I Know First Presentation (February 2013)

How Can We Predict the Financial Markets?

Investing in the time of uncertainty

By Dr. Lipa Roitman

IKnowFirst.com

February 2013

2Slide

IKnowFirst.com

I Know First - Israeli Forecasting Technologies is an Israeli start up company. Our main product is a financial market forecasting algorithm that predicts daily more than 200 markets: Stocks, world indices, Currencies & Commodities.

The company-I Know first-Introduction

• The system is a predictive model based on artificial Intelligence (AI) and Machine Learning (ML), and incorporating elements of Artificial Neural Networks and Genetic Algorithms, built with insights of Chaos Theory and self-similarity, the Fractals.

• I Know First tracks and predicts the flow of money from one market or investment channel to another

I Know First Predicts 200

investment channels daily

Tracks the flow of money Artificial

Intelligence (AI) and

Machine Learning (ML),

Artificial Neural

Networks

Genetic Algorithms

3Slide

IKnowFirst.com

Lipa Roitman PhD

Entrepreneur, algorithmic trading system developer 20 years in the AI (artificial intelligence) and Machine

Learning FieldsConsultant for startup companiesR&D Chemist with record in computer modeling of process,

new product and process development

http://iknowfirst.com/Stock-forecast-articles

l k rc ic he e

4Slide

IKnowFirst.com

"it is exceedingly difficult to make predictions, particularly about the future"

Niels Bohr.

5Slide

IKnowFirst.com

Old Fallacies About Markets

Efficient: “Markets are efficient and unpredictable: today’s information is already reflected in price. No one stock is a better buy then the other”.

Random walk: “The patterns of stock market prices are purely random. Markets are unpredictable because they are random. Playing markets is a gamble“

Markets are neither totally efficient, nor totally random. They are complex and chaotic.

6Slide

IKnowFirst.com

Trading the Stock Market

Fact: big trading houses like Goldman Sachs, Morgan Stanley consistently make profit trading stocks.

Some fail spectacularly, like Lehman Brothers, when they make big bets that go wrong, and we read of them in the paper.

7Slide

IKnowFirst.com

High-Frequency Trading

High-frequency trading HFT: quick arbitrage (in milliseconds)

Computer algorithms place orders based on information that is received electronically, before human traders can react.

Example: arbitrage from the bid-ask spread →Supposed to provide liquidity, but sometimes errors in the

algorithms cause huge distortions in prices. →Example: May 6, 2010 Flash Crash→January 23, 2013 AAPL plunge

8Slide

IKnowFirst.com

Risk Management: Fat Tailed Distribution

AAPL Flash Dump

High-frequency trading algorithms.

How else could 800,000 shares worth nearly $300 million be sold in 17-second intervals?

9Slide

IKnowFirst.com

High-Frequency TradingProfits from high-

frequency trading HFT in American stocks have peaked in 2009, and going down since: → $1.25 billion in 2012, down

35 percent from 2011 and 74 percent lower than $4.9 billion in 2009

→ The percentage of stock trades handled by firms that specialize in H.F.T. fell to about 51 percent in 2012 from 60 percent in 2009.

10Slide

IKnowFirst.com

High-Frequency Trading

Why HFT is slowing down?Lower trade volume retail investor is not back yet (begins to come back recently)Mainly institutional investors are participatingTechnological costs: financial resources to compete in this

field are now enormous (high latency, proximity to exchanges, investment in hardware and software)

High competition-low profit

11Slide

IKnowFirst.com

Trading the Stock Market

Stock market is not a walk in the park, but it’s not a random walk either.

Everyone can get lucky with stock market sometimes.

To do it consistently requires knowledge of the market, the market direction and risk management strategy.

12Slide

IKnowFirst.com

Chaos is the Result of Complexity

Why do stock prices move?→news stream constantly injects new information. →mixed messages →different market players psychology creates patterns.

