ASX Stock Pairs Trade Journal

[Pairs Trader]

Pairs Trader,

Good Morning from Atlanta. I have a question. Is it reasonable for a trader to trade this method on an end of day basis, placing orders at night to be filled at the open? Or, do they need to be placed intraday?

Yes many traders use it on an EOD basis, you just gotta make sure your orders filled the next day are still within reason for taking the trade. Its more than possible to take trades based on signals received yesterday, I do it all the time, a pair might have generated a signal yesterday and the pair continued to divergence from the mean allowing for a better entry.

 

[Pairs Trader]

Hey pairs

First time ive looked at this thread

Long LGL @ 3.23
Short SGX @ 6.03

As a long term Gold watcher and investor i can see where this combination
comes from...also can see how the short could go very very wrong.

By backing 2 stocks arnt u doubling the risk...u need 2 directions to go your
way...i find getting one stock to move my way hard enough.

GL

I would be interested to hear how it could go very wrong. You may have a fundamental bias and thats fine, I am however trading the technicals looking for a quick snapback within a day or two, taking advantage of share price volatility, fundamentals only apply if your holding time is greater than a month imo. Pair trading isn't double the risk, its actually the opposite, because it negates market and sector risk and you are only betting on the relationship of two correlated shares, and typically you make more money on one side than you lose on the other, so you don't need 2 directions to go your way.

 

[cooper1308]

Hey pairs

First time ive looked at this thread

Long LGL @ 3.23
Short SGX @ 6.03

As a long term Gold watcher and investor i can see where this combination
comes from...also can see how the short could go very very wrong.

By backing 2 stocks arnt u doubling the risk...u need 2 directions to go your
way...i find getting one stock to move my way hard enough.

GL

An outright long in my opinion would carry FAR more risk, both on a fundamental and trade pysche level...

You don't need 2 trades to go your way, infact this is very rare... Out of the 24 profitable trades tell me how many times have both positions moved in favour?? once, twice??

All that matters is one position more than offsets the other. And 90% of the time this will happen

 

[Truth]

Will a future version of PairTrade Finder have a direct connection to a broker or a partnership with say, NinjaTrader, so that it could do the following?:

1) When a pair from your list pops up, and you decide to trade them, you just click one button and the software would auto position size the pair and send the order to the broker.

2) When the pair got back to within 1 std dev (or whatever setting), an order is auto sent to broker to exit.

 

[skc]

There is a real casino - roulette element to that, the mathematics and hedging.

I suppose success is super dependent on the entry timing and stock selection.

How's this not the case for any other form of trading?

I would be interested to hear how it could go very wrong. You may have a fundamental bias and thats fine, I am however trading the technicals looking for a quick snapback within a day or two, taking advantage of share price volatility, fundamentals only apply if your holding time is greater than a month imo. Pair trading isn't double the risk, its actually the opposite, because it negates market and sector risk and you are only betting on the relationship of two correlated shares, and typically you make more money on one side than you lose on the other, so you don't need 2 directions to go your way.

The worst thing that can happen is probably if LGL launches a takeover bid for SGX. That would be a pair traders worst nightmare. But even that, say a 30% takeover premium with a 10% fall in LGL, you take a 40% hit to your positions. If you only risk 25% of your capital for any one trade, then that's a 10% hit to the capital. Bad but not catastrophic in the long run if you win rate is high enough. Some pair traders recommend avoid shorting the smaller stock to prevent this.

I tried to trade a gold pair on Friday as well - short SGX and long KCN. SGX jumped >5% in the morning, setting off all sort of signals against other gold sector stocks, whereas KCN hardly moved as all. So I opened the long side first, and have an order for the short ready to fire. I ended up closing the long after KCN caught up with SGX on an intraday basis, and never needed to open the short... I think this is called leaning on the long (or something like that).

