ASX Stock Pairs Trade Journal

[SilverRanger]

Real time prices for ASX shares!!! I will have that over anything else.

Real time ASX prices +1, or better still generic methods to hook up to established data provider/ broker data (somewhat like what Ninja Trader can do).

Some multi-core optimisation (so that more price refreshes can be done in any given time) would be nice too

 

[fingl]

Assuming closing prices are used for entering orders how would a trader manage multiple pairs because you would require to compute the closing price whether there is a signal around market closing hr. Realtime data must be fetched according. And a signal does not necessary implies transaction. A trader would open individual chart to analyze patterns and indicators. Multiple pairs would all these be done quickly around market closing hr and not considering putting orders because only around these hr the closing price is fetched. How do you guy manage all these?

Is a nice slanting equity line be more important than fundametal consideration? Would you take a signal for two firms one on technology and the other on fashion but yet having a nice sloping equity line from testing? Some trader advocates using mainly testing statistics and ignore fundametal consideration and correlation.What your take?

 

[skc]

Assuming closing prices are used for entering orders how would a trader manage multiple pairs because you would require to compute the closing price whether there is a signal around market closing hr. Realtime data must be fetched according. And a signal does not necessary implies transaction. A trader would open individual chart to analyze patterns and indicators. Multiple pairs would all these be done quickly around market closing hr and not considering putting orders because only around these hr the closing price is fetched. How do you guy manage all these?

Is a nice slanting equity line be more important than fundametal consideration? Would you take a signal for two firms one on technology and the other on fashion but yet having a nice sloping equity line from testing? Some trader advocates using mainly testing statistics and ignore fundametal consideration and correlation.What your take?

With the right trading platform you can largely get around those problems. I use webIress which displays the anticipated match price (which can and do change alot during the closing auction) and excess volume. I also run a spreadsheet which tells me at what price ratio I should be closing the trade. I can have 10-15 pairs open at any one time but it's mostly manageable.

I only pair things with the right fundamentals... but there are many ways to skin the pairs-trading cat. Just whatever you feel comfortable with...

 

[sleepy]

Real time ASX prices +1

Real time ASX prices +2

What sort of results has everyone been getting the past 3 months?
Is 2011 better or worse than previously? ... and is pairs trading still your main strategy?

sleepy

 

[fingl]

With the right trading platform you can largely get around those problems. I use webIress which displays the anticipated match price (which can and do change alot during the closing auction) and excess volume. I also run a spreadsheet which tells me at what price ratio I should be closing the trade. I can have 10-15 pairs open at any one time but it's mostly manageable.

I only pair things with the right fundamentals... but there are many ways to skin the pairs-trading cat. Just whatever you feel comfortable with...

How many pairs do you watch for potential signals? Do you study all their chart and fundametal even they do not have signals currently?

Are you talking about fundmetal similar to p/e, sales, net profit, roc, cashflow, debt ratio and other many financial figures? And their variance should not be alot. The question is whether these fundametal provide additional edge to the trading.

 

[skc]

Real time ASX prices +2

What sort of results has everyone been getting the past 3 months?
Is 2011 better or worse than previously? ... and is pairs trading still your main strategy?

sleepy

2011 has been quiet for me until March... I made more trades in the last 2 weeks than the whole of Jan and Feb combined. Account is up a healthy 8.5% in March, ~12% for the calendar year

 

[skc]

How many pairs do you watch for potential signals? Do you study all their chart and fundametal even they do not have signals currently?

Are you talking about fundmetal similar to p/e, sales, net profit, roc, cashflow, debt ratio and other many financial figures? And their variance should not be alot. The question is whether these fundametal provide additional edge to the trading.

I have ~200 pairs so probably ~170 different stocks in total. Some stocks I know better fundamentally than others. There is not much point knowing all the charts - it doesn't take long to look up the chart after a signal is given.

I try to learn fundamentals at a higher level... things that can be looked up more quickly. What business are they in, where do they operate, are they profitable or not, a general idea about their debt levels (no debt, normal debt vs high debt) and what are some key drivers influencing the share price.

E.g. BSL and OST are in the same industry, but OST makes a profit while BSL makes a loss. So I will have a bias to long OST / short BSL (which I did when BSL spiked up 2 days after the Japan earthquake).

Dividend yield is also quite useful... a quick glance at the Big 4 yield will show you that ANZ has the lowest yield, so I might think twic about long ANZ against say CBA or wait for the pair to diverge further than what the signal says.

 

[fingl]

I have ~200 pairs so probably ~170 different stocks in total. Some stocks I know better fundamentally than others. There is not much point knowing all the charts - it doesn't take long to look up the chart after a signal is given.