13Slide

IKnowFirst.com

Chaos is the Result of Complexity

Objective factors→Different valuation

modelsFundamental

valuation Price momentum, etc.

→Different time horizons short time horizon vs.

the longer view.

Reasons for chaotic behavior

14Slide

IKnowFirst.com

Psychology of Trading

Human factor: It is difficult if not impossible to make a 100% rational decision during uncertainty.

Prospect theory: (Kahneman and Tversky, 1979): Losses have more emotional impact than an equivalent amount of gains. Risk aversion.

Reasons for chaotic behavior

15Slide

IKnowFirst.com

Psychology of Trading

Which stock do you check first when you analyze your portfolio performance daily ?→Last one?→First one?→The largest investment?

16Slide

IKnowFirst.com

Psychology of TradingLatest news biasOverreaction: can’t quantifyAnchoring: buying on dipsTrend lovers: herd mentality

Why people lose money

17Slide

IKnowFirst.com

Chaos is the Result of Complexity

18Slide

IKnowFirst.com

There is an Order in the Chaos.

Basic Money Law: Money is looking for:

highest returns with lowest risk

It constantly flows from one market to another. → However the way the flow

occurs is rarely smooth, but marked with periods of turbulence.

19Slide

IKnowFirst.com

What is Chaos?

Chaos is a complex non-linear evolving system that is sensitive to initial conditions. It has “memory”.

There are times when the path is well defined and predictable

There are points in time (instability regime) where a minor perturbation can switch the future path between two opposite directions.

Chaos can appear as randomness, but it is not. This is the way we could tell them apart:

20Slide

IKnowFirst.com

Random behavior can't be learned. It is random. →Past patterns don't repeat.

Chaotic systems have memory. What happened in the past affects the future.→ Can be learned and predicted to some extent (quasi-

deterministic chaos).

What is Chaos?

21Slide

IKnowFirst.com

Chaos is a law of nature. It appears in complex dynamic systems where each element affects the others.

Astronomy: many-bodies system.

Weather

What is Chaos?

22Slide

IKnowFirst.com

Stock Market and Quantum Mechanics

Double slit experiment Particle-wave duality

23Slide

IKnowFirst.com

Stock Market and Quantum Mechanics

Five years chart:

Patterns can be seen.

The prices jump from 1 level to another

24Slide

IKnowFirst.com

Stock Market and Quantum Mechanics

One month chart:

No patterns apparent

25Slide

IKnowFirst.com

Stock Market and Quantum Mechanics

Three days chartGranularity: Single

transactionsNo patterns apparent

Stock market exhibits deterministic chaos, making the short-term movements of prices

extremely impossible to predict.

26Slide

IKnowFirst.com

Stock Market and Quantum Mechanics

How is the Stock Market like Quantum Mechanics? A single photon (a particle of light) or an electron

behave like a particle (quantum), an assembly of them behaves like a wave.

Discrete quantum levels of electrons in an atom.Market is an assembly of individual transactions

(quanta). Together they show similar patterns.

• Both are Probabilistic• Discrete levels• Granularity: unpredictable on microscopic

scale, but predictable on large scale.

27Slide

IKnowFirst.com

Chaos Example

What is the pattern?

28Slide

IKnowFirst.com

Financial Bubbles

Big Bubble

iPhone 5 release

29Slide

IKnowFirst.com

Chaos Theory and Financial Bubbles

Feedback mechanisms: Positive feedback amplifies trends (bubble

formation):→Crowd mentality, chain reaction: chasing one stock, →Go with the trend!

30Slide

IKnowFirst.com

Chaos Theory and Financial Bubbles

Feedback mechanisms: Negative feedback limits the trend

→bubble bursting, → range-bound trading → “price got too high: sell!, → “price got too low: buy!”

