strategy Archives - XMonetae Capital

The XMonetae Assembly Line: An Inside Look

5th March 20201st October 2021
by Julian Marchese

XMonetae Blog Post written by Julian Marchese (CEO & CIO) & Nigel Noriega (CTO)

In my last post, I touched on the concepts that lay the systematic foundation of our investment philosophy. In this post, I’d like to show how these concepts are implemented in practice – more specifically, XMonetae’s “assembly line” framework.

One of our main objectives is to provide our investment team with the necessary tools to most efficiently go from a strategy idea to a live production trading algorithm. These tools fall into two broad categories for our assembly line:

Rigorous conceptual strategy development cycle
Technology infrastructure facilitating a robust workflow

The diagram below describes the XMonetae assembly line at a glance:

The Strategy Development Cycle

A rigorous cycle for testing and experimenting with new strategy hypotheses allows for Idea Generation at every step in the process. Each stage of this iterative system gives the investment/research team feedback on when to accept, reject, and create new hypotheses. It’s critical, however, to be thoughtful in the early part of the assembly line.

Bias early on in a strategy’s development is when its most harmful. As a result, our cycle mitigates bias at various steps during the conceptual assembly line’s first half:

Unambiguous Hypotheses: The strategy hypothesis (i.e. the economic intuition for why you should be paid to take some kind of risk) must be well defined prior to testing. Conclusions derived from testing the strategy cannot influence the initial hypothesis and bias strategy optimization.
Representative Data Samples: The in-sample period(s) should include varied market regimes for a complete picture of true performance (i.e., if you seek to develop a long-only stock trading strategy, be sure to include 2008 in your backtest!).
Limited Optimization After Out-of-Sample Tests: If backtesting results are too poor relative to assumptions prior to testing, unapologetically move on to the next idea in the spirit of iterative design.

The Technology Infrastructure

Implementation of industry-standard technologies enables all the measures taken in the conceptual assembly line. Data acquisition, storage, and proper backtesting techniques are all an integral part of XMonetae’s process. The various key services the infrastructure provides are:

Comprehensive Databases: A variety of collected and easily accessible financial data (e.g. orderbook depth, ticks, OHLC, fundamentals).
Rigorous Data Cleaning Methodologies: Data from many different sources unified into one format and pruned for outliers/inconsistencies.
Data Science Tools: Standard statistics software for investigating and validating behaviors across asset classes.
Realistic Backtesting Processes: A mirror of live trading environments to minimize time and unforeseen obstacles between strategy development and deployment.
Interactive Visualization Dashboards: Real time graphs to depict different dimensions of strategy and portfolio performance.
Extensive Alerting System: Automated text/email alerts for monitored fund performance, risk metrics, volatile market movements, and atypical trading behavior.

As complex as a hedge fund’s infrastructure may be, new investment team hires should be able to intuit how to work with the existing assembly line framework. All documentation needs to be comprehensive so researchers & developers are able to improve functionality without significant time loss. With thoughtful documentation, any infrastructure can be easily refactored and improved with the advancement of relevant technologies and data science.

Once the major components of the technology are in place, efforts can be re-allocated towards furthering data acquisition. The assembly line and the alphas it produces are only as effective as the data quality fueling the process.

Reproducible Success

Ultimately, the assembly line process is a powerful tool when wielded properly. This is why it is equally important to maintain robust procedures to mitigate bias when discovering new alphas. We establish rigorous pipelines and workflows at XMonetae to ensure we are efficiently using the tools at our disposal, maximizing our portfolio’s chance of producing risk-adjusted return in a systematic and reproducible way.