 

[Pairs Trader]

Will a future version of PairTrade Finder have a direct connection to a broker or a partnership with say, NinjaTrader, so that it could do the following?:

1) When a pair from your list pops up, and you decide to trade them, you just click one button and the software would auto position size the pair and send the order to the broker.

2) When the pair got back to within 1 std dev (or whatever setting), an order is auto sent to broker to exit.

Yes we will be having those capabilities in the near future.

 

[Truth]

Yes we will be having those capabilities in the near future.

Excellent.

 

[Truth]

I'm checking out a current thread on EliteTrader by Jonnysharp, whose using your program, and there's a quote by "forsalenyc" that sort of bothers me:

" Johnny, as a fan of your thread, I want to ask you if you're convinced that Pairtrading is the way to go, or vice versa? I still pairtrade occasionally, but I try to keep it intraday.

In one of the posts, you quoted that you're down so far using this strategy. well, when I actively pairtraded, I was 39-0. and when it became 39-2, it pretty much wiped out all my gains. I reckon its the case with you.......running the losing pairs. "

What's your opinion on this? Is this just a case of someone using horrible position sizing judgement and putting too much of his account on one or two trades? Or is this something that has a decent probability of occurring since stops are not used with this system. I know there's not a whole lot of info there, but maybe you could still comment.

http://www.elitetrader.com/vb/showthread.php?threadid=134253&perpage=6&pagenumber=42

 

[Pairs Trader]

Yes that may have been that guys experience, and I can see how it can happen, by simply putting to much on your account in one trade, now what I would call a black swan event of one of your stocks opening 50% higher or lower than previous close if you only have 20% of your a/c in that stock your overall a/c would be down 10% which is well within acceptable drawdown levels and you can mitigate the chance of this occurring by not trading ''hot'' stocks like BNB was after **** hit the ceiling or not trading stocks just before earnings release or stock with pending news, like drug stocks awaiting government approval, stick to the well know, liquid big cap stocks and this chance of this occurring is very minimal, once again you can blowup your a/c using the most successful trading system in the world if you have the wrong risk management, anyone that does blow an a/c was over leveraged with concentrated exposure plain and simple.

 

[Edward deVeres]

Dear PT:

I've owned your software for the past month, during which I've papertraded avidly pairs of stocks in the US. I've made a couple of bad choices in that time, which were a great learning experience, and cheaper than cash! Every week that I've put off actually trading I find I've learned another bunch of lessons that make me thankful I'm waiting.

On the other hand, even with the 2 big losses, (I didn't pay enough attention to the charts) I'm ahead about 3% on 23 trades, with closed trades of 12, all with good profits, and (as usual) the open trades at much more of a loss--however, none of them are over 4 days old, so I'm hoping.

I've learned a huge amount from this Forum. No other software developer has had the courage to trade his system in real time on the Net. This puts you light-years ahead of almost all others. It's a great sales tool as well as great teaching device.

A few questions, none to challenge you, but for understanding, as I've read this thread 3 times struggling to gain an insight into how you are able to read so much into these charts.

(1) Since you started, you have changed the spread chart to the RSI Ratio chart & back to the spread chart, and I wondered if you are still tinkering with your own understanding of which 4 charts to put up? Despite years of eyeballing such RSI charts, I admit I still can't make much of divergencies on the RSI chart--I find divergencies to be in the eye of the beholder...so any gain your experienced eye may behold may be due to your experience. But, have you given up on the RSI chart in favor of the 50-say spread chart? Also I notice that sometimes you seem to ignore the "fade-the-50-day" spread rule? Any comment?

(2) A few days/weeks ago, you changed the Plus-Minus Chart from 365 days to 150. I initially put mine at 91 days and counted the movements from plus-minus 1.50 to gain a sense of how much oscillation the pair had and kept track of that. Why did you change from 365 to 150? Have you now settled on 150, or is all this viewing and settings of the charts a work in progress?

(3) At the same time, you changed the %-from-Mean chart from 365 to 150days. Same comment and same question--are all these chart settings a work in progress? I'm going to imitate your charts on my computer so I will be looking at the same view you have in an effort to gain the understanding you have.

