail just happened to occur."

Nope.

Buddy, it's 100%, not 1%.

The model regarding LTCP was dead wrong. Completely flawed and as false as a false thing can be. That's 100% dead wrong. Fake.

That leaves what? Model was accurate exactly no percent. Zero percent. Nadah! Nil! Zip! never got it right or even a little bit right at all. Did not know what was happening. had not got a tiny wee clue. 0% correct.

The physics guy stuffed up big time. His model was 100% wrong and it 100% failed to correspond with the reality of LTCM. It failed 100%, not 1%.

And today there is no LTCM around any more.

And that modle is still dead completely fake wrong.

Or mebbe it was a furphy all along.

Guess the physics wallah what made the model failed to notice some really important stuff. Like his modle was 100% w-r-o-n-g.

100% failure.

He asked whether you would "use horoscopes to guide your portfolio management, perhaps just stick to your best judgment or use models?"

Answer: Horoscopes may have been more accurate. Let's say that would be about 50% of the time. What means that they are a lot better than 100% dead freaking fake and flat bluddy wrong.

LTCP scores. Horoscopes score 50%. Models score............ 0%.




Your best judgement would likely have been much better than the model if you bothered to do some background research into the company and its dealings and who key management and staff were. What is called due diligence. That's the ticket to knowing about where your investment is going.

Model? With the startling lack of knowledge about the specifics about the company that the physics wallah suffered from his model was 100% useless. It was a 100% misleading. A waste of time it was.




100%


Got it 100% fucked up the bottom wrong.

Ouch!


Roger and out!

3/28/2008 06:33:00 pm  
Anonymous Falafulu Fisi said...

Roger Rogerson said...
Falafulu Fisi you is taking the piss man

LGM, are you resurrecting as Roger Rogerson now?

You can hide your name, but man I do use mathematical modeling to text-mine most of the posts here at Not PC to recognize patterns that belong to a specific name or pseudonyms. The text of your last post matches pseudonyms as Abdul, Banker, LGM, with a confidence of 87% probability.

I hope that my model is wrong, but as whoever is the person that is frothing above, as you said, the model was just 100% wrong.

If you doubt, then I am willing to show PC how the pattern recognition software works, because it is what I do everyday, mathematical modeling and algorithm development, be it economics/finance, spam detection, face/image recognition, speech recognition or whatever anything in this world that could be mathematically modeled.

BTW, text-mining and text-classification is not new, they've been around for a few years. Spam detection is text-classification ,ie, binary classification (2 class) an email is either spam or no_spam. What I do use in mining the Not PC blog is multi-class, ie, many users with their corresponding many messages.

3/28/2008 07:31:00 pm  
Anonymous Falafulu Fisi said...

Also the name Cleetus comes up very close to the clusters that formed by users, Abdul, Banker, LGM, including the Roger Rogerson who had just posted above.

3/28/2008 08:04:00 pm  
Anonymous Anonymous said...

Well done FF.

Someone get a horse head.

3/28/2008 08:43:00 pm  
Anonymous Craig said...

Falafulu, don't you have better things to do than argue with these pompous blowhards? For instance, punching yourself in the face for a couple of hours might be more fun.

This is not a blog for debate, its purpose is simply to espouse the immutable views of PC and his ilk.
If you are trading and doing well, I would just leave Elijah to his fantasy world.

3/29/2008 08:16:00 am  
Anonymous Anonymous said...

I'm pleased he joined in - this thread has been the best so far this year. And a couple of people got pwned. Lineberry is probably at this very moment phoning his broker b4 he runs off to wire the money to pay for the trade - hilariously out of date nonsense.

People's minds are set in concrete about modelling - for finance anyway - and no amount of debate or evidence will break the concrete.

It would be good if people got this riled about about things that REALLY matter in this country, instead of this bullshit.

3/29/2008 11:03:00 am  
Anonymous LGM said...

Falafulu Fisi

Try this.

