Page 1 of 2

Wall Street is flush with precocious 20-somethings filled with the sense of their own shattering brilliance, but if old Street hands are right, three Harvard- and MIT-educated millennials who’ve never come within a state’s throw of a Friday night midtown Manhattan happy hour may be on the cusp of engineering a genuine revolution in financial technology. And they’ve arrived at this point, appropriately enough, because of a bet.

A couple of years ago, Daniel Nadler, 29, told his friend Peter Kruskall, 26, that it was possible to model equity markets to show that individual stocks move in highly predictable cyclical and seasonal patterns. Nadler, who is completing a Ph.D. in political economy at Harvard University, had recently taken up residency as a visiting scholar at the Boston Federal Reserve, where his work focuses on tracking the asset allocation strategies of the 300 largest U.S. institutional investment funds. Kruskall had graduated from the Massachusetts Institute of Technology with a masters in computer science and was employed, at the time, as a software engineer with Google.

The two, Nadler says, had become acquainted during Kruskall’s time at MIT over their shared interest in Zen Buddhism. Kruskall’s reply to Nadler’s claim about financial market modeling was very un-Zen-like. “He said, ‘That’s garbage,’” Nadler recalls.

A challenge was born. “I said to him, ‘Give me one model that I can test locally myself with programming, give me a day to test it, and if it works out well, we can go from there,’” says Kruskall.

Top to bottom: Peter Kruskall, Daniel Nadler,
& Brandon Liu of Kenshō Technologies

What the pair very quickly discovered was that there was no way for Kruskall to evaluate Nadler’s claims. It was difficult to get access to the data needed and, more importantly, the software to model markets simply didn’t exist for anyone outside the financial industry. More easily distracted minds might have given up there. But for Nadler and Kruskall, these two children of what Nadler terms “the Facebook effect” — the thriving culture of technological entrepreneurship that has emerged at Harvard and MIT in the wake of Mark Zuckerberg’s success — the aborted bet set them on a more ambitious path: Kruskall began to build out a software environment that would allow them to model and test Nadler's boast.

Today, the fruit of that first failure is a website, Seasonal Odds, which launched this week, that Nadler says will be “transformative to markets” and to the way multiple different categories of investors approach portfolio management — “from the doctor or lawyer with his Charles Schwab account,” as Nadler puts it, to brokerage firms, family offices, banks and even small and mid-sized hedge funds.

It will do this, Nadler says, by offering a “Bloomberg terminal in your browser,” a purely web- and cloud-based service that will offer many of the standard technical, risk analysis and portfolio optimization tools used by the largest, most sophisticated hedge funds. The website currently displays data for over 7,000 individual stocks, sectors, equity indexes and ETFs, across the New York Stock Exchange, the American Stock Exchange and Nasdaq. The plan is to eventually roll the service out to include stocks listed on other exchanges, throughout the U.S. and beyond, and eventually, a broader range of asset classes and instruments, including bonds and derivatives. “We are seriously democratizing the tools of high finance,” Nadler says. In the process, he adds, they hope to pioneer a new field of outsourced statistical computing for Wall Street.

This may all sound preposterously grandiose, but others unaffiliated with the company who have seen and tested the product in beta form are singing from the same ecstatic hymn sheet. “It’s potentially a very big deal,” says Josh Abramowitz, chairman of Deep Creek Capital, a New York–based investment firm active in financial technology projects. “It potentially opens up a whole new toolkit, in a way that many different investors, at different levels of sophistication, can access and understand.”

Abramowitz, who worked as a portfolio manager and research analyst at hedge fund firms Viking Global Investors and Elliott Associates before launching Deep Creek Capital in 2009, has not invested in Nadler’s company, but he says he may. According to Nadler, the company has already secured deals for $750,000 in venture capital funding and is in discussions with several venture capital firms and hedge funds and one major Wall Street financial data provider regarding further investment once the platform launches.

Today, only a small number of elite hedge funds, such as AQR Capital Management, Bridgewater Associates, Renaissance Technologies and Two Sigma Investments have access to the types of tools Seasonal Odds features, Nadler claims. “Large quant funds spend lots of money on these technologies, because they’re very powerful,” explains Abramowitz. “If you can replicate that with much less effort and cost, that’s the kind of thing that could potentially be a game changer on Wall Street.”