(4) Maybe because I'm new at this, but I'd love to see in PTF not one page of 4 charts, but 2 pages of 4 charts, so that I can FLIP from one set to another as I'm studying a pair. I like about 6 of the 8 available charts, so I have to make an agonizing choice of what to leave out when I set up my software.

I've found that the NWS trades here in the States, and I'm hoping to trade it too.

I see you're preparing to be able to trade directly out of PTF in Ver. 3.0...I hope this will connect to Interactive Brokers here in the States, as they seem to be the best for pair trading. A trader can obtain more margin from them with a balanced portfolio than from other brokers. Without committing you to the event, when do you think Ver. 3.0 will be ready? Weeks? Months?

And lastly, a surprising request: I think you should charge for updates. The work you and your programmer have put into PTF is huge, that is evident, and exceptionally well-thought-out. I for one am willing to pay more for the continued development of the product. I'd hate for you to find that you let the development of the program languish for lack of purchasers' dollars. I think the program is worth several thousand dollars--although I'll grant it is hard to sell that to traders right now. For reasons that escape me, the average trader is obsessed with directionality. However, if you continue to improve it, in time it will become a tremendous success. That might mean that its value will diminish (paradoxically) as more people use it, but on the other hand, there are so many thousands of pairs world-wide that that's unlikely.

Lastly, you might think about 2 versions: a PTF Light, and a PTF Pro. Those who want to try the product could buy in at $500, and those who want the advanced features, pay quite a bit more. Perhaps the Light version would only do 50 pairs in the Console, and only 4 charts, whereas the Pro version would be unlimited, with 2 pages of 4 charts each. The new Port. Manager might only carry 10 pairs, while the Pro version could be unlimited. I suppose there could be several other differences, but I can't think of them just now.

Many thanks again,

Yours, Edward deVeres.

 

[Pairs Trader]

Yes some consider it a risk making real time trades in front of everyone because it might not work, or I may hit a losing streak, or plain doesn't work, however I have a high level of conviction in the software, plus with my experience I wasn't worried at all starting this journal and full confidence in being profitable and crushing the market averages, it is quite doable contrary to popular belief, people forget that the benchmark ASX200 index that so many people ''invest'' in is a trading system in itself, comprised of a market weighting mechanism of blue chips, rotating into and out of stocks with the index adds/deletes, and its long only, without a stop loss or profit target, who says this is the best trading system or the best way to invest? look at the drawdown's such a system gives, where is the competitive edge if everyone is doing it? who says you can't do better in terms of return and drawdown than the ASX200 trading system?

This is how I filter my trade signals, first I have an optimal universe of pairs on my radar from backtesting, a have diverse lists of pairs from all reaches of the trading world, I only post my ASX and index pair trades here.

Then I have audio alerts for entry and exit signals enabled, when I get a alert the first thing I look at is the pairs correlation, I want it to be above 70%, then I analyse the pair, I want to see a non-trending ratio chart that displays a regular oscillating pattern in that the pair diverges and converges in a repeating pattern, then I what to see the % from the mean in the top quarter of its historical extreme, so if I pair regularly diverts from its mean around 5%, I would be reluctant to take a signal at 2%, I may wait for it to diverge more if it does at all, then Il do the same thing with the +/- chart but less strict, ideally the more extreme the +/- the better especially if there is no price sensitive news just been released that could explain it, for eg....if you had a pair with a 3stdev divergence on no news released, that is an anomaly and should be exploited. Then finally I like to fade the 50day spread chart, if I have a signal too go long the cheap share and short the expensive share, Its a bearish spread signal, we are betting the spread between the two stocks is going to decrease, in this scenario I want to see the 50 day spread chart at new 50day highs or near abouts, vice versa if we are bullish the spread. If a signal passes these tests, about 1 in 5 do for me, takes me about 30 seconds or less to determine if il take the trade or not, all thats left is to execute the trade.

