Lets start with the basics then we’ll get into the specific strategies:
Quantitative investing is an approach for implementing investment strategies in an automated (or semi-automated) way. This approach lends itself well to (1) using large or unique data sets, (2) refining them into explanatory information, and (3) deploying that information via as trades using technology. Every quant investor is looking for an edge, so we’ll explain how these elements are used to capture edge.
The best quants employ the scientific method; they come up with a hypothesis based on a real-world observation, then they test it. All individual strategy edge erodes over time, but the key to coming up with a winning hypothesis is to understand the most profitable themes in finance. Then, develop a process for sourcing and expressing those themes. Some specific examples:
1. Data sets
Most people equate quantitative investing with advanced PhDs and high-level cutting edge math, but advanced math is often not
the core driver of edge. Many of the most profitable quant strategies are actually very straightforward to understand. Sourcing and implementing unique data sets is often one of the most intuitive ways to extract an advantage. Some examples:
overnments create a multitude of opportunities for pure gamification. They are motivated by politics rather than profit, and there are numerous agencies and national regimes that create messy, often contradicting rules. Quants look to reverse-engineer these rule structures and pinpoint inefficiencies within a system (and across systems) so they can capitalize on them. Governments tend to be slower to respond than profit-motivated actors, so the inefficiencies they create can persist for a long time (often indefinitely). For these reasons they are a profitable source of low-hanging fruit for quants:
(i) Classic Regulatory Arbitrage.
Financial actors often scour the rule systems of regulators in an effort to find inefficiencies. One example: the electricity regulators have a reputation for being so incompetent that their complex rules and regulations provide electricity traders with innumerable opportunities. As Bloomberg put it, “FERC (the electricity regulator) builds markets with so many bells and whistles and buttons and valves that some of the buttons end up having no function but to dispense money. If you can find those buttons, what you do is just keep pressing them until the FERC notices and gets mad at you… 
(ii) Cross-Border Regulatory Arbitrage.
Example: U.S. regulators require companies to report quarterly financials, but companies listed on the Taiwanese exchange are required to report monthly sales. Quants take the monthly sales for companies listed in Taiwan 
and develop signals based on that information to gain an understanding of expected U.S. company performance. They can then use those signals to trade equities, options or ETFs more effectively compared to most other U.S. investors who simply wait for the quarterly and annual numbers to come out.
(iii) Inter-Agency Regulatory Arbitrage.
This is when quants use the fact that rules have a tendency to conflict across different regulators within the same system. Example: The implementation of Dodd-Frank legislation (which hedge fund titan John Paulson artfully referred to as “gobbledygook”
) has been a field day for quant investors who decode rules. FDIC-insured banks & thrift institutions must now report ‘Call Data’ to the FDIC, which requires disclosures of earnings, among other things. Many banks are now inadvertently reporting earnings via FDIC reports ahead of their quarterly SEC 10-Q reports, providing quants an opportunity to take advantage of the timing discrepancy.
Much like governments, exchanges come up with specific trading rules that can be gamed:
(iv) Time Zone Arbitrage 
. An arbitrage existed whereby some international mutual funds could be gamed due to differences in time zones. Per the rules, all mutual funds had their prices set end of day at 4pm EST, when U.S. exchanges closed. The problem was that for some international mutual funds, their markets had already closed prior to 4pm EST, which meant that investors could see the closing prices before the actual close. They would then simply algorithmically buy funds that they knew would be priced higher than the price being paid. (This practice was later shunned after some mutual funds had accepted bribes to mark trades that happened after
the close as happening prior
(v) Flash Pricing. Several quantitative approaches that are oft cited in discussions of high frequency trading are actually based on exploiting exchange rules. Quants use “flash pricing”  to get a sneak peek on large order flow and then trade microseconds ahead of other participants by stepping in the middle.
(vi) Rebate Arbitrage.
This is a tactic that uses an exchange rule that seeks to reward market participants who provide liquidity to the exchange versus those that remove liquidity. Several high frequency approaches seek to take advantage of this ‘money button’ by placing trades that neutralize the market impact of their bets while maximizing their free rebates.
Market Participant Rules
In addition to the inefficiencies created by governments and exchanges, market participants have their own rules to trade against, whether it be institutions with their own unique investment protocols or individuals with behavioral biases. Examples:
(vii) Algorithmic Pattern Recognition. One significant area of market innovation of late has been in pattern recognition. Back in the simpler days, if a big institutional order came into a brokerage house, the broker would shop the order around to multiple other brokers to fill the big trade. If broker Mike at Morgan Stanley called broker George at Goldman Sachs, George might be able to intuit that a big order was being processed and keep some shares for himself while selling some of the others to Mike to fill his order. Brokers would track volume moves and the information at hand to “Read the Tape”  to try and take advantage of big directional moves in a stock due to these block purchases.
Nowadays, all institutional trading is done via electronic algorithm, where orders are routed in staggered patterns to multiple exchanges as well as different brokers, dark pools, and crossing networks, all in effort to fill big orders with minimal market impact. Instead of reading the tape, modern quantitative funds now work on the other side and try to “break the code”. In other words, they seek to recognize and isolate custom trade execution patterns in an effort to trade against them.
