How Curve Fitting Could Impact Your 2023 US Stock Market Investing

by Plotinus

The sudden recent implosion of FTX provides an opportunity for investors to explore the importance of the role of human narrative in investment decision-making.

As the saga evolved—saga being a very appropriate term as the story is shaped by the retelling of it—the magical hero to villain transformation of the company was, as always in these types of situations, instantaneous. With the unacknowledged benefits of hindsight quickly and seamlessly being absorbed and processed, the uninterrupted flow of words in the financial press and general media more broadly, trampolined smoothly from glowing praise to damning criticism. The problem with this “tracking” of investments made in FTX is that there is no pause for thought, no recognition that a very familiar pattern was playing out once again. This problem is not about the specific, main triggers for FTX’s downfall, but rather the investors’ self-deception involved in hearing and taking in the company’s story over time as though it were one singular, sensible flow.

For investors who are unaffected, they’ve made the right call, knew it all the time, and are now vindicated in the decisions they made to steer clear. Yet for those investors hit with losses, it is time for hand wringing and regrets as to how were they so blind to what was clearly, patently wrong. Neither of these positions reflect reality but both can be neatly molded to fit into the story.

How have you handled the a priori and a posteriori investor knowledge challenge?

The FTX debacle is another investment example of the confusion of a priori and a posteriori knowledge. In other words, a failure for the investor to maintain a consistency in delineating deduction and observation in the presence of new information.

This is a problem for anyone involved investing. The deductions an investor makes on, say, a stock on which he or she conducts a due diligence analysis is the theoretical curve that fits an observed data set; be that quantitative analytics data only, fundamental data only or a combination of these. The balance challenge for the investor is to be able to draw reasonable conclusions from the information at hand to date about what may be expected while acknowledging that future data is unknown and may defy the theoretical curve. The objective is to avoid the pitfalls of over-fitting or under-fitting the curve.

This is a familiar problem for any investor involved in AI-based investing.

Is your US equity investing possibly over- or under-fitting the curve?

Points we made above relating to the current tale of woe in crypto investing equally relate to the most common, core holding you likely have in your investment portfolio: its exposure to the US stock market.

As we have mentioned on previous occasions, we at Plotinus have opted for using an AI derived data approach for giving our investors exposure to the US stock market. A primary rational for doing so in our statistically based trade decision-making strategy has been to have a translucent rather than black box; both for ourselves as the portfolio manager and developer of our proprietary model, and for our investors. By opting to restrict the amount of data available to the system we reduced its potential to go off chasing the curve. Furthermore, by structuring the quantitative strategy to have a translucent box methodology it offers the potential to understand why the system is coming to the decisions it is making; in particular, those trading decisions that end up being incorrect. Being able to observe and understand why our AI-based model is making decisions allows us to build and refine a tolerance for error. This, in turn, forms part of the broader learning process of our approach to employing an AI trade decision-making system, allowing it to become progressively more robust.

How might an investor reconcile the empirical of what has happened with what goes on to take place? Statistically based decision-making is not a guarantee of correctness. It is, however, a method on which to base decisions, which is grounded (hopefully) in empirical measurement. In simple terms, the purpose of statistically based decision-making is to provide a logical basis for taking action A rather than action B. For this, it is necessary to build a tolerance for error, accepting that a statistically based trade decision-making system will—and should—get things wrong at times. (Just like any other investment decision-making process, for that matter.) What would be of more concern to the savvy investor, however, would be any investment system or strategy that appears to be very highly accurate, as this is akin to FTX whilst in hero mode. This is when one should be wary, when one should be left wondering when it will go off? And, when it does, how bad will that be?

How will you divvy up your US stock market exposures for 2023?

As we enter into the final month of the year, investors’ attention is being drawn to 2023. US stock market exposure is subject to a wide array of uncertainty. Morgan Stanley’s chief equity strategist, for instance, recently suggested that the S&P 500 could bottom out in the first quarter of next year at 3000, 25% lower than current levels. That target number view aside, there is question enough being generated by domestic factors such as US inflation and recessionary fears as well as what the potential effects on the US markets might be based on what is happening elsewhere in the world.

This is made all the more difficult by there being a range of situations that could play out, and with potentially different consequences. The conditions in the Eurozone and the UK are not looking very bright. There is geopolitical instability, stubbornly high inflation and recessions that are beginning or have already begun. All of these factors have the potential to affect the US stock market, whether for the better by helping attract European capital to it or for the worse by hindering it via weaker demand for US products. This is the kind of dilemma which —depending on the angle from which it is perceived—could have the likelihood of both scenarios explained through previous empirical data. Add to that storyteller’s flair and then try assessing the influence the narrative at any given moment may have on the ups and downs of your US stock market exposure.

For many investors their US stock market exposure is a core holding, regardless of how that may be allocated, actively or passively. So, the concerns for such long-term investors regarding the US markets are not necessarily about if and when to reallocate US stock market allocated money into some other asset class. It is more so a question of how might the investor seek to buffer some of the downside market volatility—in the event of any of the bad case scenarios playing out—while keeping a full US stock market exposure in play.

Diversification by method: a new option for a 2023 US stock market allocation

In our option, adding an AI trade decision-making strategy to a core US stock market allocation (be that in actively managed or passive index form) provides a diversification by method, as it is a new and different trade decision-making approach. This statistically based trade decision-making investment process provides a novel and differentiated take on diversification, which normally implies an allocation away from the core investment order to protect it, but here remains allocated to that very core element: the S&P 500.

A useful AI trade decision-making strategy has the potential to be a tool for achieving the long-term investment goal of achieving better risk-managed returns. As the foundation of the approach is to generate more statistically relevant trade signals (separating signals from noise), it has the benefit of being detached from the vagaries of human narrative. Statistical analytics decisions are not stories, and if such an AI based investing model has a translucent structure, it can provide explanation for its trading outcomes both good and bad.

If deployed as a component of an investor’s total US stock market exposure, it has the potential to deliver a different pathway in pursuit of an improved risk-adjusted return exposure to the US stock market.

A New Year’s Resolution action we propose the savvy investor consider is to aim to avoid the pitfalls of letting their investment strategies fall prey to over-fitting or under-fitting the curve.

© 2022 Plotinus Asset Management. All rights reserved.
Unauthorized use and/or duplication of any material on this site without written permission is prohibited.

Image Credit: Max776 at Can Stock Photo.