31Slide

IKnowFirst.com

Chaos Example

Negative Feedback

Positive Feedback

32Slide

IKnowFirst.com

Apple Inc. AAPL Bubble Crash

Financial Bubble Detection

33Slide

IKnowFirst.com

Chaos Theory and Financial Bubbles

Randomness is also part of the market: Occurs when the market is indecisive

→Low volume→ Increased volatility→Randomness is stronger near turning points

Hidden variables: Albert Einstein VS. Niels Bohr: "I am convinced God does not play dice"

→A trader placed big order at the low liquidity time→Computer error→Market manipulation (HFT program trading, etc)

Increased volatility (randomness) is a warning: stay away from the market or adjust your tactics

34Slide

IKnowFirst.com

What are financial bubbles?

Stock market is a bubble machine on all time scale levels. Big bubbles can last years.

A bubble is a very basic part of the market and can’t be eliminated, but if it is recognized it could be exploited!

Part of the price discovery process in the presence of uncertainty.

To get to the “fair” market price the price has to “overshoot” in both directions.

The market constantly behaves like a drunk driver

35Slide

IKnowFirst.com

The Key to the Market

Markets are chaotic, they alternate between three regimes: positive feedback, negative feedback, and randomness.

The three regimes could be present simultaneously at different time scales

The one who can recognize these regimes has the key to the market.

The “buy” or “sell” decision depends on what regime you think the market is at now, and at what time scale!

36Slide

IKnowFirst.com

The Key to the Market

“For every complex problem there is an answer that is clear, simple and wrong.”→– H. L. Mencken

37Slide

IKnowFirst.com

The Key to the Market

Market fallacies:“The trend is your friend —

(until that nasty bend at the end)”.“Buy low, sell high”.

The “buy” or “sell” decision depends on what regime you think the market is at now, and at what time scale!

38Slide

IKnowFirst.com

The Key to the Market"I took economics courses in Harvard College for four years, and everything I was taught was wrong."

Franklin D. Roosevelt (1882 –1945)

"The established theory has collapsed but we haven't actually got a proper understanding of how financial markets operate”.

George Soros

World Economic Forum in Davos 2013

39Slide

IKnowFirst.com

I Know First Algorithmic System

I Know First algorithmic system was developed to discover the laws of the market that could show which way the market is going.

40Slide

IKnowFirst.com

I Know First - Israeli Forecasting Technologies is an Israeli start up company. Our main product is a financial market Forecasting system that predicts daily more than 200 markets: Stocks, Currencies & Commodities.

The company-I Know first-Introduction

• The system is a predictive model based on artificial Intelligence (AI) and Machine Learning (ML), and incorporating elements of Artificial Neural Networks and Genetic Algorithms, built with insights of Chaos Theory and self-similarity, the Fractals.

• I Know First tracks and predicts the flow of money from one market or investment channel to another

I Know First Predicts 200

investment channels daily

Tracks the flow of money Artificial

Intelligence (AI) and

Machine Learning (ML),

Artificial Neural

Networks

Genetic Algorithms

41Slide

IKnowFirst.com

I Know First Algorithmic System I Know First Algorithmic System is based on the realization

that a stock value is a function of many factors which interact in a non-linear way and affect the future trajectory of the stock creating waves in prices.

Being completely empirical, the I Know First self learning algorithms analyze the inputs and rank them according to their significance in predicting the target stock price.

Then they create multiple models, and test them automatically on the historical data.

The robustness of the model is measured by how it performs in different market circumstances.

The best predicting models are kept and the rest are rejected. Such refinement has continued daily as the new market data is added to the historical pool.

42Slide

IKnowFirst.com

I Know First Algorithmic System

Adaptable and balanced algorithm: →Empirical

does not depend on any human assumptionsself-learning

→ learns new patterns daily, →adapts to new reality, →but still follows the general historical rules.

