Author: shyam

Financial News Sentiment Analysis: Qwen vs. Jev

Ever since Typesafe’s Jev came out, people have been raving about its “one-shot” classification capability. We decided to put our news archives through Qwen (a “regular” llm) and Jev (a classifier) to bucket the sentiment expressed in each of them into Positive, Neutral and Negative.

Is Jev lazy? Picking “Neutral” is the path of least resistance, isn’t it? What if we split it by media type?

It looks like Jev loves being neutral. Either Indian financial media is mostly unbiased or Jev is not as “one-shot” as claimed to be. Thanks to Jev’s neutrality, the two models agree only 43% of the time.

Given their differences, are they useful?

There is no evidence that extreme media sentiment reliably predicts subsequent index direction. The positive post-event paths may simply reflect the general upward drift of the indices.

It appears that sentiment is in the eye of the analyzer and largely useless in predicting low frequency market direction. Details about the methodology and results are here.

Code and charts are on github.

Categories Tags

Filtering and Momentum Signals

The paper Shenhao Zhang, A Low-Frequency Quantitative Trading Strategy Based on Trend Filtering and Momentum Signals: Empirical Evidence from A-Share Banking Stocks, published at ICDEIT 2025, is a small, low-frequency technical-strategy study on two Chinese bank stocks. We wanted to check if it generalizes to a larger universe of Indian and US stocks.

tl;dr: it doesn’t (summary).

For Indian stocks, you may want to give this a wide berth. From 2016 through now, the model barely beat a fixed deposit.

However, by the virtue of it side-stepping the 2008 drawdown, it managed to put a decent number on the board for US stocks.

We tested this out because the premise seemed promising. Sadly, it turned out to be another one of those narrow, overfitted papers with an SEO friendly title.

Code, charts, findings are on github.

Categories Tags ,

SEBI Surveillance Measures

SEBI introduced the Additional Surveillance Measures (ASM) and Graded Surveillance Measures (GSM) Framework in 2018. They serve as “early warning systems” to protect investors. ASM focuses on price volatility, while GSM focuses on weak corporate fundamentals.

We have been tracking these since they were introduced. They are helpful “no-go” zones while constructing portfolios primarily because the additional margin requirements with narrow price bands and the corresponding fall in volumes makes risk-management impossible.

The question is, can this black list be turned in to a white list? Is there alpha in trading the transitions of stocks in & out of this list?

Could there be an effect that can be exploited?

Nope.

You can read a more detailed introduction to surveillance and research results here.

Code and charts are on github.

Categories Tags

Turbulence in Financial Markets

Can the concept of turbulence in physics be applied to financial markets? It could be. We did a small test and it does appear to be helpful (summary).

There are some instruments for which it works in some regimes…

… and fail in trending markets.

The biggest problem with the entire approach was the sheer number of parameters that needed to be specified. There are no “right” ones and you will almost always overfit.

And walkforwards fell apart.

Its a promising approach to be revisited once we have more material to work with.

Code, charts and summaries on github.

Categories Tags

Book Review: The Trend Following Mindset

The book Trend Following Mindset: The Genius of Legendary Trader Tom Basso by Michael Covel (Amazon,) is not really about trend following at all. It doesn’t give away any techniques or road maps.

There was one outline of a strategy in the book that we put to test and nothing much shook out of it (summary).

It has some general ideas about investing, risk management and the nature of markets that a new investor might find useful.

Categories Tags ,