Richard Olsen: Tuned to high-frequency Part 3

WE: Why is such a large part of the forex market operating intraday?

RO: Very simple. The biggest thing that determines how long you have to be in a position is the spread you pay. If the bid-ask and the fees are .5 percent, then you’ll have to wait until price, on average, has moved at least .5percent.

Without people realizing it, that’s what dictates how long positions are held in any financial market. In foreign exchange, the spread is 0.02 percent, so you can be out with a profit — or a loss — after, literally, five minutes.


WE: Does the forex market have characteristics distinct from other markets?

RO: First, it is ideal for speculative trading, exactly for the reason many people have shied away from it — because it’s so efficient. But it also offers far more opportunity than the typical equity market.

Opportunity potential is defined by annual volatility divided by the current spread. For example, if annual volatility for an equity market is 30 percent and the typical bid-ask spread and ticket costs amount to, say, 0.3 percent, you get a factor of 100 (.30/.3). If you are correct every time you trade, with this 30-percent volatility you can do 100 times your spread.

Take the same approach in foreign exchange. Assume annual volatility is 12 percent and the spread is .02 percent — you get a factor of 600, which means is it is potentially six times more profitable to trade foreign exchange.


WE: Tell me about your money management firm, Olsen Invest.

RO: We’ve been doing it for two and a half years. Prior to that we were offering banks a kind of trading signal service, as a purely advisory activity. Our return, I think, is quite impressive. We run a Sharpe Ratio of 2.23. Last year’s return for a standard risk profile was 15 percent. This year we’re up around 8 percent.


WE: What kind of trading approach do you use for this?

RO: Purely quantitative. We have computer models that read in tick-bytick data, and take positions based purely on this data.


WE: What kind of time frame?

RO: The length of trade, on average, is six hours. Our objective is to shorten that. The maximum period might be up to several days, but in general we try to get in and get out as soon as possible.


WE: Can we discuss a representativestatistical observation or pattern that would provide the basis for a trade?

RO: Our trading strategy is based on a broader economic theory. We are able to earn money because we provide a service — or value-added — to the market and indirectly to the economy as a whole.

Our value-added is that we provide liquidity and are ready to warehouse risk when many other market participants are running for cover. We try to open positions whenever there is a price shock or “tail event” on an intraday basis. The size of a position is determined by the model’s overall assessment of the market environment: Are we in a trending or sideways market? We close out our position regardless of the overall market environment as soon as we have reached our profit objective. We do not believe in sitting on positions too long.

Positions are like inventory in a warehouse: We like to make sure we clear the inventory at a regular frequency. If our trading idea has proven to be premature and we incur an unrealized loss, we increase our exposure — provided the model has not changed its longer-term assessment. We continue to rebalance our positions to optimize the risk profile of our overall exposure. We do not like stop losses and use them only to prevent large-scale drawdowns.


WE: This sounds like you react to “order flow” — meaning, simply, the aggregate OANDA trading activity at a given moment by going long when the market had overextended itself to the downside and selling on the opposite condition. In effect, you’re profiting by being an effective self corrective mechanism for the marketplace, this being something that is difficult for the smaller retail trader to do. Am I in the ballpark?

RO: I have to stress that our models are exclusively driven by the price feed of OANDA FXTrade. Anyone subscribing to the OANDA price feed could build our type of model. Our models analyze microvolatility, which is volatility computed using tick-bytick prices, and on the basis of this makeinferences about market state and the likelihood of self-corrective mechanisms, as you call it, in the marketplace. The small retail trader is unlikely to build such an elaborate system as we havedeveloped. Having said this, the retail trader has other advantages over the professional money manager. He can adopt a higher risk profile and thus achieve a higher return, albeit at a higher risk. Long-term, I see a strong demand for sophisticated trading indicators, which incorporate some of the refinements that I have mentioned earlier.


WE: Does the money management work overlap with OANDA.com at all?

RO: We execute exclusively through OANDA. One important thing is the investment management is a completely independent business that we trade through OANDA. The reason I participated in the founding of OANDA was because, as a trader, I always thought it would be great if there was a brokerage that offered me these features.

