Seeking Advice: Has Pairs Trading Become an Overcrowded Red Ocean?

Quant Trading Strategies & Ideas
1

We are a small quantitative team of 1-2 people, each with over 10 years of software development experience.

This letter is a reflection on our experiences in 2025, and a sincere request for your professional insight on a core challenge we face.

:memo: Background: Our 2025 Journey ā€“ Statistical Arbitrage

Method: ADF Test + correlation test + hurst + half life + RSI + zscore
Hour level data, day level trade

Our direction for much of the year was rooted in a ā€œcontrarian investment, focus on arbitrageā€ philosophy, specifically employing Pairs Trading as the primary arbitrage method.

We pursued this in three main areas:

  1. US Stock Pairs Trading: We found that factors like earnings reports often caused cointegration relationships between individual US stocks to break down frequently.

  2. US ETF Pairs Trading: The cointegration and correlation here proved very stable. However, the annualized returns were in the single digits, making it less appealing than a simple buy-and-hold strategy on the S&P 500 (SPX). Historically, a simple dollar-cost averaging into SPY has yielded around 10% annualized returns with only a few hours of backtesting effort.

  3. Crypto Perpetual Contract Pairs Trading: We observed that their cointegration was unstable and frequently broke down.

Across these three attempts, we invested over half a year and concluded: We used the correct methods to pursue the wrong things.

:bulb: Key Takeaway & Core Challenge

Quantitative trading is inherently a research-based, trial-and-error process. However, each potential direction typically requires 1 to 1.5 months of trial and validation. For a small team like ours, this becomes a financially unsustainable cost.

Our core challenge stems from a realization: Sometimes, knowing what not to do is more important than knowing what to do.

The underlying reason for this is that some theoretical directions, while sound, are simply untenable for a resource-constrained small team. Our competitors possess extreme hardware advantages and elite teams. Given these overwhelming resources, our small teamā€™s effort, no matter how great, is unlikely to generate a significant Return on Investment (ROI).

:pray: Seeking Your Advice

We would be genuinely grateful if you could share your seasoned perspective on the following:

Question 1: The Pairs Trading Landscape

  • Are there market participants who are consistently profitable in the Pairs Trading space today?

  • Or, has this field become so highly crowded that low returns and losses are the norm?

Question 2: Strategy for Small Teams

  • What are your thoughts on the specific challenges we are currently facing as a small, resource-limited team?

  • We welcome any candid critique or constructive suggestions you might have for us.

Thank you for taking the time to read our reflection. We look forward to hearing your valuable insights.

2

Answer to your Ques1
Yes, some participants are still consistently profitable, but mostly at scale and with strong infra (execution, borrow, costs, portfolio construction). Pairs trading is no more an easy alpha strategy like in early 2000s.

Yes the space is crowded which tends to compress returns, and profits often hinge on smarter pair selection, regime filtering, and cost control.

Answer to Ques 2
Yes might be with such small team, you may take more time in research, in that case the opportunity might get missed.
Our suggestion -

  1. Big firms dominate popular markets/strategies. Look for smaller, less crowded areas, where they donā€™t bother or canā€™t move fast.
  2. Set clear ā€œpass/failā€ rules early. If an idea doesnā€™t hit basic targets in a few weeks, drop it and move on.
  3. Many small edges beat one big bet. Pairs trading works better as lots of tiny trades together, not trying to find one magic pair.

Again, these are not core facts, just our perception based on industry patterns.

3

Thanks for your response, Sanya.
And here are some update for the data for Forex
Per experience, the AUD and NZD should be correlated and cointegrated.
However, There are not, per the calculation result in the hour level, datasource: exness

7 months work been proven to be meanless, frustrated. Donā€™t know why.

4

Thanks for sharing the chart. Even for ā€˜intuitively linkedā€™ pairs like AUD/NZD, relationships can temporarily decouple due to macro/regime factors (rate differentials, China exposure, risk-on/off, commodity shocks, etc.), so this could simply be a decoupling phase right now. Iā€™d treat the hour-level result as a signal, not a hard fact. Might be worth checking past decoupling episodes: when they started, what macro backdrop drove them, and how/when the pair re-coupled. That historical pattern could itself become a useful filter or feature.