I try to learn fundamentals at a higher level... things that can be looked up more quickly. What business are they in, where do they operate, are they profitable or not, a general idea about their debt levels (no debt, normal debt vs high debt) and what are some key drivers influencing the share price.

E.g. BSL and OST are in the same industry, but OST makes a profit while BSL makes a loss. So I will have a bias to long OST / short BSL (which I did when BSL spiked up 2 days after the Japan earthquake).

Dividend yield is also quite useful... a quick glance at the Big 4 yield will show you that ANZ has the lowest yield, so I might think twic about long ANZ against say CBA or wait for the pair to diverge further than what the signal says.

Do you choose these 200 plus because of their equity curves from historical trades are nicely trending or they are chosen because of other reasons or their nature of businesses and industry are the same. The best equity curve is from cross industries companies having little relation to each other. Would you take the trade for cross industries having little relation because of their excellent equity curve but may due to spurious relation?

Assuming closing prices are used do you place orders on the close or on the next open? Because for 200 plus manually placing orders for 5 or more may require speed considering the market about to close and orders have to be placed on close.

 

[skc]

Do you choose these 200 plus because of their equity curves from historical trades are nicely trending or they are chosen because of other reasons or their nature of businesses and industry are the same. The best equity curve is from cross industries companies having little relation to each other. Would you take the trade for cross industries having little relation because of their excellent equity curve but may due to spurious relation?

Assuming closing prices are used do you place orders on the close or on the next open? Because for 200 plus manually placing orders for 5 or more may require speed considering the market about to close and orders have to be placed on close.

I will answer these after you share something fingl... Have you started trading pairs? How's that going? Who do you use as your broker/provider? Which markets you trade? What's your approach?

Just feeling a bit one-wayed

 

[fingl]

I will answer these after you share something fingl... Have you started trading pairs? How's that going? Who do you use as your broker/provider? Which markets you trade? What's your approach?

Just feeling a bit one-wayed

skc, did not try trading pairs but has been interested. There are a few reasons why not. One reason is the reward/risk profile for this field. Done quite a lot of backtesting on us securities and the r/r figures varies 0.3 between 1.5. No doubt there are exceptional r/r > 1 but the bulk average is 0.5 to 0.8. And the max downdraw figures are not beyond what a trader should accept. On 1000 securities, > 70% has max dd > 40%. Max dd of 25% accounts for < 5%.

On using more data for backtesting, these r/r and max dd figures become more unfavor for a trader. Same for the win rate 55% to 65%. With these win rate and on r/r average 0.5 to 0.8 any more large losing trades would steer the equity curve south. This field maybe alot of winners but a few large trades would give back most of the profit. Exiting using trades bars do not seems to solve the problem. The main reason is exiting on trades bars might be more unfavor compared to let the trade exit using the original rules. The trade might exit on the lowest pnl.

Using a simple rsi rule for backtesting, on 1000 securities, there are definitely a certain % with exceptional good equity trend. Is this due to the individualize nature of the security suiting the rsi rule or spurious? A random rule would also generate a certain % of good equity trend.

You have a good equity trend which seems to deviate from mass stat figures for this field. Do you use alot of discretion on trades selection? Any recommendations to prevent large losing trades beside the 3 rules you have. Because unless the trader is really really really experience, there are definitely large losing trades and these would decrease the r/r to very unfavor.

Alot of traders advocate allocate a small amount of equity on each trade but multiple small losing trades consecutively with each losing amount to the largest losing dollar would also meant max dd. Take 6 trades with with each losing 6% consecutively would meant a 36% down. And 6% is a very conservative figure. Off course 6 consecutive losing trades is remote but they do occur no matter how experience the trader is.

 

[skc]

skc, did not try trading pairs but has been interested. There are a few reasons why not. One reason is the reward/risk profile for this field. Done quite a lot of backtesting on us securities and the r/r figures varies 0.3 between 1.5. No doubt there are exceptional r/r > 1 but the bulk average is 0.5 to 0.8. And the max downdraw figures are not beyond what a trader should accept. On 1000 securities, > 70% has max dd > 40%. Max dd of 25% accounts for < 5%.

On using more data for backtesting, these r/r and max dd figures become more unfavor for a trader. Same for the win rate 55% to 65%. With these win rate and on r/r average 0.5 to 0.8 any more large losing trades would steer the equity curve south. This field maybe alot of winners but a few large trades would give back most of the profit. Exiting using trades bars do not seems to solve the problem. The main reason is exiting on trades bars might be more unfavor compared to let the trade exit using the original rules. The trade might exit on the lowest pnl.

