“I suppose,” said Jo.
The following Monday, Jo arrived at the office having enjoyed a normal quota of sleep and looking forward to carrying out some proactive market analysis based on a trading idea she had. The analysis required sifting through large volumes of data, a task which Jo knew could be facilitated by using some software that the traders employed.
Jo called Adam to see if he could come and install the program on her computer. Half an hour later Adam stood in the doorway, hesitating in mock apprehension.
“Now, you haven’t called me to do anything illicit again have you? Hack Bray’s personal files, read Wright’s e-mail maybe?”
“Shh,” Jo ushered him in, “I don’t know what you’re talking about Adam. I would never ask to see files I wasn’t authorized to see, and you would never give me the access.”
“Absolutely. That’s the official story and we’re sticking to it.”
“Seriously Adam. Let’s not mention it again.”
On a day to day basis, Jo had managed to block the incident from her mind when at work. Adam bringing up the subject, even in good humor, had irked Jo. She wasn’t sure whether her disquiet was caused by being reminded of what she had found, or by the fact that she hadn’t acted upon it.
Adam went to work and as the awkwardness passed, their conversation resumed. Jo warmed to Adam even more as she discovered that they had many interests in common, including a passion for scuba diving. Jo’s diving experiences had been confined to the Mediterranean and she was awed by tales of Adam’s escapades of diving with sharks at the Blue Hole in Belize and a near encounter with a sting ray on Australia’s Great Barrier Reef.
Adam finished installing the software and was leaving when Jo called him back.
“Adam, even though we’re not talking about ‘the incident’, I do still owe you a drink.”
“Yes, I think maybe you do. In fact, you may be advised to pay for a few extra drinks on account. You never know when you’ll be facing your next predicament and in dire need of IT support.”
“I don’t know that I’d ever admit to being in dire need of anything to do with IT, but I’m sure the price of a few beers would be a good insurance policy.”
“Let’s do Friday then. I’ll meet you downstairs at six-thirty, if you can sneak out for an early night. I’ll pick the venue.”
“Friday it is.”
Jo smiled to herself once Adam had left, wondering if she had finally found something close to a friendship at Butterfly.
CHAPTER FIFTEEN
That Thursday afternoon, Charles Bray invited Jo into his office. He announced that the time had now come for her to start the study of something truly interesting. Bray ceremoniously took down a thick blue leather bound volume from a shelf behind his desk and placed it in front of Jo. He motioned for her to pick up the tome and open it.
Jo handled the volume somewhat gingerly, aware of the hawk like vigilance with which Bray was observing her every move. She opened the first page with a reverence she sensed was expected.
In large black font near the head of the page was the title, “The Inefficient Market Hypothesis.” In the same font, but rendered more prominent by its central placement was written, “By Professor Charles T. Bray.” Toward the bottom was the longer, but still somewhat opaque explanation, “The Application of Neural Network Analysis to Anomalous Behavior in Financial Markets.”
When Jo looked up, the expectancy on Bray’s face had not abated. He motioned for her to turn to the next page, which contained a single quote. “Does the flap of a butterfly’s wings in Brazil set off a tornado in Texas?” The author of the question was identified as Edward Lorenz, famous for his seminal work on Chaos Theory. At least, pondered Jo, she appeared to have solved one of the mysteries of her secretive employer; the origin of its name.
Jo continued to turn the pages. She skimmed through some introductory paragraphs which provided the background to the proposed theoretical framework. Words quickly ceded way, however, to advanced mathematical formulas which were given no explanation. The target audience was obviously assumed to have graduate level knowledge of statistics, probability and certain aspects of quantum theory as they relate to the prediction of future outcomes.
Jo experienced a gratifying mix of curiosity and eagerness as she realized she was facing a formidable academic challenge and that she was one of the few capable of taking it on.
“Well?” Bray eventually asked. “Do you realize what you are looking at?”
“An accelerated doctorate course in financial market analysis?” suggested Jo.
“Your aim is too low. Believe me, if this were published, the word Nobel would be more apt than thesis. Care to guess again?”
“Well it looks like the theory behind some kind of financial model.”
“We all know looks can be deceiving, don’t we? Just like the markets. I’m slightly disappointed by all your literal thinking. It’s not a course, or a theory, or a model. That is what they teach those useless MBA clones who are incapable of figuring anything out for themselves. No, you’re holding something much more valuable. It’s a key, a gateway if you will.”
“To what exactly?” ventured Jo.
“To the future, of course. The underlying theory has a vast number of potential applications to many fields. We live in an information driven world, where the distribution of power is directly related to the distribution of information. In the world of finance there is the tacit belief that power comes from wealth. The big players have the ability to influence trading. But where most people fail is in completing this logic. What is the ultimate source of this wealth? How did these players accumulate it in the first place?” Bray looked at Jo for an answer. By now she was detecting a theme.
“Through information.” Jo offered obligingly.
