Decay of Stat Arb

[Tom Ford]

I've love to get some advice and experience from other more knowledgeable members of the forum.

I've read so much out there that pair trading no longer generate attractive returns anymore. Is this true? Much of the performance has dropped to lower time frames (high frequency) from what I hear. Love to hear if anyone is generating consistent returns in this space.

Thanks,
T

[admin]

Yes, it is true that it is much more difficult than in the past, at least for well-known methods described in the literature. But it still works, if you are smart.

This is why we run our own research to find methods still having some edge.

One thing I can already recommend is to look at pairs traded in a single direction (long only, for instance). Like one direction stat-arb pairs based on one stock and one ETFs. Trading pairs like this is not really described in the literature much.

Example here:
https://www.pairtradinglab.com/backtests/Vdrn5bhwpZRml5vl

This is an arbitrage between Sherwin-Williams Company and an ETF fund. Long positions opened only.

Also what you have read usually applies to US stock market only (they often don't consider ETFs in the literature). There are also many opportunities in other stock markets (LSE, Singapore, Indian stocks, ...). PTL platform will support more markets in future.

If you are a small retail trader, you may want to try pair trading stocks having just enough liquidity to be tradeable by yourself, but illiquid to be traded by big players.

[Tom Ford]

From reading threads, you seem to have been trading live. may I ask, if you don't mind, how you've performed? What's the biggest challenge? It would be awesome to know your live performance stats, ie sharpe.

Thanks a lot

[admin]

Unfortunately not anymore, I personally needed to use the capital to quit my regular job and to dedicate full time to this project (PTL). But I will trade again later (we mostly develop this PTL project for ourselves anyway

I know few pair traders who trade live personally, their performance is something like 1-3% per month on average. They don't use PTL (they use their own proprietary systems) but we share our know-how together.

We have some customers who went live using PTL recently, I will be personally curious about their performance. I might disclose it if they will agree later.

The biggest challenge is to manage the pair portfolio (select good pairs and later decide which pairs you keep and which pairs you replace). This is the key.

PTL project will focus on this issue in the future (we would like to automatize this process as well and make it "backtestable").

[admin]

Just a side note: after more than one year, the arbitrage strategy between SHW and PFF still works somehow (equity not ideal, some drawdown, but apparently still works).

https://www.pairtradinglab.com/analyses/WCpI3NojKWrf3QMf (look at the long-side-only profit analysis)

[sal mhd]

I see a previous post, in regards to your suggestion of

This is an arbitrage between Sherwin-Williams Company and an ETF fund. Long positions opened only.


When you say long only, are you saying both legs long or always long Sherwin. can you please explain. As i am trying to see if there are other new methods .

Thanks

[admin]

Always long Sherwin and short the other one...