Heidi Johnson, managing director of collaboration services at Markit, the London-based technology and financial data provider that serves many of the main players on Wall Street, says the true innovation in Seasonal Odds is its adoption of many of the same technologies at the heart of the revolution in social media and consumer technology: high-speed search algorithms, machine learning (the technology that allows Netflix to make intelligent suggestions for films you might like to watch, based on your previous viewing and website navigation habits), “big data” — or its more ancient corollary, the ability to crunch large datasets at high speed — and cloud computing. Johnson says that Markit is in the process of integrating Seasonal Odds’s content into Markit Hub, a research, data and analytics platform, accessed by 200,000 users in the financial industry, of which she is the product lead.

“This is Silicon Valley technology being brought to Wall Street,” Johnson says. “That’s exactly what the markets need.” The Seasonal Odds team takes Google as “their gold standard,” she adds. “That is not typically the standard one sees in financial services today.”

Nadler and Kruskall took on a third partner, Brandon Liu, 20, after beginning work on the project in earnest around a year ago. Nadler describes Liu, 20, a junior at Harvard majoring in computer science, as a “very talented hacker in the true Social Network style.” The trio call their company Kenshō Technologies. Kenshō, according to Nadler, is a Zen Buddhist term roughly analogous to the concept of “seeing reality with a perspective you didn’t have before.” The company is currently engaged in a battle to get the regulatory authorities in Delaware, where the company’s registration is pending, to recognize the macron over the “o” in Kenshō, Nadler says. (The business of engineering shifts in the perception of reality starts small.)

The use of seasonality as a tool to maximize returns is well known. Many of the main equity indexes, such as the S&P 500, have historically performed better at certain periods of the year than others, meaning investors can optimize their portfolio allocations by cycling in and out of their index positions according to historically sticky seasonal performance patterns. This is the wisdom behind the so-called “January effect,” which holds that stock prices generally increase in January, as well as the dictum, “Sell in May and go away,” the idea of which is that equities historically underperform from May to September and it is best for investors to seek alternative asset classes during that time.

The innovation of Seasonal Odds is to give investors a simple, self-updating snapshot of how individual stocks have performed on a monthly basis since 1980. In the process, it is hoped investors will themselves gain kenshō, or the ability to perceive a reality that previously had been hidden amid the undifferentiated mass of ever-changing stock data.

Richard Ross, a market technician at Auerbach Grayson, a New York–based brokerage firm, says that while he employs seasonality as a strategy across indexes and asset classes, he’s yet to hear of anyone applying it at the level of individual stocks. But “I don’t know why it wouldn’t work on a more micro, stock-specific level,” he adds. “If you get a big enough sample size, there’s certainly an edge to be gained here.”

To prove this thesis, Nadler, Kruskall and Liu have since November been conducting a live, controlled test, across the 20 most widely held stocks in the U.S., of the kind of seasonally responsive portfolio cycling investors could use Seasonal Odds to pursue. Whenever the seasonal odds almanac told them that there was a greater than 65 percent likelihood of a stock recording a positive monthly return, they bought it; if there were greater than 65 percent odds of a negative return, they sold the stock. From November to March, of the 36 times that the odds almanac sent them a signal to buy or sell a stock, it was right 34 times.

“No known pundit came anywhere near to calling the month-by-month performance of the most widely held stocks with such accuracy,” the trio note.

This strategy of picking seasonal winners, factoring in broker fees of $20 a trade, yielded a simple cumulative return of 15.7 percent over the five months from November to March. The S&P 500, by contrast, finished up 8.6 percent.

Seasoned equities traders to which Institutional Investor described the Seasonal Odds project suggested that anyone could generate the same seasonal returns charts for individual stocks by downloading the relevant data off Bloomberg, Reuters, Datastream or even Yahoo Finance. Nadler’s response is that calculating the averages manually, across a family of 7,000 stocks as Seasonal Odds has done, is simply not an option for most people on Wall Street — and that his company’s use of sophisticated computation tools means that its system automatically updates as time goes on and stock prices change, a task essentially beyond the realm of manual human tolerance.

“It’s not something you can do in Excel manageably,” he argues. “That’s where computers can become very useful. They can take manual processes and automate them, and then scale them very, very quickly.”