“Exactly!” said Bray, pleased with the success of his tutorial. “Well there’s one more step. Information is obviously a vague term. Everyone can claim to have access to information, even access to all the same information. The key, the magic, the genius, if you will, is in connecting this generic information and creating more information that others either don’t have or cannot see. And you, my dear, are holding that genius in your hand.”
Jo unconsciously wiped her hand on her skirt, as if not wishing to contaminate this grail of power.
Bray checked his watch and determined that it was time to descend from his philosophical tower. “Anyway, let’s get started shall we? I am to be your personal tutor for the next few months, before you put any of this into practice.”
Unlike the time when she was confronted with her move to the Quantitative Group, Jo’s response was quick and appropriate.
“What a privilege. I am extremely grateful that you would spend that time with me.”
“Well, I am meticulous in ensuring that the theories are fully understood and that no third party interpretations leak into the source material.”
Jo thought that Bray also wanted to ensure that all the accolades flowed to the source too.
“Let’s start at the beginning then.”
Jo turned to the first page of text and equations, but Bray came over and turned back to the first title page. “Tell me, what do you know of the inefficient market hypothesis?”
Jo looked nonplussed. She had hoped to delve directly into the math.
“Well how about its opposite then?”
“The efficient market hypothesis?” asked Jo. Bray nodded, if a little over patiently. Jo began to regurgitate sections of textbook theory recently learned. “Well, its most well-known proponent was Fama. He postulated that financial markets are efficient in so far as all the information concerning a commodity, stock or financial instrument, including expectations of future events, is already reflected in its price. Prices are always therefore fair, or efficient. The only reason a price moves is if something unexpected happens, which cannot be predicted from current information.”
“And what does this mean for inve
stors or fund managers?”
“Well the implication is that it is impossible to consistently outperform the market through active selection of investments. Because all the information to which you have access has already been factored into the price of the instruments you want to invest in, all you are really doing is taking a gamble that some unexpected event will change the price in the future. And in the long run, you cannot gamble and consistently beat the market. Sooner or later, your luck will end.”
“Yes, yes, all very commendable. Pass the Risk Analysis section of your exam and collect a job in a big investment bank as you move to the Derivatives section of the paper. But you’re missing the most important part of the efficient market hypothesis.”
Again, Jo was flummoxed. Bray helped her out, vehemently.
“It’s complete and utter rubbish spouted by a bunch of high minded academic recluses who form a fatuous intellectual circle where everyone agrees with the dinosaur on their left until they get back to the beginning and suddenly the theory is irrefutable.”
Bray marched to a refrigerator and took a few gulps of water from a bottle, still shaking his head in disgust at the faceless professors. He failed to offer Jo any refreshment.
When Bray faced Jo again, his composure had resurfaced. “What I mean is that the theory is fundamentally flawed, but impossible to refute on its terms. If you were to point out Butterfly’s unprecedented record, these theorists would attribute it to luck that would one day expire. No, the only refutation is to build a theory that proves the opposite. That a mathematical and scientific methodology exists which allows those who correctly apply it to consistently predict and outperform the market.”
“Hence the inefficient market hypothesis?” asked Jo.
“Precisely. The point is that the markets are efficient in the sense that prices reflect the aggregate expectation of participants. It’s kind of like the odds at a betting track. The more punters who think that a particular horse is going to win, the safer the bet is considered, but the less upside if it wins. Just like stocks. The more confidence investors have in them, the higher the price in terms of multiples of earnings or book value, so the safer the bet, but the lower the upside. But these odds, or prices, are just what the market thinks. A kind of emperor’s new clothes consensus. If everyone agrees this is the right price, then it must be.”
Jo was beginning to catch on, “But everyone agreeing doesn’t always make them right.”
“Of course not. People may draw a consensus opinion based on the information available. They may even have all the information, and it may all be correct, but it doesn’t follow that their conclusions aren’t flawed. The fact is, some people are just smarter than others. And some people, such as us, are infinitely smarter. We have what I like to call the informational edge. We have the ability to take vast amounts of seemingly disparate data and extract conclusions that are both correct and different from what the market as a whole thinks. And if what you believe is different, and you are always right, then you have found the key. The key to anything you may want.”
Jo was now rapt. She felt as if she had just joined the Evangelical Church of Bray. What he said made so much sense. The refutation of sheep like logic that had been entrenched in financial theory for decades was liberating and enticingly subversive. Jo had been converted.
Jo went to turn past the cover page, but Bray stopped her.
“Wait, we haven’t finished. What do you know about neural networks and anomalous market behavior?”
“Neural networks are part of those black box financial theories aren’t they? And anomalous market behavior means highly unusual movements.”
“Again, impressive textbook repetition. Remember, financial theories are only opaque to someone who is outside, or who is incapable of understanding them. What we achieve isn’t random, it is pure brilliance. Secondly, anomalous market behavior is another idiotic catch all. Markets can move in any direction, by any magnitude at any time. It just gets labeled ‘anomalous’ when the fund manager is justifying to a client why the client has just lost half his retirement capital. The only anomalous market behavior would be no movement; anything else is opportunity for us.”
