Realizing reliable return by studying echoes from the past
We are a quantitative asset management and research firm which helps investors achieve consistent sources of return through security prediction science.
Guided by institutional investment experience, we use machine learning AI science to predict, combine predictions, and construct portfolios in our continual effort to explore new sources of systematic return.
In every way possible, invest scientifically.
For a differentiated approach, turn the world on its head
Many quantitative managers seek return from strategies designed to exploit the behavior or "theme" of large cohorts of instruments, such as by market or factor, considered a "top down" approach.
Instead of specializing in the market or cohort, we study the individual stocks of which a market is comprised. Our approach models investor behavior at the stock level, systematically assigning each instrument its own unique model based on its own historical patterns. Relating the market as a multitude of prediction models distinguishes our perspective as "bottom up."
Our systematic, "bottom up" approach is statistically driven. We detect the direction and change in direction of investor behavior through a number of statistically driven methods that measure changes in price and other related statistically measurable phenomena.
“Bottom up” approach
Statistically driven
Highly liquid equities
Our approach
Our approach
1. Universe design
Our approach is most effective in any market comprised of highly liquid, large cap equities.
We include market constituents with historically relatable statistically measurable price history.
2. Predict risk & reward
Features are constructed that describe the price patterns and other characteristics of individual stocks across time.
Predictions are calculated from the features to set future expectations for each individual stock.
3. Portfolio Construction
Using predictions of reward and risk, Markowitz-diversified Long-only and Long-short portfolios are built.
Bespoke portfolios are constructed in line with client preference (such as risk tolerance and ESG).
4. Study outcomes & adjust
Predictions are continually studied and related to portfolio outcomes to help inform ongoing evolution.
PM intervention in client-traded portfolios may occur to ensure optimal client outcomes.
Pick the market and we'll seek to beat it
Our stock selection is optimal in any large subset of liquid stocks. In any robust equity universe, we will find a meaningful proportion of winners and identify those best to avoid.
Our stock predictions are used to create Long-only and Long-short equity portfolios that seek to beat the market index of the selected universe by the amount of active risk, over time, before fees.
US Long-only
Using predictions of reward and risk, Markowitz-diversified Long-only and Long-short portfolios are are built.
Bespoke portfolios are constructed in line with client preference (such as risk tolerance and ESG).
US Long-short
Long-short returns are uncorrelated to long-only portfolios, portfolio excess returns, and to the market.
Portfolio returns and portfolio excess returns have low correlation to interest rates, and are not correlated to any market, style or hedge fund type.
