Pair Trading Strategy Journal

Quote from jonnysharp:
01/12/2009

2 new trades;

Long BAS @ 10.80
Short HAL @ 17.93

Long ETE @ 17.47
Short MGG @ 16.16
More...

BAS/HAL ended up being one of your largest losses. I've studied the chart you attached to the January post, and studied my graphs of the pair, I've checked the news for the pairs back in January, and I cannot see why the trade should not have been taken. (Except that historically, this pair has returned less than $85/trade).

One item, on January 13, BAS reported "selected operating data for Dec 2008." The stock price did not seem to be affected, yet from then on BAS - HAL divergence grew. I lost big on one paper trade because I did not check when earnings were to be reported and the report the day after I entered the trade caused 1/2 of the pair to go bad. Could this be what happened?

Have you any insights in retrospect? Would you take that same trade again?

WB
 
Quote from waltbx:

BAS/HAL ended up being one of your largest losses. I've studied the chart you attached to the January post, and studied my graphs of the pair, I've checked the news for the pairs back in January, and I cannot see why the trade should not have been taken. (Except that historically, this pair has returned less than $85/trade).

One item, on January 13, BAS reported "selected operating data for Dec 2008." The stock price did not seem to be affected, yet from then on BAS - HAL divergence grew. I lost big on one paper trade because I did not check when earnings were to be reported and the report the day after I entered the trade caused 1/2 of the pair to go bad. Could this be what happened?

Have you any insights in retrospect? Would you take that same trade again?

WB
More...

Yes this was a ugly trade I know, I have backtested this pair and have it as producing as avg $337 profit per trade not $85. yes the day after I opened the trade BAS released a statement saying earnings were likely to be down more than forecast and they would be lowering their pricing, anytime a company lowers its pricing its a red flag, im thinking about implementing a 10day time stop for my trades, im still testing this idea, this would have significantly reduced the loss on this trade, however like you say the divergence was persistent and prolonged, ive attached the stdev chart and you can see that for a pair that regulary oscillates it stayed below its mean for sometime before returning. Ive discovered there is correlation between trade length and profitibility, the best trades are the shortest in length and the worst trades are the longest, im trying to incorprate this fact into my trading plan.
 

Attachments

Quote from total_keops:

"Paulson Buys AngloGold Stake"
Thanks man, I appreciate.
More...

Paulson buying in is probably what caused the divergence, it may pop some more wednesday, but could be shortlived as it will mostly be daytraders unless he keeps buying. did you currency hedge this trade?
 
Quote from jonnysharp:

Paulson buying in is probably what caused the divergence, it may pop some more wednesday, but could be shortlived as it will mostly be daytraders unless he keeps buying. did you currency hedge this trade?
More...
No, I did not expect it to last that long. I also had a view on the currency exchange that was on my side.
 
Quote from jonnysharp:

Ive discovered there is correlation between trade length and profitibility, the best trades are the shortest in length and the worst trades are the longest, im trying to incorprate this fact into my trading plan.
More...


Agreed. I've noticed the same thing. I've made the best profits on short length trades and I've closed trades early that showed a decent profit without waiting for an exit signal. I have yet to regret doing that as the longer the trade went after I booked the profit, the less the profit turned out to be.

I have started looking at pairs that are in trade and looking for the turn up from the bottom of the ratio with RSI involved, but haven't taken any. I think this subject is worth pursuing further.

Thanks,

Lazy
 
Lazy, yes if you downloaded the latest update of pt finder, they have a +/- column in the console and chart in the analysis window which you can use to scale out before an exit signal, like if you enter a pair at -2.20 stdevs, you could exit when it goes above -1.00, trade length reduces dramatically, also win rate would pickup too, avg win would probably decrease a little, something im looking into.

Yeah ive looked at similar things for pairs currently in trade, i decided to steer away from those because like we have noticed trades that don't work out straight away tend to not be the most profitable, I would take notice if a pair currently in trade went 3+ stdevs from the mean on non-specific stock news, that would be a awesome trade setup I think.
 
exited trade

Sold ATN @ 10.78
Covered RIG @ 59.30

new trade entered on wednesday

Long PUK @ 7.63
Short AXA @ 11.72

I was reading a research paper on pairs trading and it stated that pairs that diverged on industry news as opposed to stock specific news tended to converge much quicker. The theory behind this was the lead-lag affect between two similar stocks, each stock may have different institutional holders that may digest & act on information differently, plus liquidity constraints may not allow one stock to be priced in the same timely manner as another, whereas stock specific news that heavily affects the stock tends to have a more persistent divergence as a fundamental re-pricing may be warranted.
 

Attachments

You must log in or register to reply here.
Back
Top