The Unreasonable Effectiveness of Monte Carlo Simulations in A/B Testing

Marton Trencseni - Sun 20 October 2024 • Tagged with ab-testing

The article illustrates how Monte Carlo simulations serve as a powerful method for exploring statistical concepts in A/B testing, enhancing understanding, improving experimental design and analysis.

Monte Carlo casino

Continue reading

A simple telnet based Taboo game in async Python

Marton Trencseni - Sun 06 October 2024 • Tagged with taboo, python, async

I built a simple Python-based Taboo game server using asyncio to provide a customizable multiplayer experience via Telnet.

100

Continue reading

AI Strategy in 2024 for the Rest of Us

Marton Trencseni - Fri 04 October 2024 • Tagged with meta

My review of the AI landscape in late 2024, and what the road ahead looks like for those of us not building foundational models.

100

Continue reading

Techtalk: Why are A/B tests the gold standard of causal inference?

Marton Trencseni - Sun 29 September 2024 • Tagged with ab-testing, techtalk, kohavi

Recently, I delivered a techtalk on A/B testing to an audience of non-technical Product Managers and experienced Data Scientists.

100

Continue reading

200 articles

Marton Trencseni - Sat 28 September 2024 • Tagged with meta

A review and introspect on the second 100 articles written on Bytepawn.

100

Continue reading

ChatGPT's critique of OpenAI CEO Sam Altman's vision of the AI future

Marton Trencseni - Fri 27 September 2024 • Tagged with openai, chatgpt, sam, altman

A few days ago OpenAI CEO Sam Altman published an article titled The Intelligence Age. I used OpenAI's new o1-mini model to critique Sam Altman's writing of its own future.

Continue reading

How I manage my 1v1s

Marton Trencseni - Sun 22 September 2024 • Tagged with people, management

I detail my structured approach to managing one-on-one meetings within a 40-person data team, emphasizing people management principles like radical transparency, tailored meeting cadences, and strategies to navigate common managerial challenges to foster a supportive and productive work environment.

Meeting distribution

Continue reading

Breaking Bell's Inequality with Monte Carlo Simulations in Python

Marton Trencseni - Sun 01 September 2024 • Tagged with physics, quantum, mechanics, qubit, entanglement, epr, bell

The article explains the Bell inequality using Monte Carlo simulations in Python, and shows how non-local action-at-a-distance can be used to break it with entangled qubits.

John Stewart Bell

Continue reading

Quantum Entanglement, Locality, Determinism

Marton Trencseni - Sun 18 August 2024 • Tagged with physics, quantum, mechanics, qubit, entanglement, epr, bell

This blog post explores foundational quantum mechanics concepts such as superposition, entanglement, and the challenges posed by locality and determinism, while highlighting the historical efforts of Einstein and Bohm to reconcile these phenomena with hidden variable theories.

Double slit experiment

Continue reading

Why I don't live in Hungary

Marton Trencseni - Wed 07 August 2024 • Tagged with hungary, politics, orban, dubai

In February 2016, I made a life-changing decision to leave Hungary and seek opportunities abroad, starting in London and later moving to Dubai. This essay explores the multitude of reasons behind my choice and why I have no immediate plans to return to Hungary.

Orban no migration

Continue reading

How to solve it

Marton Trencseni - Thu 01 August 2024 • Tagged with polya, problem, solving

I discuss Polya's How to Solve it approach to mathematical problem solving and the 2010 paper Teaching general problem-solving skills is not a substitute for teaching math.

How to solve it

Continue reading

How I screen CVs

Marton Trencseni - Mon 01 July 2024 • Tagged with cv, screen, hiring

Screening CVs is a critical step in the hiring process, used to determine which candidates progress further in the hiring funnel. While the approach to CV screening can vary by company and industry, the primary goal remains the same: to identify candidates with the potential to succeed in subsequent interview stages.

.

Continue reading

Writing a simple Javascript async message queue server - Part III

Marton Trencseni - Mon 17 June 2024 • Tagged with javascript, async, message, queue

In this final post on the toy Javascript async message queue server implementation, I make it compatible with the Python version. I use the library of unit tests developed previously to ensure identical behaviour between the two codebases.

.

Continue reading

Writing unit tests for the async message queue server

Marton Trencseni - Sat 15 June 2024 • Tagged with python, async, message, queue, unit, test

I write a small library of unit tests for the message queue servers.

.

Continue reading

Testing randomness extractors

Marton Trencseni - Sun 02 June 2024 • Tagged with randomness, extractors, biased, coin, fair

I apply NIST's suite of statistical tests to my randomness extractor implementations from the previous posts.

Coin flip

Continue reading

Randomness extractors: making fair coins out of biased coins

Marton Trencseni - Sun 26 May 2024 • Tagged with randomness, extractors, biased, coin, fair

I discuss various models of biased bit sequences, and how to extract uniform random (or close to it) output bit sequences from them, illustrated with Python code.

Coin flip

Continue reading

Writing a simple Python async message queue server - Part III

Marton Trencseni - Wed 22 May 2024 • Tagged with python, async, message, queue

In this final post on the toy Python async message queue server implementation, I make further feature and code improvements.

.

Continue reading

Writing a simple Javascript async message queue server - Part II

Marton Trencseni - Sat 11 May 2024 • Tagged with javascript, async, message, queue

I write a somewhat more complicated, but still relatively simple async message queue server in Javascript.

.

Continue reading

Writing a simple Javascript async message queue server - Part I

Marton Trencseni - Sun 05 May 2024 • Tagged with javascript, async, message, queue

I write a simple, bi-directional async message queue server in Javascript.

.

Continue reading

Paper review: A Comparison of Approaches to Advertising Measurement

Marton Trencseni - Wed 01 May 2024 • Tagged with ab-testing, facebook, stratification, propensity

Why are Randomized Controlled Trials (RCTs, known as A/B testing in much of the industry) testing is widely regarded as the golden standard of causal inference? What else can a Data Scientist do if A/B testing is not possible, and why are those alternatives inferior to A/B testing?

This papers shows, using 15 experiments (for ads on Facebook) where a RCT was conducted, that common observational methods (run on the Facebook data, by ignoring the control group) severely mis-estimate the true treatment life (as measured by the RCT), often by a factor of 3x or more. This is true, even though Facebook has (i) very large sample sizes, and, (ii) very high quality data (per-user feature vector) about its users which are used in the observational methods. This should be a major red flag for Data Scientists working on common marketing measurements (such as marketing campaigns) using observational methods.

Continue reading