“The first difference is that the Key Trends trading program has a probabilistic approach to trading 50 markets,” Baguenault de Vieville says. “Probabilities and risk are calculated thanks to two trend-following tools applied on two time scales; thus four sources of information are combined to estimate the predictive distribution function.”
What results is a probability number above or below 50%. If it is above 50%, a long position is initiated, if below, a short. The probability score also determines position size and is recalculated every day. This allows them to get into a trend early with low risk and build positions as it moves in their favor.
“The position is sized based on the quality of the trend and the perceived risk of the market,” Baguenault de Vieville says. “For example, a large long position in the portfolio could be the result of a high score value, for example a 60% chance of going up, along with a low perception of risk for that market.”
While it will add to winning positions as the scores improves, often a large acceleration will cause the model to reduce a position as it increases the probability of a sharp reversal. This came in handy last August in the Treasury complex as it reduced exposure as bonds spiked higher and gave back less during the correction.
The result of its reversal model allows KeyQuant to better manage trends at the beginning and end, which often makes the difference between profit and loss. “Our strategy is very continuous and has an exposure [to] almost all the markets, all the time. This allows us to profit from the beginning of trends, and not to be over-exposed at the end of a trend. The beginning and the conclusion of trends are where trend-followers distinguish themselves from each other,” Gelrubin says.
The strategy also aggregates the probability percentage across all 50 of the markets it trades in what it calls the “Global Economic Factor” (GEF) to alter its overall exposure.
“The GEF is an indicator of the quality of global trends. The GEF can oscillate from 0.5 (low confidence) to 1.5 (high confidence),” Gelrubin says. “Trend-following investors understand that CTAs have some very appreciable run-ups, some large drawdowns and most of the time is just oscillating. The GEF tries to increase exposure during good times, being at the average on oscillations and [decrease exposure] at the very beginning of reversals.”
Another big difference is their approach to risk. KeyQuant does not use standard volatility or value-at-risk models, which Gelrubin says are ill-fit applied to trend-following. “We developed risk measures that fit with the tools used to determine the probabilities of price direction,” Gelrubin says.
As noted, KeyQuant is not a typical emerging trend-follower and people are taking notice. It has been nominated for a first-ever Pinnacle Award as one of the Top Emerging Managers by BarclayHedge and CME Group.