ASX Stock Pairs Trade Journal

[bochman]

Samg:

I failed to see the logic/reason for your new pair, they are not even at stddev 1?

 

[samgribbles]

yeah I know. I've just run some data through a spreadsheet at work. I think the PT program must have used 2 days old data. It was my intention to enter at -1.6 SD which is what my spreadsheet shows from 25 Sep data.

I'll have to be more diligent when looking for entries from the PT program at 11 at night after a beer or two!

Will monitor this one. May exit at the close for BE just to get out of a position I shouldnt be in

 

[tigger36]

Hi,
I've recently downloaded the Pair trade finder trial, and having some fun trying all sorts of filters to find good pairs, some successful, some not... I don't seem to be anywhere near as good as some of the posters on this forum with 75% plus winning trades. Is anyone willing to giveup any winning pair selection filtering...? (Please!)
Thanks

 

[samgribbles]

in addition to the advice from PairsTrader on this forum I would say:
- read PT's posts from the start
- print off the screen plots and study them
- understand the (very simple) stat functions behind the software
- go for low volatility pairs
- stick to stocks with higher prices (+$1) and liquid stocks


bochman - you should post your trades

 

[tigger36]

in addition to the advice from PairsTrader on this forum I would say:
- read PT's posts from the start
- print off the screen plots and study them
- understand the (very simple) stat functions behind the software
- go for low volatility pairs
- stick to stocks with higher prices (+$1) and liquid stocks


bochman - you should post your trades

Hi, thanks for the info.
When you say "stat functions", are you refering to all the different columns for %Win, AverageWin, Win/Loss ratio, etc..? or it maybe something I haven't discovered yet?
Thanks

 

[bochman]

Samg:

I will, if I manage to get in, none so far, I only have like 5 pairs on watch, still need more.

 

[samgribbles]

NAB/CBA - out at 31.26 for NAB, 52.02 for CBA. Both moved in the right direction but with the fees I made a loss of 0.9% on ROC => BE trade

Tigger:
no I meant the SMA, standard deviations, % from mean concepts.

And be wary of the backtester. It does not take into account drawdown and trade closeouts that would actually happen in reality. eg, in a trade you would not accept a 150% loss on capital to hang on and after mean reversion make a 10% profit - you would cut it earlier for a loss. The screener backtester hangs onto trades until the price ratio moves back to the SMA.

Eg. If the price ratio moves outside the 2xSD limit and does not come back to the SMA per the desired theory/outcome (volatility about the SMA), and it continues to trend away from the SMA then the backtester hangs onto the trade until the ratio revets or the SMA catches up with the new ratio range.

If this doesnt make sense then you need to spend a couple of hours investigating what the screener actually does. Don't think its too hard - it is VERY EASY theory.

Just research Bollinger Bands theory.

Even better. Set up a spreadhsheet for say NAB/CBA and calculate and plot the SMA and the Bollingers for the price ratio => you will then see the PT Finder signals in action over time - it will all come together

 

[tigger36]

NAB/CBA - out at 31.26 for NAB, 52.02 for CBA. Both moved in the right direction but with the fees I made a loss of 0.9% on ROC => BE trade

Tigger:
no I meant the SMA, standard deviations, % from mean concepts.

And be wary of the backtester. It does not take into account drawdown and trade closeouts that would actually happen in reality. eg, in a trade you would not accept a 150% loss on capital to hang on and after mean reversion make a 10% profit - you would cut it earlier for a loss. The screener backtester hangs onto trades until the price ratio moves back to the SMA.

Eg. If the price ratio moves outside the 2xSD limit and does not come back to the SMA per the desired theory/outcome (volatility about the SMA), and it continues to trend away from the SMA then the backtester hangs onto the trade until the ratio revets or the SMA catches up with the new ratio range.

If this doesnt make sense then you need to spend a couple of hours investigating what the screener actually does. Don't think its too hard - it is VERY EASY theory.

Just research Bollinger Bands theory.

Even better. Set up a spreadhsheet for say NAB/CBA and calculate and plot the SMA and the Bollingers for the price ratio => you will then see the PT Finder signals in action over time - it will all come together

Thanks for the pointers... it does all make sense.

 

[skc]

Samg:

I will, if I manage to get in, none so far, I only have like 5 pairs on watch, still need more.

Only 5 pairs on watch?!

Pair up the big 4 against each other and you have 6 pairs. Here are a few others to get your data base populated.

Building materials - ABC, BLD, JHX
Engineering services - UGL, DOW, MND, TSE
Media companies - APN, FXJ, TEN, WAN
Retailers - HVN, DJS (and Myer to come)

Do your own backtest and research

 

[samgribbles]

can you post your trades SKC?

 

[skc]

can you post your trades SKC?

I will post a summary of my trades tomorrow, but I won't post individual trades.

 

[samgribbles]

skc - PT and rozella may have disappeared due to ASIC RG 162 but they were seeking to make money from their posts, and I suspect they did not have an AFSL.

We are not. We are individual posters sharing ideas. When I post my trades I am making a statement of fact. I went long X and short Y. I don't recommend others mimic my trade, nor do I seek income from my posts. I seek critique. It idea sharing.

I dont think its an issue for us to post trades.