Using a simple rsi rule for backtesting, on 1000 securities, there are definitely a certain % with exceptional good equity trend. Is this due to the individualize nature of the security suiting the rsi rule or spurious? A random rule would also generate a certain % of good equity trend.

You have a good equity trend which seems to deviate from mass stat figures for this field. Do you use alot of discretion on trades selection? Any recommendations to prevent large losing trades beside the 3 rules you have. Because unless the trader is really really really experience, there are definitely large losing trades and these would decrease the r/r to very unfavor.

Alot of traders advocate allocate a small amount of equity on each trade but multiple small losing trades consecutively with each losing amount to the largest losing dollar would also meant max dd. Take 6 trades with with each losing 6% consecutively would meant a 36% down. And 6% is a very conservative figure. Off course 6 consecutive losing trades is remote but they do occur no matter how experience the trader is.

Interesting stats. On those numbers you wouldn't pairs trade that's for sure. My numbers are better so I guess it shows that I haven't wasted all my time in understanding the company fundamentals and watching market news.

The simplest way to reduce drawdown is to reduce your position size. I have done ~1000 trades in the last 2 and a bit years and my single largest losing trade was 2.2% of my account. That particular trade went 20% against me, but my position size was only ~10% of my account. It wasn't a pair I know well and they were smaller, more volatile companies. I also exited at the worst possible PnL - but $hit happens and I had to follow my stop rules.

Position sizing is the key to survival in most trading strategies. Trade bigger and you have to brace for bigger drawdown. Personally I have a specific annual return target in mind, and so I know how hard/not hard I need to push my account to achieve that. This means that I rarely have a position that is >20% of my account. If you trade with a smaller account however it will be a different story.

Trust me you don't have to worry about consecutive losses. Without any doubt your largest drawdown will occur when you hit a patch of low win rate, not when you hit consecutive losses. You see consecutive losses mentioned a lot in many trading books but it's simply wrong.

My largest drawdown was ~8% for the account. It took me 60 trades to get there from the previous equity high. Within that time, the longest losing streak was only 4 losses, but the win% fell to 55% (compared to the usual 75-80%). Avg win/avg loss was 0.35, compared to the usual 0.6. Consecutive losses have nothing to do with it.

All I can say is somehow I've managed to make pairs trading work most of the time, and I personally don't know how one can do it without a bit of investment in fundamental research. In my view that goes a long way to improving your win rate and avoiding large losses.

 

[Pairs Trader]

Guys, thanks for the suggestion and I agree live ASX data would be a bonus, but here is the brutal honest truth why I haven't done it, I think by telling you exactly why, you may understand and hopefully this comes across the right way, not that i'm under-committed to us down-under traders;

Finding a suitable data feed;

Believe me I've searched high and low to find a suitable data feed, but for ASX this doesn't seem feasible, aside from the high cost to implement this feature there would be a very high cost to the trader (about $150 per month) for a suitable feed, it's not as simple as hooking into a master feed or some website(that's illegal), each trader would need a subscription, because the feed we would have to implement needs to be API supported and be well documented for my programmer to integrate, there is a lack of this type of service here in aussie.

So it's a cost vs benefit argument, I know from experience with our IQfeed for real time data on US stocks, forex, commodities at $60 per month that not many traders take that up and the majority of our clients are US based, because not many people like subscriptions, they like one-off costs, so at a high cost to me to develop it I don't think many ASX traders would take it up.

That's the honest reason. Now, if a bunch of you are really keen and we can find a suitable provider that has good support, maybe we can all chip in to get my programmer to develop a custom version, he charges $100 per hour, however at this stage it would be a loss-maker for me, please don't take offense, it's just business.

 

[fingl]

Interesting stats. On those numbers you wouldn't pairs trade that's for sure. My numbers are better so I guess it shows that I haven't wasted all my time in understanding the company fundamentals and watching market news.

The simplest way to reduce drawdown is to reduce your position size. I have done ~1000 trades in the last 2 and a bit years and my single largest losing trade was 2.2% of my account. That particular trade went 20% against me, but my position size was only ~10% of my account. It wasn't a pair I know well and they were smaller, more volatile companies. I also exited at the worst possible PnL - but $hit happens and I had to follow my stop rules.

Position sizing is the key to survival in most trading strategies. Trade bigger and you have to brace for bigger drawdown. Personally I have a specific annual return target in mind, and so I know how hard/not hard I need to push my account to achieve that. This means that I rarely have a position that is >20% of my account. If you trade with a smaller account however it will be a different story.