Once you get a thorough understanding of pair trading and why it works these filters will come naturally, Im not saying my filters are the best or there is even a best set of filters, pair trading(or trading in general) is part science, part art, there are no hard and fast rules.

Thanks for the kind words and feedback, its refreshing to hear a client that's willing to pay more for the product, you obviously know the value of quality, we have spent 1000s of hours to arrive where we are at now, in terms of design, structure, framework, feasibility studies, programming, testing, re-design, more testing, debugging, adapting, documentation, support and implementation, and we have a combined 26 years of programming/trading experience, its been no small feat and I can assure you we have some more big developments in the pipeline. All the features have been built based on client feedback/requests and quite a few clients have co-sponsored the cost of developing those features. Its always a work in progress of continual improvement.

 

[skyQuake]

Hi Pairstrader, with regards to a trade, how would you take into account technicals and overall trend? For example SGX has formed a nice breakout at the 5.70s while LGL is chopping around.

 

[Pairs Trader]

I don't look at the tech/a of individual stocks when analysing a pair, plus you say there was a breakout of SGX, well if you program that breakout and try and see if there is any predictable pattern after that breakout that would influence your decision in taking a trade you would find out that price breakouts have absolutely no statistical predictive measures in forecasting price, doesn't mean you can't profit from breakouts(you can marginally with the right setup), just means that because a price breaks a 20, 50, 100 day high/low doesn't mean price will follow through on subsequent days, actually the opposite is more true, when I traded directionally I use to fade price breakouts on an intra-day scale.

The largest profit potential is when the largest amount of people are wrong, you can't make money by following what everyone else is doing, it amazes me how many people trade breakouts or think that is the way to profitably trade, now we can argue all day long about this or we can look at the facts, I don't know of any hedge funds that trade breakouts, I know of thousands that successfully use arbitrage, actually the last most popular breakout trader was richard dennis, the guy who incubated the turtle traders group, Richards hedge fund blewup about 7 years ago, and he claimed ''markets had changed, information was being disseminated and priced in quicker, making trend following obsolete'' in other words he said ''the market changed and I didn't adapt in time''

Tech/a on individual stocks is good for determining trend, support, resistance, chart patterns, which can be used to give a loose forecast of a trading range, likely direction, etc....however it doesn't mean you can profitably trade using it though....largely because by the time you recognize the formation or setup, the attractive risk/reward profile has been sucked out of the equation, take for example the ''hammer'' formation on a candlestick chart, now its does have some predictive characteristics with other filters, however its not a profitable trading system, if you enter right at the close when the hammer formation will be shown on the daily chart, if you could enter in the bottom 10% of the hammer formation, i.e....near the low for the day, the system becomes profitable, however when the stock was at the low during the day, the hammer formation was unknown.

judging by the 100'000s of views on the numerous breakout threads around here, id imagine I would have stirred quite a few cups of tea out there by saying breakouts and tech/a is close to a load of crap and thats fine, its no co-incidence 90% of traders lose and 90% of traders only buy stocks, and most of them use ''breakouts'', well pro's buy and short stocks, the biggest hedge fund in the world is based on the pair trading concept and I can assure you all no successful hedge funds out there are trading ''hammers'' on the candlestick chart, lol

p.s. this isn't a stab at you, skyquake, more just stating my general opinion to the many subscribers of this journal.

 

[skyQuake]

The largest profit potential is when the largest amount of people are wrong, you can't make money by following what everyone else is doing, it amazes me how many people trade breakouts or think that is the way to profitably trade

Agree so much with you there.

No offence taken, I traded SPI and my fave play is false brks... But disagree slightly with you on Brkout trades. Sometimes the break is legit and it can run really hard with spec money jumping in, esp in market conditions where there is a lot of spec money sloshing around.

