THREAD: Stop Paying for Beta; How to Fin-Dom Beta ▼😳
1\ The Volatility Effect is a well journaled factor return, with empirical evidence suggesting that in US, European, and Japanese markets, low volatility assets in equity markets have higher risk-adjusted returns.
2\ This indicates that investors tend to overpay for risky stocks. Some of the explanations include leverage, behavioural biases and “beta-chasing” funds. For example, we know that in the CAPM model, excess returns may only be obtained via beta of the market net on risk-free.
3\ In many cases, factor returns exist both cross sectionally and in time-series analysis. For example, both cross-sectional and time-series momentum are known effects.
They do not exist on equal planes, however. TS-MOM is tends to be more persistent, particularly in equities.
4\ While literature covers the Vol Effect cross sectionally across equity baskets, our alpha research shows that such idiosyncratic preferences (and overpaying) for beta/volatility exists in time-series data.
5\ That is, equities showing lower volatility relative to its own historical average also have positive expected returns. This alpha exists regardless of the absolute levels of volatility compared to its basket. This is quite effect is quite persistent.
6\ We think that the same reasons in the cross sectionality apply to the time-series effect. In addition, in time-series data, we believe that other factors such as exhaustion of trends/explosive movements (simulating some variant of mean-reversion) play a part in excess profits.
7\ Such excess returns may be harvested by overlaying cross sectional and time series effects on a basket of equities.
8\ FIN. This was from our weekly alpha program at the HangukQuant newsletter. For premium subscribers, see implementation and report:
1\ While I have criticized many facets of applying AI in trading, the appeals fall on deaf ears. If you are going to use them anyway, let’s do it correctly.
A deeper look at one of the most common pitfalls of “AI quants”.
THREAD; The Curse of Dimensionality.
2\ In an earlier thread, I argued the boon of having synthetic data where datasets were sparse, for example in the crypto space. This technique is particularly important when we consider applications of statistical learning algorithms.
3\ The favourite garbage AI model is one when people just spam 10 different features into a neural network to get some price signals. Lets argue why this often fails.
1\ Buffett on hindsight: “a lot of things I wish I’d done in hindsight. But I don’t think much of hindsight generally in terms of investment decisions. You only get paid for what you do.”
Thread: Hindsight Wisdom on Trading; A Treasure Dismissed as Hoax.
2\ One would be remiss to ignore the value of experiential learning and hindsight wisdoms.
In trading, hindsight is dismissed as the most useless wisdom. However, as agents of the world, there is information in all we perceive. Hindsight can make us better after all.
3\ Often, in finance, significant number of problems often have sparse rewards; when training intelligent agents (Reinforcement Learning/Neural Networks), we do not have sufficient data to learn our environment.
THREAD: Stochastic Optimization in Dynamic Environments: Portfolio Allocation by a Quant ▼
1\ Combining alpha signals are an essential part of portfolio management, with extensive literature on integrating alpha. Famous examples include the (Half) Kelly, Markowitz portfolios.
2\ We provide a review of these methods and offer glimpses at our unique, proprietary robust signal-weighting scheme. Let us consider the problem statement and inherent characteristics of dynamic optimisation.
3\ The obvious, and most problematic behaviour is the presence of stochasticity in a dynamic environment.
For an academic treatment of stochastic optimization, a lesson from the Department of Statistics at Columbia.
Why should you care?
Is it going to “break Bitcoin”?
A Trader’s Overview of the Quantum World
2\ You probably have heard of quantum teleportation. No, it is not going to make your girlfriend appear on your bed whilst your parents are sleeping. Whilst it cannot currently destroy your digital coins, it might just be able to do that, sometime soon.
3\ So, what is Quantum Mechanics? It is a fundamental theory in physics that provides a description of the physical properties of nature. The branch of quantum theory dealing with manipulation and computation of quantum bits (qubits) is known as quantum computational theory.
1\13 One of the biggest challenges in quant and alpha research is obtaining clean, error-free data. Models need be built, using assumptions to reduce “dimensions” of reality for tractability.
A THREAD ▼ ▼ ▼
Machine Learning, Sparse Datasets and Error-Free Simulations
2\ Mathematical models, by definition are built to simulate and capture some phenomena, practical or abstract. Often, they are built on data, which themselves are derived from some unknown, statistical distribution.
3\ For example, an alpha model, is backtested upon data where prices/returns are drawn from some distribution. Widely in quantitative finance, they are assumed to be drawn from a log-normal distribution.
Been a fulfilling ~1 month since our launch. For hitting 1k followers, we have a special thread for you, including a premium alpha report and a case study. 🔥
MEGA Thread (N = 60+) : Robust Alpha Research Processes; HangukQuant