Trust me you don't have to worry about consecutive losses. Without any doubt your largest drawdown will occur when you hit a patch of low win rate, not when you hit consecutive losses. You see consecutive losses mentioned a lot in many trading books but it's simply wrong.

My largest drawdown was ~8% for the account. It took me 60 trades to get there from the previous equity high. Within that time, the longest losing streak was only 4 losses, but the win% fell to 55% (compared to the usual 75-80%). Avg win/avg loss was 0.35, compared to the usual 0.6. Consecutive losses have nothing to do with it.

All I can say is somehow I've managed to make pairs trading work most of the time, and I personally don't know how one can do it without a bit of investment in fundamental research. In my view that goes a long way to improving your win rate and avoiding large losses.

Felt there are not much significant things to contribute compared to many of veterans in this field except to learn more from you guys. skc what were the reasons for your 8% dd which decreases your win % to a mere 55%? Any steps you took to prevent this max dd after reaching new equity high? Consecutive losers may not be the main factor for dd. More losing trades compared to winning trades on whatever sequences might be more contributing to dd. Your r/r averaging on 0.6? Thats imply you are putting a dollar to profit 0.6. Your win % must be really good to maintain the current equity trend.

About reducing position size with relation to max dd. No doubt small allocation of equity would reduces dd but this would imply your win would reduce too. Most trader would concern what is the duration to reach new equity high. Smaller allocation would imply proportional more trades and more duration to cover the previous max dd. How does position allocation helps with the < 1 r/r for pairs trading?

Good to consider fundametal. There are alot of different ways to use fundametals. Some uses value and growth rules, accounting statements, cashflow positions, debt to equity, pnl determination and not considering more ratios derive from more calculations. But which is more relevant to pairs trading?

 

[fingl]

Is trading trading signals from cross but related industries a good idea?

 

[edman79]

Does anyone want to post some trades in real time and (more importantly) explain why they took the trade. In order for me to learn, the profit of the trade is less important than why it was actually taken.
--Looking at you SKC

There are many real time trades on this thread but they dont really explain by saying "I took this trade because the PE ratio was out ... etc. etc.

The ones that are taken are based on technical data (charts etc) but the explanations dont seem to be consistant ie. some are justified with spread, some with RSI, some at all time highs/lows but never 2 the same.

 

[skc]

Is trading trading signals from cross but related industries a good idea?

What is cross but related industries?

Does anyone want to post some trades in real time and (more importantly) explain why they took the trade. In order for me to learn, the profit of the trade is less important than why it was actually taken.
--Looking at you SKC

There are many real time trades on this thread but they dont really explain by saying "I took this trade because the PE ratio was out ... etc. etc.

The ones that are taken are based on technical data (charts etc) but the explanations dont seem to be consistant ie. some are justified with spread, some with RSI, some at all time highs/lows but never 2 the same.

*Looking away....*

Too time consuming, and the more I post the worst entry/exit I am going to get! I struggle with liquidity on some pairs already.

There is nothing wrong with looking at multiple charts/data points... the more things that support a trade, the higher chance of a profitable outcome.

 

[fingl]

Cross but related industries

One security might be dealing with oil the other deal with oil equipment. Both are on oil industry and amount of oil business affect both securities profit.

 

[fingl]

skc you are trading the asx 200 and having liquidity issue? You have problem having fill on the selling? What do you do with one security fill on buy but the other not getting fill on sell?

 

[skc]

Cross but related industries

One security might be dealing with oil the other deal with oil equipment. Both are on oil industry and amount of oil business affect both securities profit.

Sure. I would pair those up.

skc you are trading the asx 200 and having liquidity issue? You have problem having fill on the selling? What do you do with one security fill on buy but the other not getting fill on sell?

I trade outside ASX200 as well. The volumes I do are not that great, but some ASX200 shares can have bad liquidities on certain days. I consider liquidity as poor if I can't close my position without taking out more than 1 level (when I hit the market). Or if I put my full size in a single price level (as limit order) and it changes the market depth balance a bit too much.

If I have 1 leg open but can't get a fill on the other leg, I either close the open leg (i.e. abandon the trade - which I do if the profit target is thin), or I reach down to open the other leg (more acceptable if the profit target is high). Or I sit in the queue and keep waiting, if the stock is just bouncing between bid and ask and the whole market is steady. But this can be risky as well - I can't tell you the number of times I waited and waited for 1m shares to trade in front of me, and only to have the price moving away from me when there were only 20K left... it's like someone knew I was there! Then I have to decide whether to wait again, or reach down over 2 price levels (which is significant if the share price value is under $1).

 

[kingcarmleo]

Just read most of this thread from the start, awesome!