43Slide

IKnowFirst.com

I Know First Algorithmic System

The stocks, indexes, commodities and currencies represent most of the liquid forms of investment

“What if” scenarios: →Effect of raising or lowering interest rates on the

markets and on real estate prices. →Effect of currency exchange rates→Effect of higher or lower oil prices →Effect of price of oil on oil company stocks→More….

A Model of the World Economy!

44Slide

IKnowFirst.com

I Know First Algorithmic System

Many inputs from different sources go into each algorithmic forecast.

Up to 15 years of data go into each model. Powerful computers process and learn the data, and create

forecast.

45Slide

IKnowFirst.com

I Know First System

running

Cycle

The Product- I Know First System Basic Principle

Daily stock Data

Get the daily market update, and add it to the 15-years database

Run a learning & prediction cycle with new combined data (Takes about 8 hours per cycle, it runs non-stop around the clock).

15 years stock

database

Reporting Module-Software as a service (SaaS)

The daily prediction for each stock/ currency/ commodity is produced for the following periods :3 days, 7 days, 14 days, 30 days, 90 days & 365 days

AgenTeamIQSHIP

Learning & Prediction

Cycle

Learning & Prediction

Cycle

Generate Results”

procedure

Generate Results”

procedurePredictionsPredictions

Database

3 Days prediction

7 Days prediction

14 Days prediction

30 Days prediction

90 Days prediction

365 Days prediction

46Slide

IKnowFirst.com

I Know First Algorithmic System

FeatureTech. analysis

I Know First

Algorithm Self learning Adaptable Learns new patterns daily Looks at many different stocks,

indexes, etc Signals at different time horizons Quantitative Predictability indicator Artificial Intelligence Neural Networks Genetic algorithms

No

No

No

No

No

No

No

No

No

No

No

Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

The system is self learning, which sets it apart from the technical analysis.

47Slide

IKnowFirst.com

I Know First Algorithmic System-Performance

48Slide

IKnowFirst.com

I Know First Algorithmic System-Performance

49Slide

IKnowFirst.com

The Product- I Know First System

Basic PrincipleGTI CHKP X F VNQ

16.69 14.58 13.90 13.42 12.620.622 0.528 0.676 0.655 0.679GT KEY Platinum ^HSI ORBK

11.82 11.80 11.09 9.75 9.450.709 0.209 0.591 0.707 0.685AMD BA IFN WAG DD

9.20 8.77 7.87 7.50 6.760.363 0.553 0.575 0.484 0.576DIS CSCO VUG UN POT

5.68 5.56 5.34 5.31 4.760.592 0.642 0.694 0.585 0.551^VIX ZRAN CSX VPL VXF

4.71 4.55 4.41 4.26 4.010.526 0.469 0.643 0.66 0.629WFR ^TA100 AAUKY VDE BAC

3.86 3.86 3.77 3.73 3.680.557 0.731 0.634 0.618 0.463VTV VGK S&P500 INTC HAS

3.64 3.61 3.38 3.22 2.900.605 0.696 0.628 0.595 0.439DOX WFC RTLX LMT ADM

2.81 2.79 2.36 2.02 1.700.448 0.269 0.456 0.39 0.265EWG TEVA EWC NICE WMT

1.54 1.25 0.68 0.56 0.510.59 0.487 0.373 0.411 0.125EWA JNJ KO VWO T

0.34 0.26 0.24 0.21 0.190.741 0.406 0.491 0.707 0.028K BMY AIP NIS/EUR NIS/GBP

0.09 0.07 0.03 -0.09 -0.090.402 0.048 0.614 0.272 0.476TRP NIS/$US CVS FCX CAT

-0.18 -0.22 -0.29 -0.29 -0.350.536 0.485 0.381 0.29 0.583ENB NSC JY/$US ESLT TKF

-0.38 -0.51 -0.77 -0.78 -2.210.625 0.479 0.203 0.594 0.418OTEX AAPL Crude Oil IRL ISRL

