Pair Trading Strategy Journal

Quote from cooper1308:

I am not talking about the ratio chart...I am talking about the spread chart.

ie..... ABC minus XYZ

not ABC divided by XYZ
More...

i can't imagine what else you would get out of it if the prices are NOT similar. i mean, if the spread is huge, there could be times where the spread chart is showing a convergence when the actual ratio is going out further from the mean. thus, potential confusion.
 
Yeah exactly what I was thinking...

Just wonder why it is even there, unless I have completely missed something

Quote from Podimer:

i can't imagine what else you would get out of it if the prices are NOT similar. i mean, if the spread is huge, there could be times where the spread chart is showing a convergence when the actual ratio is going out further from the mean. thus, potential confusion.
More...
 
1. quick update on the finviz (visualization) - based paper trade i mentioned earlier. looks like showing a small profit, but will keep it open to see what happens further.

not going to hijack jonnysharp's excellent journal with untested etf ideas, so i'm hereby disappearing from the thread :D good luck, traders! :D

"based on the finviz global etf vizualization alone - i don't have access to statistical tools yet (but i am play with the gummy now) - i have just placed the following paper trade:

BHP -167@ $60.02 (australian stock)
EWA +593@ $16.80 (australia etf)

1:1, total position $20K"

2. a couple of other ideas w etfs i'm make look at now, for paper trading:

XLE short
USO long

XLE short
EWC long (Canada etf)

GDX short (gold producers)
EWC long (Canada etf)

GDX short
DIA long (dow johns)

GLD short
EWC long (Canada etf)
 
I am rid of my most horrid trade - CEPH/VRX.

It came back from a really terrible level to only a 16.47% loss on invested equity.

I'm still +ve with only 35 trades under my belt, so I'm happy thus far.

Hope everyone else is doing well.

Adrian
 
Quote from cipherscribe:

I am rid of my most horrid trade - CEPH/VRX.

It came back from a really terrible level to only a 16.47% loss on invested equity.

I'm still +ve with only 35 trades under my belt, so I'm happy thus far.

Hope everyone else is doing well.

Adrian
More...

WHat went wrong with the trade? I just looked at it and noticed the pair has a pretty low correlation so you could get some big trendy moves.

Despite the low correlation, the pair ratio deviation is screaming go long the pair at 2.5. It looks pretty stretched so perhaps now is a good time to go long.
 
Quote from jonnysharp:

Exited one trade for profits

Sold CLR @ 32.17
Covered ESV @ 40.04

New trade

Long SPG @ 54.51
Short TCO @ 27.95
More...

Quick question Johnnysharp regarding this trade. WHat do you expect to make on it? 5%? How does PTF determine exit points for trades?
 
Quote from cipherscribe:

I am rid of my most horrid trade - CEPH/VRX.

It came back from a really terrible level to only a 16.47% loss on invested equity.

I'm still +ve with only 35 trades under my belt, so I'm happy thus far.

Hope everyone else is doing well.

Adrian
More...

too bad you didn't do CEPH/VRTX
that's a sweet pair
 
Quote from sysre:

too bad you didn't do CEPH/VRTX
that's a sweet pair
More...

Today I spent some more time looking at these trades...ceph-vrx and ceph-vrtx. Why you guys are trading them as a pair is beyond me. They both have very little correlation which generally means once a directional trend begins it will continue for a long time. The other reason also is that you are not getting the hedging affect that correlated pairs provide. You are in essence simply long and short stocks.

I am curious as why you guys entered the trade to begin with. I am curious primarily because it raises the question of trading none correlated pairs and a possible strategy of mean aversion when a pair breaks outside of its bands.
 
Agreed the correlations are low for the pair CEPH/VRTX and CEPH/VRX.

But I did a very rough 10 year backtest using a rolling 2 year z-score (with 2stdev trigger) and CEPH/VRTX actually looked somewhat reasonable with information ratio of around 1.0. It was a crude test without considering T-cost.

The pairing with VRX didn't look as good.

Now, I know high correlation implies that the reversion tendency is good, but I cannot explain why a low correlated pair can sometimes work for a reversion strategy.

A question to ponder: if high correlation is a requirement for a good pair, then why is it difficult to pair trade spy/dia (correlated 94%)
 
You must log in or register to reply here.
Back
Top