Falafulu Fisi is S - - - e. No maths necessary. No maths model or tricky software required. It's a 100% fact of reality. Cast iron guaranteed. How about that!

There have been a number of points raised for your consideration and response. You fail to directly address them, instead dragging along any old red herring you can find in order to obscure the issues to hand. It is now obvious that you don't have a clue about the subject of economics (and certainly not about philosophy of science- you'd do well to check out Harriman's "Philosophic Corruption of Physics").

Actually, the point about the model being 100% wrong is important. Consider it.

It is not a matter of being 1% incorrect. The modellers were completely incorrect and the model did not reflect reality. It never did correspond. Not even when they first developed it. They had not a clue. The model was worse than useless. How many people were deceived? How much wealth was malinvested? Who is still fooled by it?

Yes, it is possible to make all sorts of excuses, including the 1% one. That does nothing to alter the fact that the modellers were not aware of the real situation. Instead what they did was construct a model in spite of it. They were ignorant.

The point is not that all models are invalid. It is that unless the modeller actually knows (that is, he holds specific knowledge of reality) about what it is he wants to model, then the model is a little fiction or worse, a deceptive lie. It is a seductive one, but remains false nonetheless. They may as well have thrown dice or cast horoscopes.

This is important as it means that the idea of setting up a model (or a mathematical equation such as Schroedinger's) on the basis of arbitrary notions (as you previously have claimed) is wrong; doomed to deceive.
[BTW your previous characterisation of Scroedinger's approach is not factual. Perhaps you should read his correspondence.]

In order to successfully model something, one must know about the attributes of the entities that one would like to model. Understand that the maths is nothing more than a descriptive language. If one want's the model to work out (or to have a reasonable chance of actually being useful), then the description must be of reality, of that which exists. In other words, possession of real knowledge is a necessity, not arbitrary notions.

You must know about the real entities before you abstract certain of their attributes into your model. You must know about the real entities in order to determine which attributes are key and must be taken into account as opposed to those which are not significant for the purposes of your particular model. You must know about the real entities before you make your assumptions. You must understand the significance of the assumptions you make and why they are valid to make. You must know what your techniques are and what the nature of those actually is, including the shortcomings and limitations. You must already know quite a lot about the reality of the situation. (One reason none of the climate models get it right is because they do not incorporate sufficient knowledge about the Earth's atmosphere and what drives its weather systems and so forth.)

This brings us back to the topic of economics. Economics is a study of the actions of volitional individuals and the results. It is a study of how individuals should be allowed to act and WHY they should be allowed to act. It is a study of what occurs when they are restricted from acting and what results. It is a study that relies on key understandings regarding volitional people. That is the issue all the maths in the world cannot obscure or overcome.

An objection to the notion of "mathematical" modelling in economics is that the practioners loudly proclaiming their technical skills have not got a fundamental understanding or knowledge of economics. They assume individual people as mindless particles, described by a few lines of code, some letters and numbers. That is not going to work out as it does not correspond with reality. "I am not a number", as the quote goes. Just as the climate modellers have not got the knowledge (and hence their models are unsuccessful in describing reality), so too with the mathematical "economists" and their castings. See again PC's essay and that interesting table he reproduces.

In the material you have presented a glaring omission exists; the fundamental of economics- individuals and their actions. Auto-trading software and the development thereof is not nearly the same topic.

And now we come to this. You have an 87% confidence in a claim you make, on the basis of someone else's software, that I am Roger Rogerson. Fair enough. But I have 100% knowledge of the fact that Falafulu Fisi is also S - - - e. Understand the difference?

So now, how about admitting that mathematics does not define economics.

LGM

3/29/2008 11:26:00 am  
Anonymous LGM said...

Craig

You wrote, "This is not a blog for debate".

Actually it is. It's just that you've not contributed anything of value or worth to the debate. That's likely because you have nothing. What a worthless toady you are!