(Yes, we did quote the Dalai Lama in an article about quantitative hedge funds.)
(viii) ETF Rule Trading. When a stock is being added to an index, the ETFs representing that index often MUST buy that stock as well. By understanding the rules of index additions/subtractions, hedge funds can trade ahead of the forced buying and capitalize on those rules.
(ix) Prospectus Arbitrage. Many mutual funds & hedge funds have their own investing rules. For example, many mutual funds arbitrarily set rules for themselves that they cannot own a stock under $5. Others must only invest in stocks that meet their specific “growth” or “value” characteristics. For a savvy quant investor, they can use textual analysis to scan for these types of rules across prospectuses, source the publicly available information on mutual fund holdings (via form 4s and 13F filings), determine which funds hold assets close to their stated thresholds, and trade against those constraints.
(x) Behavioral Biases. Many retail investors have well-established psychological biases. For example, retail investors have a tendency to cut winning positions add to losing positions due to a loss-aversion bias. Quants can identify general behavioral biases among certain classes of investors, isolate which stocks express those biases and are favored by the class of investors, then trade against the irrational behavior as a source of return.
(xi) Other Pure Informational Advantages
Some funds focus on finding unique data sources to extract an edge. A very interesting WSJ article 
- Some funds use satellite imagery to determine whether crops are growing at the expected rate in order to estimate commodity supplies & prices.
- Others use satellite imagery to gauge whether parking lots are full or empty at specific retailers as a way to anticipate sales.
- Others measure the shadows cast from buildings to estimate the rate of new construction in major cities.
The list goes on. There are an innumerable number of clever, legal ways to find better, faster information rather than simply waiting for the quarterly & annual reports to come out.
2. Refining The Data
All of the above thus far describes different types of trades and data sets that can be used to extract an edge. Often the uniqueness of the data alone is enough to confer an obvious advantage, but additional edge can be extracted using the best techniques to scrub & refine the information. This is often where those PhD-level mathematics can provide a leg up. Monte Carlo simulation, machine-learning algorithms, refinements to traditional regression analysis or other means can contribute to higher predictive values for a given date set.
(xii) Deep Math Applications
The inefficiencies & data sets above are intuitive to understand and do not fundamentally require advanced math applications. That being said, there are strategies that are only explained with advanced math. E=mc^2 is not a fundamentally intuitive concept, but it has been used to explain a vast array of knowledge that didn’t exist prior to its discovery. The same discoveries can be unearthed in finance (though no one would broadly disseminate them as long as they are effective.) An example of a financial field where advanced math is almost mandatory is options.
Options (and derivative securities in general) have more complex mathematical underpinnings than traditional stocks. As opposed to stocks that move primarily in relation to the health of the company and the broader economy, the value of options are are also affected by (i) the passage of time, (ii) the volatility of the underlying security, (iii) the movements of the broader market, (iv) the volatility of the broader market, and other key factors. All of the factors change non-linearly with the movement of the others, so the higher-order moments of each variable can have a meaningful impact on value of the option. Add these complications to the fact that the options market has its own unique opportunities for trading & rule gamification and you suddenly have a very intense math problem.
Sometimes there is no simple underpinning to solving these inefficiencies and it comes down to building the best mathematical mousetrap to assess differences in price vs. value.
3. Deployment methods
The last main category of quantitative edge can be found through deployment methods. If a fund can source & refine data on par with other quant funds they can still lose if they are slower to deploy those algorithms and trades. Additionally, there is valuable information released every day that can move the markets (from company specific information to economic indicators), and those who can process that information and trade on it faster can win.
The reality here is that there is such a diversity of profitable quant strategies that deployment is one of the hardest edges to maintain. High-frequency trading (HFT) has become a veritable arms-race to zero latency (ie: trading at speeds approaching zero microseconds.) That being said, there can be winners in an arms-race and there have been firms that have benefited from highly profitable advancements:
(They use microwave towers, not actual microwaves FYI.)
(xiii) Microwaves. For HFT firms, fiberoptics are a painfully slow way to communicate. The problem is that the speed of light is somewhat hampered by all that bouncing around inside the optic cable, and it slows the information down. To solve for this, firms now use microwave transmitters, which communicate directly from point A to point B via a less convoluted route. The result is a transmission time that is up to 50% faster, which saves precious microseconds.
(xiv) Server Co-Location.
Another way to trade faster with an exchange is to ‘co-locate’ or to purchase a server directly on site at the exchange. HFT firms pay top dollar not only to co-locate servers but also for the front-row seats which cut down the physical distance by multiple centimeters (or maybe even meters!).
(xv) Better Algorithms. Beyond the hardware considerations, HFT firms are constantly looking for faster ways to process their algorithms and shave off processing time. This is done via a combination of software (and hardware) advancements that eliminate every microsecond possible.
We hope this has been helpful and interesting! We look forward to any comments or feel free to email us if you’d like to discuss this topic further: firstname.lastname@example.org