-2.32 -3.19 -5.09 -7.09 -11.640.459 0.644 0.668 0.47 0.398WDC DOW

-12.53 -43.430.692 0.602

Negative Negative Positive Postive++ ++

SignalSymbol

Predictability

VWO0.21

0.707

• For each stock/ currency/ commodity the following data is calculated by the system:

•Predictability - The "strength" of the prediction

•Signal- the movement direction (increase/decrease)

•The daily prediction is produced for the following periods :3 days, 7 days, 14 days, 30 days, 90 days & 365 days

50Slide

IKnowFirst.com

The Product- I Know First System

Basic PrincipleGTI CHKP X F VNQ

16.69 14.58 13.90 13.42 12.620.622 0.528 0.676 0.655 0.679GT KEY Platinum ^HSI ORBK

11.82 11.80 11.09 9.75 9.450.709 0.209 0.591 0.707 0.685AMD BA IFN WAG DD

9.20 8.77 7.87 7.50 6.760.363 0.553 0.575 0.484 0.576DIS CSCO VUG UN POT

5.68 5.56 5.34 5.31 4.760.592 0.642 0.694 0.585 0.551^VIX ZRAN CSX VPL VXF

4.71 4.55 4.41 4.26 4.010.526 0.469 0.643 0.66 0.629WFR ^TA100 AAUKY VDE BAC

3.86 3.86 3.77 3.73 3.680.557 0.731 0.634 0.618 0.463VTV VGK S&P500 INTC HAS

3.64 3.61 3.38 3.22 2.900.605 0.696 0.628 0.595 0.439DOX WFC RTLX LMT ADM

2.81 2.79 2.36 2.02 1.700.448 0.269 0.456 0.39 0.265EWG TEVA EWC NICE WMT

1.54 1.25 0.68 0.56 0.510.59 0.487 0.373 0.411 0.125EWA JNJ KO VWO T

0.34 0.26 0.24 0.21 0.190.741 0.406 0.491 0.707 0.028K BMY AIP NIS/EUR NIS/GBP

0.09 0.07 0.03 -0.09 -0.090.402 0.048 0.614 0.272 0.476TRP NIS/$US CVS FCX CAT

-0.18 -0.22 -0.29 -0.29 -0.350.536 0.485 0.381 0.29 0.583ENB NSC JY/$US ESLT TKF

-0.38 -0.51 -0.77 -0.78 -2.210.625 0.479 0.203 0.594 0.418OTEX AAPL Crude Oil IRL ISRL

-2.32 -3.19 -5.09 -7.09 -11.640.459 0.644 0.668 0.47 0.398WDC DOW

-12.53 -43.430.692 0.602

Negative Negative Positive Postive++ ++

SignalSymbol

Predictability

VWO0.21

0.707

Two components to the stock action

→ Stock-specific action.→ Stock action related to

general market action. A stock is a part of the

market and responds to the general market news.

The forecast table shows how the stock forecast is positioned relatively to other stocks forecast.

51Slide

IKnowFirst.com

The Product- I Know First System

Basic PrincipleGTI CHKP X F VNQ

16.69 14.58 13.90 13.42 12.620.622 0.528 0.676 0.655 0.679GT KEY Platinum ^HSI ORBK