Cheers

 

[Truth]

You, as well as, other pair traders have said you like to see a "non-trending ratio chart". I don't think I'm understanding this correctly. Your QAN/MAP ratio chart on pg. 3, appears to be trending up. I suppose it's ok for oscillation as a whole to be trending up or down, as long as the ratio continues to cross the mean, back and forth? Is this right?

https://www.aussiestockforums.com/forums/showthread.php?t=14508&page=3

 

[So_Cynical]

Long LGL @ 3.23
Short SGX @ 6.03

As a long term Gold watcher and investor i can see where this combination
comes from...also can see how the short could go very very wrong.

For example SGX has formed a nice breakout at the 5.70s

Noticed that too.

There is a real casino - roulette element to that, the mathematics and hedging.

I suppose success is super dependent on the entry timing and stock selection.

How's this not the case for any other form of trading?

The hedging is something ive never seen in trading before this thread...very common in
Roulette systems, still the bottom line is u have to get it right "the punt" with any multiple
punt system there are degrees of winning.

Pairs trading would appear to be no different...its like having a punt on Red and evens on
a roulette table....there's low chance of winning both bets, even money to win one and get
your money back, and low chance to lose both...especially if u hedge with a small punt on 0
and perhaps a small (large square) or column bet.

 

[cooper1308]

I don't get it

Talking breakouts is irrelavent when these trades are based on quantative measures?

I mean, if you tell me "Double bottom on XYZ" and I come back with something fundamental such as " XYZ's Tobins Q Ratio is out of whack relative to peers" what would you say? Assuming you are profitable you wouldn't really care, you have your system and I have mine.

It's two completley different forms of analysis

Roullete analogy isn't the greatest either as you still have a negative expectation if you bet a colour and 0

Noticed that too.



The hedging is something ive never seen in trading before this thread...very common in
Roulette systems, still the bottom line is u have to get it right "the punt" with any multiple
punt system there are degrees of winning.

Pairs trading would appear to be no different...its like having a punt on Red and evens on
a roulette table....there's low chance of winning both bets, even money to win one and get
your money back, and low chance to lose both...especially if u hedge with a small punt on 0
and perhaps a small (large square) or column bet.

 

[tatankov]

PairsTrader,

Many thanks for your valuable contribution.
I still have some questions and doubts (regarding different time-frames) I should be able to understand after some practicing.

Since you have been a such a nice person I would like to hear some recommendations regarding the following:

* How often do you add/remove stocks to your "optimal universe of pairs" ?
* How many stocks do you have on your universe of pairs?
* Besides ASX and USA, what other stock markets do you recommend?
* Do you use layers on your trading? Do you recommend average down on losing positions?
* How much do you allocate in each pair? 5% ? 10 % ?
* How important is to you, RSI overbough/oversold levels?

Many thanks!

 

[Pairs Trader]

You, as well as, other pair traders have said you like to see a "non-trending ratio chart". I don't think I'm understanding this correctly. Your QAN/MAP ratio chart on pg. 3, appears to be trending up. I suppose it's ok for oscillation as a whole to be trending up or down, as long as the ratio continues to cross the mean, back and forth? Is this right?

https://www.aussiestockforums.com/forums/showthread.php?t=14508&page=3

You could say its slighty uptrending, once again its up to the trader to identify a good ratio chart, its really not that hard, and at the end of the day the system is still profitable with un-filtered trades.

 

[Pairs Trader]

* How often do you add/remove stocks to your "optimal universe of pairs" ?
* How many stocks do you have on your universe of pairs?
* Besides ASX and USA, what other stock markets do you recommend?
* Do you use layers on your trading? Do you recommend average down on losing positions?
* How much do you allocate in each pair? 5% ? 10 % ?
* How important is to you, RSI overbough/oversold levels?

Many thanks!

I do a complete delete/backtest/add new pairs about every 2 months, I have about 200 pairs on my radar, LSE and TSX are good exchanges to trade on, yes some trades I layer, not in this journal though, layering is for more experienced pair traders, I allocate between 20-40% per trade, RSI divergences are good, not so much overbought/oversold levels.