The idea of Seasonal Odds is not to use seasonal portfolio cycling as a standalone strategy. The service includes many other features, including a risk/return plot — with Google maps-style zoomability — that shows how the ratio of risk to return of each stock for which they have data compares against every other one in their current universe of 7,000 securities. Nadler and others, including Johnson, describe this graphic, despite its apparent simplicity, as a public resource unique to the marketplace.

Other features include a “daily odds almanac” that shows, in one figure, the odds of each of the stocks covered finishing that particular trading day positive, based on the real-time price of the stock and the probability of a price reversal under similar historical conditions; a curated Twitter feed of the social media chatter around individual stocks, which Nadler compares to the Twitter service that Bloomberg recently introduced on its terminals; a patent-pending sentiment signal that visually depicts, with the help of a proprietary natural language processing algorithm, how social media sentiment around a particular stock is shifting in real time; and a live “smart money grid” that shows, as prices shift, the stocks in which the biggest 25 institutional investors, as ranked by Institutional Investor, are building positions. Johnson describes these elements, especially the sentiment signal and the risk/return plot, as Kenshō’s “special sauce,” and as potentially even more transformative to markets than the seasonality data.

Nadler, who says he has been “obsessed” with finance since the age of 16 and whose doctoral research at Harvard focuses on the intersection of sovereign bond pricing and political risk analysis, concedes that investors will not see value in or want to use each and every one of these features. He cites the suggestion put forward by George Soros, in his book “the Alchemy of Finance,” that successful investment is part science, part alchemy. “The science part is where we fit in,” he says. “We don’t pretend to be the brilliant hedge fund manager. That’s the alchemy.”

Why is this data not more readily available? The answer, according to Nadler, comes down to the fundamental selfishness of Wall Street’s biggest firms: The data — and the computational science to crunch that data — is what gives them an edge, so there's no incentive for them to open it up to a broader audience.

“It’s not that this data is not out there; it’s out there, but it’s behind closed walls,” he says, adding that Seasonal Odds pays for the cleaned data it feeds into its system from a well-regarded data provider, whose name he will not disclose, used by many financial firms. “For most people what we’re doing seems kind of irrational,” he explains. “On Wall Street, if you have some kind of differentiating edge, you trade on it. Silicon Valley thinks in terms of giving people tools: you don’t become a gold miner, you develop a better pick axe and sell that to people. On Wall Street you just start the gold mine.”

Nadler believes that he, Kruskall and Liu are pioneers in another sense: They are young, would-be tech barons who have turned to Wall Street, rather than the ping pong tabled fun rooms of Silicon Valley, to turn their talents to profit. “Facebook and Google did and do provide a lot of social pluses to working there,” explains Kruskall, who worked for two years as a software developer at travel website Kayak before moving to Google in 2011, where his work touched heavily on the core search function. “But times are changing; they’re both getting to be very big companies, and while they’re still wonderful places to work, it’s a very exciting time to be involved in startups. So I figure, ‘Just go for it.’” For Nadler, the recruitment of a person like Kruskall, who resigned from his job at Google earlier this month to work full-time on Seasonal Odds, represents a huge coup, because most young software developers have no interest in applying their wit to financial problems.

The trio have yet to decide exactly how far the “democratization” of their service will go. Nadler insists that “stuff that you can’t get on a Bloomberg terminal we will give away for free,” either directly on their website or through partnerships with popular consumer finance websites, but a tiered, fee-based service offering deeper data and customization is also in the works, aimed at the sophisticated professional retail investor at its cheaper end and financial institutions at its more expensive end. Guaranteeing the reliability of securities data as the service expands to less developed, emerging markets will also be a challenge. And, perhaps most important, the task of dislodging Bloomberg and the other legacy big beasts of the financial data jungle from the Wall Street firms that rely on them is immense.

“This is a huge issue for all tech startups selling into institutional finance,” says Deep Creek Capital’s Abramowitz. “But if you have something which is truly revolutionary — which you may have here — then it’s just a question of time before it reaches a wider audience.”

The challenges are large, but so is the appetite to meet them. The question now is whether Seasonal Odds’s usefulness to investors can prove itself to be at the same level as its founders’ bounding ambition.

Single Page    1 | 2