11.82 11.80 11.09 9.75 9.450.709 0.209 0.591 0.707 0.685AMD BA IFN WAG DD

9.20 8.77 7.87 7.50 6.760.363 0.553 0.575 0.484 0.576DIS CSCO VUG UN POT

5.68 5.56 5.34 5.31 4.760.592 0.642 0.694 0.585 0.551^VIX ZRAN CSX VPL VXF

4.71 4.55 4.41 4.26 4.010.526 0.469 0.643 0.66 0.629WFR ^TA100 AAUKY VDE BAC

3.86 3.86 3.77 3.73 3.680.557 0.731 0.634 0.618 0.463VTV VGK S&P500 INTC HAS

3.64 3.61 3.38 3.22 2.900.605 0.696 0.628 0.595 0.439DOX WFC RTLX LMT ADM

2.81 2.79 2.36 2.02 1.700.448 0.269 0.456 0.39 0.265EWG TEVA EWC NICE WMT

1.54 1.25 0.68 0.56 0.510.59 0.487 0.373 0.411 0.125EWA JNJ KO VWO T

0.34 0.26 0.24 0.21 0.190.741 0.406 0.491 0.707 0.028K BMY AIP NIS/EUR NIS/GBP

0.09 0.07 0.03 -0.09 -0.090.402 0.048 0.614 0.272 0.476TRP NIS/$US CVS FCX CAT

-0.18 -0.22 -0.29 -0.29 -0.350.536 0.485 0.381 0.29 0.583ENB NSC JY/$US ESLT TKF

-0.38 -0.51 -0.77 -0.78 -2.210.625 0.479 0.203 0.594 0.418OTEX AAPL Crude Oil IRL ISRL

-2.32 -3.19 -5.09 -7.09 -11.640.459 0.644 0.668 0.47 0.398WDC DOW

-12.53 -43.430.692 0.602

Negative Negative Positive Postive++ ++

SignalSymbol

Predictability

VWO0.21

0.707

Customized forecastChoosing stocks from

particular industry into the table composition

52Slide

IKnowFirst.com

The Product- I Know First System

Short term predictions: 3 days, 7 days, 14 days

Example

53Slide

IKnowFirst.com

The Product- I Know First System

Long term predictions: 30 days, 90 days & 365 days

54Slide

IKnowFirst.com

I Know First Algorithmic System

More uses for algorithms:Forecasting demand for products and servicesModeling complex chemical processesGlobal climate trendsAgricultural forecasting: crops, water demand More….

Send us requests for your forecasting needs

iknowfirst@iknowfirst.com

55Slide

IKnowFirst.com

What are the stocks for 2013?

56Slide

IKnowFirst.com

December performance

57Slide

IKnowFirst.com

December 2012 performance-aggressive

58Slide

IKnowFirst.com

Recent performance

59Slide

IKnowFirst.com

October 2012 performance

60Slide

IKnowFirst.com

Google Stock Forecast: Case Study http://iknowfirst.com/Google-stock-forecast-case-study

61Slide

IKnowFirst.com

Google Stock Forecast: Case Study

62Slide

IKnowFirst.com

Google Stock Forecast: Case Study

63Slide

IKnowFirst.com

Signal Analysis of a Group of Stocks

The following chart shows a combined signal of all stocks in the system, calculated for each of the six time ranges.

http://iknowfirst.com/S-P-500-May-12-2013

Send us requests for your forecasting needs

iknowfirst@iknowfirst.com

64Slide

IKnowFirst.com

Signal analysis of a group of stocks

65Slide

IKnowFirst.com

Signal analysis of a group of stocks

The stocks and indices in the I Know First system are a good representation of a broader market.

The forecast for the plurality of stocks in the I Know First system can serve as a proxy for the S&P500 forecast.

66Slide

IKnowFirst.com

Predictability: Measuring Chaos

Predictability is the indicator that tells predictable chaos from randomness.

Just like the market rises and falls in waves, so does the predictability. And the waves are not synchronous.

The focus shifts between gold, stocks, oil, bonds. Some become more predictable, while the others retreat to randomness.

67Slide

IKnowFirst.com

Predictability: Measuring Chaos

By monitoring predictability one can get advance warning that the market paradigm change is in progress.

68Slide

IKnowFirst.com

Which stocks can be predicted Most of the major stocks in the S&P 500 index are

predictable to some extent.

Start-Ups are Unpredictable→ Investor hopes: ‘This is going to be the new Google, the new

Facebook.’ → Some start-ups make it, some don’t — nobody knows in

advance

But the main reason our algorithms can’t forecast start-ups: no history. No way to predict future moves.