LGM

3/29/2008 11:33:00 am  
Anonymous Falafulu Fisi said...

As I said in a previous post that quantitative economics (in which you and PC don't like) where pioneers like Sharpe, Markowitz and others is a sub-discipline of economics (Microeconomics & Econometrics ). What Mises , Reisman and others do, is another sub-discipline, exactly as Physics or Engineering discipline for instance (or other fields). There is Civil, Mechanical, Electrical & Electronic, Chemical, Engineering Science, Software Engineering, and no doubt that there will be new sub-disciplines emerging in the future.

I do understand economics, from the perspective of the sub-disciplines I have quoted above. I don't know other sub-disciplines of economics (Macroeconomics and others), but it doesn't mean that it is something that is totally unknowable or hard for me to grasp if I choose to be self-taught in those areas. It is that I just read what is relevant to what I am doing and that is computational finance/economic models. The point is, that it doesn't mean that Mises know his sub-discipline quite well, and all of a sudden, he is expertise in other economics sub-disciplines and I suspect that this is what you and PC think. Mises & Reisman are always right and everyone else is wrong. Just the same as an engineer who has been trained Civil, can't be expected to know Electrical & Electronic, Mechanical, etc. It doesn't mean that one can't, but you find those experts in their respective sub-disciplines don't try to dismiss others from other sub-disciplines . I've never heard of an Electrical Engineer who had criticized a Chemical Engineer because the Electrical guy thinks that what Chemical Engineers do are all bollocks. Economics is no different. When a guy from one sub-discipline takes a snipe attack on another guy from another sub-discipline (where he has no knowledge about the subject except his uninformed opinions and not based on verifiable facts) with no justification at all because of what he perceived to be impossible, then the attacker needs to be attacked back.

See, the attacker offered no evidence (none) except his/her opinions and the person being attacked could only point to evidence that has been verifiably observed and been published in the literatures. Again don't confuse static models with dynamic models.

I do like Mises & Reisman's philosophies from reading Not PC's blog (even though I have'nt read their books). It itches me, that PC (supported by you) is very dismissive about concepts that doesn't suit him (or doesn't support Mises philosophies), but stealthily used those same concepts that he dismissed to support his views on other unrelated subjects or topics.


LGM said...
But I have 100% knowledge of the fact that Falafulu Fisi is also S - - - e. Understand the difference?

But I also have 100% knowledge of the fact that LGM is G..... R...

Anyway, how is your boat? Have you sold it or you still have it in your possession. I remembered going for a boat ride (with you & your wife) over at the Hauraki Gulf a few years back and the engine stopped. If you didn't manage to restart it, I would have jumped and swam back to the Auckland Harbour, because it looked like that the boat was drifting further & further away. You didn't know what the weather was gong to be like, when you mentioned to your wife, that she should have tuned in to the forecast for that afternoon.

See, you need to listen to weather forecast (even though the models are not perfect) because it gives you a rough guide of what the weather is going to be like if you plan on taking that boat out frequently.

Finally, your book "Schrodingers Cat", that I borrowed was returned to PC a few years back to give it back to you. I think that it is still sitting at the PC's collection at the Castle. I haven't seen you at the Castle since you've been back.

3/29/2008 01:49:00 pm  
Anonymous Anonymous said...

In light of the last few posts I am going to have to retract my corporate offer to you FF.

Looking back on my files I note you forewarded my email to the xtra address that was supplied to me to a pseudonynmous web based address, and replied under that pseudonym. I had told the respective bank that your name was falafulu fisi. 'LGM''s post confirms that it is not.

This is highly unprofessional and we want no part in it.

All the best for the future.

3/30/2008 04:31:00 pm  
Anonymous Falafulu Fisi said...

Ruth,

Thanks for the info you have offered me and I appreciated your effort.

I am known to all the Libertarianzs even though I am not Libertarianz. I was PC's neighbour at the castle years ago, and I happened to enjoy popping downstairs to join their regular BBQs & drinks (used to be last Friday of the month) outside in the garden.