 

[skc]

Here is my pair trading result summary. Started around April this year but really got into it from June onwards. The stats are:

Total trades: 97
Wins: 77 (79%)
Losses: 20 (21%)
Avg win / avg loss: 0.62
Avg days held: 5.52

Total net return from starting capital: 26.6%
Commission paid / total net return: 42.8%

Some general comments about my trading so far:
- Commission is a big part of the strategy unfortunately. Even though I believe I already use the cheapest broker out there.

- I started with average leg size around 20-25% of total equity, but has now increased it ~30-35% depending on the trade and liquidity. Using smaller leg size allows more positions to be opened at the same time.

- I use CFD for short leg, but direct share holding for the long leg. Had I used CFD for the long leg as well I would have paid additional interest equating to ~10% of the total net return.

- I use market depth a lot to see which leg I will enter first and where to use limit orders or hit the market.

- I always read company news etc before entering, and I never short potential takeover targets, no matter how unreliable such rumours may be.

- I only pair up stocks in the same sector.

- I sometimes read the individual share charts before entering, so I will happily long a share that looks like ending a downward correction, or avoid shorting a share that is positioning for a potential breakout.

- I sometimes exit before the exit signals. Looking back at all the trades, I found that I underperformed the system by ~10% if I had followed all the signals. I think that is acceptable given the lag in prices of the software and the spreads.

- I use 10 day stop. As you can see the trades held >10 days have dramatically reduced win%.

- On a few occasions I have closed one leg and let another leg run with very tight stops. Given the market rally, allowing the long leg to run on occasions have been profitable.

I am not posting individual trades not because I am concerned about regulations. It's just too time consuming and potentially cause me to get worse fills!

 

[tigger36]

Here is my pair trading result summary. Started around April this year but really got into it from June onwards. The stats are:

Total trades: 97
Wins: 77 (79%)
Losses: 20 (21%)
Avg win / avg loss: 0.62
Avg days held: 5.52

Total net return from starting capital: 26.6%
Commission paid / total net return: 42.8%
...

Excellent feedback, thank you for this. I too have found from my backtesting that the big losers pull the net return right down...but I suspect that's all part of the strategy (just not a nice one!).

I would be interested on people's views on what fundamentals you look at/compare when choosing pairs. Eg.P/E ratios, Price/Book, Beta, MCap, etc..
Reading the PairTradeFinder top 12 tips they suggest finding pairs with similar P/E, MCap and other fundamentals, and also ensuring trades are in the right direction for compared Price/Book ratios for each. I've started looking at the later, but I am not entirely sure it is really useful for short time frames..? The thing is with trying to find pairs that meet all these requitements you end up with only a few trades a year...! So obviously some compromise is required.
Any views anyone?
Thanks

 

[bochman]

I am not sure whether I got it right, but here it goes

Long FXJ 1.59
Short APN 2.08

in the red at the moment.

Yearly chart suggest it is back to normal, but 45-60 days chart suggest stddev of over 2.

FXJ is a bit of a worry, due to the board fight (but it can be good once they fixed it).

comments welcome.

btw, I am not using PT's software, hence only 5 pairs on watch.

 

[samgribbles]

long WBC at 25.01
short CBA at 50.38

 

[bochman]

good pick, samg, does look good, 1 year is still not near stddev2, but 60days look ok.

 

[skc]

Excellent feedback, thank you for this. I too have found from my backtesting that the big losers pull the net return right down...but I suspect that's all part of the strategy (just not a nice one!).

I would be interested on people's views on what fundamentals you look at/compare when choosing pairs. Eg.P/E ratios, Price/Book, Beta, MCap, etc..
Reading the PairTradeFinder top 12 tips they suggest finding pairs with similar P/E, MCap and other fundamentals, and also ensuring trades are in the right direction for compared Price/Book ratios for each. I've started looking at the later, but I am not entirely sure it is really useful for short time frames..? The thing is with trying to find pairs that meet all these requitements you end up with only a few trades a year...! So obviously some compromise is required.
Any views anyone?
Thanks

I don't really compare fundamentals. Market cap yes to some extent just to make sure I don't trade a $3B company with a $100M one, but these pairs should not make it to your watchlist anyway. Beta is probably useful, but I haven't spend enough time or don't have enough brains to consider its bearing.

The other valuation metrics are probably not useful in such short time frames as you said. They may be useful for much longer term holds and if you compare each with their historical average PE.

Having said that, I do find going long on a quality stock makes me feel more certain about a trade, and likely to be more patience waiting for the convergence to happen.

 

[skc]

long WBC at 25.01
short CBA at 50.38

Personally I've found Big4 don't offer a very high potential reward. They do correlate very well but, because of that, the P/L after fees tend to be quite small.

So good hit rate but low reward. Usually if you can clear $100 out of every $10K you are doing well on a Big 4 pair trade. However, not best use of capital if you have other trades on offer.

 

[samgribbles]

Agree. I havent found as much time as I would like to generate a good pairs list. I'm mindful that idle capital in my CFD a/c is making zero return so use the banks trades which seem to occur frequently just to tick over the return. I try and close my banks trades out within 3 days and target an annualised ROC (I know you dont like ROC!) of +300% for the trade, which in nominal terms is usually less than 10% on the trade.