69Slide

IKnowFirst.com

70Slide

IKnowFirst.com

71Slide

IKnowFirst.com

72Slide

IKnowFirst.com

73Slide

IKnowFirst.com

74Slide

IKnowFirst.com

75Slide

IKnowFirst.com

76Slide

IKnowFirst.com

77Slide

IKnowFirst.com

78Slide

IKnowFirst.com

79Slide

IKnowFirst.com

80Slide

IKnowFirst.com

Risk Management

So, is it a trading system that could make money consistently?

Most of the time yes, with the right risk management strategy and a bit of luck!

Luck factorWe don’t know all risk factors.

81Slide

IKnowFirst.com

Risk Management

● There are:● “known knowns; there are things we know that we

know.● known unknowns; that is to say there are things

that, we now know we don't know.● But there are also unknown unknowns – there are

things we do not know we don't know”. ● —Donald Rumsfeld● United States Secretary of Defense

82Slide

IKnowFirst.com

Risk Management: Normal DistributionNormal

distribution applies to many random events,

but it is not a rule in the markets, but rather an exception.

83Slide

IKnowFirst.com

Risk Management: Fat Tailed DistributionFat Tailed

distribution is very common in the markets.

Large swings (3 to 6 standard deviations from the mean)

Far more frequent than the normal distribution

Power Law

84Slide

IKnowFirst.com

The Danger of Fat Tails

The uncertainty about price distribution makes “rational” decision making impossible.

One could be right about the market direction, but lose money or miss an opportunity.

85Slide

IKnowFirst.com

The Danger of Fat Tails

Risk management:allocation of capital traditional allocation model) allocation of risk.

86Slide

IKnowFirst.com

Risk Management

87Slide

IKnowFirst.com

Strategies for Success

Watch the signals daily, but act only on strong ones. To minimize risk stay out of the market until you see a

great opportunity: a strong signal, extreme price.When predictability is high, invest on strong signals. When predictability goes down, expect a storm. When the signal disappears or weakens, reduce your

exposure.For a stable portfolio invest in non-correlated securities.

→ Caveat: during times of global financial crisis all assets become positively correlated, because they all move (down) together.

88Slide

IKnowFirst.com

Risk Management: Fat Tailed Distribution

What happened to Apple (AAPL) in the last minute of trading Friday, January 23, 2013

89Slide

IKnowFirst.com

Risk Management: Fat Tailed DistributionAAPL Flash

DumpHigh-

frequency trading algorithms. How else could 800,000 shares worth nearly $300 million be sold in 17-second intervals?

:

90Slide

IKnowFirst.com

What Causes Fat Tails

● Extreme risks of "high consequence", but of low probability. The risks of

● terrorist attack, major earthquakes, accidents● hurricanes, a volcanic-ash cloud grounding all

flights for a continent, ● HFT trading algorithms

the frequency and impact of totally unexpected events is generally underestimated

91Slide

IKnowFirst.com

Self-Similarity: Fractals.

Chaotic systems and fractals. Fractals are objects which are "self-similar" in the

sense that the individual parts are related to the whole. Mandelbrot. →The detail looks just about the same as the whole.

Market patterns are the same on all time scales, except the shortest times (quanta).

92Slide

IKnowFirst.com

Fractals and the Power Laws

The main attribute of power laws that makes them interesting is their scale invariance. Given a relation f(x) = ax^k, scaling the argument x by a constant factor c causes only a proportionate scaling of the function itself. That is,→ f(c x) = a(c x)^k = c^{k}f(x) is proportional to f(x).

That is, scaling by a constant c simply multiplies the original power-law relation by the constant c^k. Thus, it follows that all power laws with a particular scaling exponent are equivalent up to constant factors, since each is simply a scaled version of the others. This behavior is what produces the linear relationship when logarithms are taken of both f(x) and x, and

the straight-line on the log-log plot is often called the signature of a power law.