I believe that everyone who is a reader of Not PC blog knows that Falafulu Fisi is a nom de plume and I am surprised that you didn't know that already.

It is good to use one, just the same as you do, because as we move in to more advanced intelligent computing (mathematical modeling) in information retrieval on the internet , I can tell you to be honest that it is quite frightening. Softwares would be able to figure out who is saying what & where, regarding similar concepts when we put those infos on the internet. Google is already using one and one can only estimate that it will get better over times. The conference at Stanford in 2006, Workshop on Algorithms for Modern Massive Data Sets, revealed some new techniques (algorithms) that are able to solve the limitation of current search engines such as Google, Yahoo or Microsoft are using. All current search engines use only 2 metrics, such that Google PageRank collects the outward links and inward links. New techniques can now do more than 2 metrics (unlimited). With this type of capability search engine's indexing accuracy will dramatically improve.

See, the Google placed advertisement in blogs are already deploying such application. No one, from Google is reading Not PC's blog (or million others) to determine its content beforehand, then deciding which items (books) to be placed in a specific blog post. The software scans the blog post try to understand its semantics & concepts then automatically match it to items (or books) that are similar or relevant. The placed ads is not static, it is dynamic. If one observes the items that are advertised on Not PC's blog, mostly they're different items for different posts.

I do also specializes in this area, algorithm development in search engine, text-mining, Sentiment Analysis (SA) and as I said, it is only going to get better & better. And my software will include such capability to analyze financial news story in realtime so that users are made aware of the different possible scenarios that could take place in the market.

In fact Reuters and other commercial vendors are already doing it. There is a system called Warren developed at CMU, that it is quite effective in summrising facts from financial news and presenting them to users. Warren presented infos such as the following:

GOOD - News articles which explicitly show evidence of the company's healthy financial status.
e.g.) … Shares of ABC Company rose 1=2 or 2 percent on the Nasdaq to $24-
15/16. …

GOOD, UNCERTAIN - News articles which refer to predictions of future profitability, and forecasts.
e.g.) … ABC Company predicts fourth-quarter earnings will be high. …

NEUTRAL - News articles which mention financial facts but do not provide good or
bad aspects.
e.g.) … ABC contributes $ 700 million in stock to its pension plan …

BAD, UNCERTAIN - News articles which refer to predictions of future losses, or no profitability.
e.g.) … ABC (Nasdaq: ABC) warned on Tuesday that Fourth-quarter results
could fall short of expectations. …

BAD - News articles which explicitly show evidence of the company's bad financial status.
e.g.) … Shares of ABC (ABC: down $0.54 to $49.37) fell in early New York
trading. …


Here is another interesting article on using text-mining in the financial market analysis:

Using News articles to predict Stock Price Movements

I do test my text mining algorithm on the 20 Newsgroups, and recently, I've just set my crawler to scan the Not PC blog and cluster it by topics, by post, by user, etc,... and this was how I noted the similarity between the LGM, Abdul, Roger Rogerson, because they all lumped up together. I will continue to test its accuracy on Not PC's blog and I will include other financial news sites for the crawler to index .

If you're interested in such software capability for your industry & Forex trading, then take a look at this commercial vendor, perhaps you might want to subscribe to their service: Forex News and Sentiment Analysis | Economics News Alert

Forget about Falafulu Fisi's as a name, you want best performing algorithm that is going to give good return then your organisation should look no further than my software. It is likely that you won't be dealings with me, I believe that there will be a sales person or whoever that is going to promote it on behalf of my startup, since it would be better for me to do what I am good at, and that is researching the literatures such as my favorites (RePEC - Research Papers in Economics) to find best performing algorithms & techniques to adopt and leave the admin or sales site of the business to someone else. So, in that way it is purely professional relationship because your organisation will not be dealing directly with me.