Bray’s phone rang. He spoke briefly informing the caller he would be right there.
“Well Miss Lavelle, have to dash. We’ll get past the cover page next time then.”
Jo was disappointed.
“Can I take the book with me? To start going over it?” she asked Bray.
“If you like, but on the understanding that it doesn’t leave the department. It is the only physical copy. Of course it’s just the theory behind the model. The true beauty is in its application, which entails vast amounts of information gathering and sophisticated programming. Even so, the theory is invaluable and we insist it remain within these walls. It is after all the basis of almost all our most profitable trades.”
At the mention of profitable trades, the names of IBG and Money Trust Securities came into Jo’s mind. It occurred to Jo that this would be an opportune moment to raise some innocent questions about the application of Bray’s theory in these cases, but Bray ushered Jo, along with her courage, out of the door.
Back at her desk, Jo started going through the model. Its elegance was a great testament to Bray’s brilliant mathematical mind, and Jo was acutely aware of the privilege of studying this proprietary work and the good fortune of having the potential to fully understand it.
Jo gathered textbooks on many disparate mathematical and economic subjects from the well-stocked department library, and scanned them for sections that might help her understanding.
Although Jo was confident within the domains of mathematics and finance, she had never specifically studied their application in the field of neural networks. In fact, the definition she had provided to Bray practically represented the totality of her knowledge on the subject. After several hours of reading, however, her knowledge base had expanded significantly.
Jo learned that neural network theory emerged from research into Artificial Intelligence and that, specifically, a neural network was a program designed to mimic the learning behavior of the human brain. Based on the inputs fed into the program, the network processed the information according to the mathematical rules embedded in the model and provided numerical outputs, such as the anticipated level of a share index at some point in the future. The final output was the outcome of intermediate results produced at earlier points, or nodes in the network. Just as in the human learning process, information and conclusions were fed back upon themselves many times until a consistent pattern emerged, which allowed the system to make predictions about future outcomes or events.
Jo also found that there was also an overlap between neural networks and Chaos Theory. Chaos Theory analyzed a process based on the assumption that part of the process could be determined and part of it appeared to be random. A neural network’s ability to make predictions about both the ordered and random aspects of data made it ideal for modeling chaotic systems.
As far as Jo could fathom, the neural network resembled a large interconnected decision tree, the complexity of which was determined by the number of inputs and the rules governing the processing of the information. In the case of Bray’s model, the list of inputs were numerous and, in some cases, unexpected. Jo wasn’t surprised to see historical market trading data, interest rate information and various other economic indicators such as inflation and employment rates. Other inputs such as temperature differentials compared to seasonal norms and the number of vacation days taken by leading CEO’s stuck her as anomalous. She concluded that it must all be part of obtaining the informational edge.
About a third of the way through the main mathematical section of the model, Jo began to reach the limits of her understanding. It was a section on backpropagation and the algorithms employed to minimize the autocorrelation of residual errors between nodes. An understanding of the segment appeared integral.
Jo pushed
several text books off her desktop with a frustrated cry, giving vent to the petulance that emerged whenever she faced incomprehension in her subject area. Doggedly retrieving the books from the floor, Jo realized that something else bothered her about the whole concept of neural networks. It was the implicit claim that the network could extract rules that governed the performance of markets without the architect of the model having to formalize these rules, either at the outset or after the networks discharged its results. It was as if the network became an extension of the human cognitive functions, becoming literally a mind of its own; one that asked its creators not to question the wisdom it bestowed.
Jo was beginning to regard the whole concept as rather sinister when she noticed that it was almost midnight. She decided that the light of day would no doubt be the harbinger of dispassionate reason, casting out the unlikely ghosts Jo was imagining.
Having snuck into her parents’ apartment and her bedroom at such a late hour, Jo found a letter on her pillow. It was an invitation card from her old Cambridge college to their Boat Club dinner in celebration of the Jesus College sponsored Fairbairns Cup race. The race and the dinner were to be held this coming Saturday.
Jo had been the captain of the women’s rowing club at Jesus College the previous year. Enclosed with the invitation was a scribbled note from the current captain apologizing for the short notice. Jo smiled. Bridget was an excellent rower, but apparently the captaincy hadn’t improved her organizational skills.
The Boat Club also reminded Jo of Professor Radcliff, her brilliant mentor. He loved any activity on the water, being an avid sailor and a huge supporter of the rowing activities at the college. Jo also missed their long theoretical debates in his study, of which the Professor was invariably the winner. It occurred to Jo that he could certainly help her with the impasse she had reached in her understanding of Bray’s model. Due to her academic pride, she felt compelled to have a complete understanding of it before her next tutorial with Bray.
The Girl Inside Page 9