An important feature of OANDA FXTrade is that every customer trades at the same price, whatever his position. So, Olsen Invest customers trade on exactly the same price as every other customer of OANDA. Olsen Invest does not take advantage of inside information to “trade against the customer order flow.”


WE: How does the discontinuity and other “inefficiencies” you might have discovered — in tick data translate into trade opportunities?

RO: The example I just gave is such an opportunity. The short-term discontinuity, or price extreme, is embedded in the long-term memory effect, or mean reversion. I use these terms in a fuzzy way. I want to refer to a broad range of different types of memory effects — they exist in many different shapes and forms. An overbought or oversold indicator is a means to quantify the size of a price discontinuity and determine when the price is ready for a bounce back.


WE: Based on your understanding of price behavior, what kinds of trading ideas or approaches — especially the better-known ones — do you think lack merit or effectiveness?

RO: Traditional moving averages are overrated — they fail to account for the dynamic nature of price evolution. Someone who uses such an indicator gets stuck in the rut of a particular analytical time range and forgets about the nonlinear interplay of market participants trading on very different time scales — day traders, regular speculators, portfolio managers, treasurers and central bankers.


WE: Do you concentrate on particular currency pairs?

RO: No, we spread our trading across a wide range of currencies.


WE: What kind of leverage do you use?

RO: We run programs with different risk profiles. Our standard risk profile has an average leverage of approximately 0.8.


WE: What do you think of the growth in the retail forex industry over the past few years?

RO: I was always frustrated by how the big banks discriminate against small traders, making it [prohibitively expensive] for them to trade. So I believe it’s a great opportunity because, in the end, why should only the big guys earn good money? So [the development of the industry] is not just good for the small trader, but for the industry as a whole because longer-term it will contribute to overall market stability. If the big guys all [rely on the same information] they’ll drive the markets in one direction. The more smaller players you have involved, the more diversity you’ll have in trading styles.


WE: Is a corollary of this that the massive long-term trends that used to develop will be diffused to a certain extent?

RO: Yes — this is my hope. I have to add a point here: My big surprise with the OANDA FXTrade platform was the observation that the trading behavior of the retail market was far closer to that of the professional market than I would ever have guessed. For example, from the very outset, we at OANDA FXTrade would have a low volume day, when UBS or Deutsche Bank had a low-volume day.


WE: Are there any risks unique to forex that typical stock traders should be aware of?

RO: All the defects a trader has in equity markets will be exaggerated in forex trading.


WE: Is that just a matter of the high leverage available in forex?

RO: Yes. That’s point No. 1. But it’s also because forex is a 24-hour market and prices move with unbelievable speed relative to the spread. The return and risk potential are very big. It’s like driving on the highway: you can get very far, but make sure your seat belt is on and you know how to drive.


WE: Have you applied your tick-data analysis to markets other than forex? Do you have plans to branch out into other markets?

RO: Yes, we’ve researched the fixed income, commodity and stock markets. They all obey the same rules, subject to complex intricacies, which are a result of their specific market structures and conventions. But for the time being, I want to remain focused on foreign exchange. As the saying goes,the devil is in the details and I want to excel in the foreign exchange market.


WE: Have you done much research regarding the specific trading applications of certain economic “events,” such as the response of forex to interest rate adjustments, trade balance numbers, GDP and other economic numbers?

RO: We’ve done some work. We know the “simple” ways don’t work — the linear indicators fail to perform.


WE: Meaning, a one-input, cause and effect, “What does Market A do in the two days after Report B is released?” type of trading model?

RO: Yes. Interestingly, the neural network type approaches people tried out over the past 15 years don’t work, either. I believe in using a staged approach. Like an archeologist, it’s important to identify different “layers” and resolve the complexities of each layer separately. The overbought-oversold indicator I mentioned has two layers — the definition of the actual indicator and, second, a method that accounts for the 24-hour seasonality of volatility. I definitely plan to develop specific trading applications for economic events, but a lot of work has to be done to make this a success.

I have lots of ambitions. I want to create a few successful companies and have some real fun. It’s so easy to be cynical, but I just want to do some good work.