Here is a general hint about its capability:

A rocket scientist from NASA or an Electrical Engineer from Motorolla might confuse my software as design tool that they could use in their work.

Yep, Signal Processing algorithms, feedback control theory algorithms (yes, feedback mechanisms do exist in the market) are included (not one, but many different types).

I have started informal inquiries to venture capitalists such as Jenny Morel and others, so that I can present my ideas in the prototype that I am writing at the moment. I am confident that they will come in and take a chunk of my startup and become shareholders. There is no guarantee that Jenny will come in, but if it is not her, then it may be someone else (there are a few of those in NZ). All startups that Jenny's company involve in, she is also a director. This is what is likely to be the case. If an external investor becomes involve in what I am doing, then corporate dealings would be done with that person and not with Falafulu. And as this local millionare business man who is a huge fan of Falafulu Fisi where I have met him a few times has advised startups in his Herald article to take the money when seeking venture funds even if it is a big chunk of your business that they want to invest in. This is true, it is better to own a small piece of a big pie than a big piece of a small one.

I am well connected to University researchers (Mathematics, Physics, Statistics, Artificial Intelligence) both local and from offshore, and for a high-end product that target international market of the likes that I am developing, one needs to keep innovating and this can only be achieved by collaborating or funding Unversity researchers. The day when one stops innovating, is the day, that the competitors perhaps overtake you.

The man from the island (me) wants to get involve in algorithm wars as stated by Andrew Lo, director of the Massachusetts Institute of Technology’s Laboratory for Financial Engineering in this article: Artificial intelligence applied heavily to picking stocks

"Now it’s an arms race,Everyone is building more sophisticated algorithms, and the more competition exists, the smaller the profits"

Falafulu Fisi is confident that his product can compete in international market with billionare fund managers as Kenneth Griffin and James Simon who are developing & using similar systems. My only limitation to take the product internationally is the funding. If funding is available, then I am able to recruit more programmers.

This software will give you good return Ruth and forget about internet pseudonyms because everyone has one.

3/31/2008 01:38:00 am  
Anonymous Falafulu Fisi said...

Paul Walker, I am interested to hear your view on the subject definition of economics. LGM & PC thinks that Econometric & Micro-economics are not economics at all. None of us has been formally trained in economics like yourself and it will be interesting to see some interpretation of the term from you.

3/31/2008 02:38:00 am  
Blogger Paul Walker said...

I'm not sure there really is a good definition of economics. See the examples below.

Most of economics can be summarized in four words: People respond to incentives. The rest is commentary
Steven E. Landsburg

My view of economics is that it is the study of the economic system, that it studies how people earn their incomes and how they spend them, and that it studies the institutions of the economic system, that is to say, banks,
markets, and firms, all influenced ... by the legal system.
Ronald Coase

... Economics is a study of mankind in the ordinary business of life; ...
Alfred Marshall

Economics is the science which studies human behavior as a relationship between scarce means which have alternative uses
Lionel C. Robbins

Economics is what economists do
Jacob Viner

The Viner one is the most honest! But if you look at any of them then microeconomics is covered by them. Econometrics (that is the use of stats in econ) is a tool that economists find useful so we teach and use it. I guess its not part of econ as such but it is used in both macro and micro to help understand whatever we are looking at.

3/31/2008 03:54:00 pm  
Anonymous Anonymous said...

as a non economist, my definition would include the desire to explain human (primarily commercial) behaviour and the impact of that behaviour and systems on economies.

I think the point LGM was making at length is that economics is qualitative, trading models quantitative unless applied by a qualitative mind.

Insider

3/31/2008 05:43:00 pm  

Post a Comment

Respond with a polite and intelligent comment. (Both will be applauded.)

Say what you mean, and mean what you say. (Do others the courtesy of being honest.)

Please put a name to your comments. (If you're prepared to give voice, then back it up with a name.)

And don't troll. Please. (Contemplate doing something more productive with your time, and ours.)

<< Home