DISCLAIMER

NOT FDIC INSURED | MAY LOSE VALUE | NO BANK GUARANTEE | PAST PERFORMANCE IS NO GUARANTEE OF FUTURE RESULTS. There is no guarantee that investment objectives, or risk or return targets discussed in this material will be achieved. The data provided is for information purposes only. This material is not intended to be read in isolation and may not provide a full explanation of all the topics that are presented and discussed. Information contained in this material has been obtained from sources believed to be reliable, but its accuracy is not guaranteed. Furthermore, there can be no assurance that any trends described in this material will continue or that forecasts will occur; economic and market conditions change frequently. This material should not be considered a recommendation or offer to purchase or sell any particular investment. Investors should consult a financial advisor/financial consultant before making any investment decisions.

About Us

The Power of Strategy Diversification Via Automation

16th December 20191st October 2021
by Julian Marchese

In our last two blog posts, I went over some examples of opportunistic & event-driven trades that fit within our wheelhouse (ie. forced liquidity arbitrage taken advantage of in a market-neutral manner). In this post, I’d like to touch on the systematic element of our investment philosophy. I would recommend that you review our blog post “Our Investment Philosophy” as a starting point.

Finding The Elusive “Holy Grail”

The Whole is Greater than the Sum of its Parts
-Aristotle

Many traders seek a “holy grail” investment strategy. We don’t believe that such a thing exists. To be clear, in the context of an active trader or portfolio manager, a “holy grail” would be an investment strategy that, over the long-run, consistently generates high risk-adjusted returns (say 2.0+ Sharpe Ratio) with little to no correlation to major market factors. It’s also important to define “investment strategy” in this context as being a defined set of rules designed to capture a specific alpha.

That is not to say that there aren’t very high risk-adjusted returning investment strategies out there – there are. But how long will they last? The more lucrative an alpha source is, the more incentive there is for other sophisticated traders (and many are very well-incentivized!) to find and capture it. In general, the better the returns have been for a investment strategy, the faster those returns could decay – an uncovered “holy grail” may tarnish.

We believe that the real “holy grail” is the effective implementation of an idea rather than a single strategy. To us, that idea is a dynamic portfolio of uncorrelated investment strategies. Ideally, these are automated algorithms that capture structural alphas across the global financial marketplace (ie. taking the other side of higher time-frame or forced market participants, with the clues of their impact ascertained through the scientific process). The individual component strategies of such a portfolio may not have “holy grail” characteristics – however, the whole is greater than the sum of its parts.

With this in mind, we would argue that, in the long-run, it is actually more fruitful to look for multiple low-Sharpe component strategies and combine them within a dynamic regime-based meta-strategy, rather than gunning for the single 2.0+ Sharpe strategy. One could also make a reasonable argument that it is exponentially “harder” to find and implement high-Sharpe strategies (ie. we reason it should take less time / resources / intellectual capacity to find a 0.5-1.0 Sharpe strategy vs. a 2.0+ Sharpe given market competition – though of course it’s tough to tell to what degree).

Implementing this philosophy successfully should drastically improve portfolio robustness when compared to a single high-Sharpe strategy. Assuming the component alpha’s are “true” (ie. the backtests are not significantly biased / curve fit), there are two major sources of the increased robustness: 1) a portfolio of uncorrelated strategies has less failure points – ex. a few strategies could be “false” or decay and the overall portfolio would still generate alpha, and 2) the lower-Sharpe individual component strategies will generally have less risk of competition eating away at the alpha availability. Assuming the strategies are also “truly” uncorrelated (ie. they trade different markets, timeframes, styles, etc.), the systemic risk of the portfolio is greatly reduced and black swan events should have less impact on overall portfolio returns.

When scaled correctly, this investment philosophy results in what we believe to be the ultimate form of diversification. Of course, its implementation is no doubt easier said than done.

One thing is clear from our perspective: it is impossible to implement this investment philosophy optimally without automation.

This is exactly why XMonetae has built out and will perpetually improve its technology-driven “assembly line” – designed to hypothesize, verify, and ultimately trade automated investment strategies within a dynamic, multi-asset portfolio in the most efficient way possible.

DISCLAIMER