ASX Stock Pairs Trade Journal

[skc]

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.

Yes. ROC figures are somewhat meaningless if not completely misleading. It is really only useful if you are very capital constrained. You are down to the last $500 for margin and have 2 different pairs you can enter, and the 2 potential trades have all parameters equal except for the CFD provider's margin requirement.

In that case, you can use projected ROC as your selection criteria. Other than that, ROC (and annualised ROC) is not a meaningful performance measure.

If you ever find yourself in that situation, however, then you probably shouldn't be pairs trading with CFD. You don't want to blow your account on a bad trade if you are so capital constrained.

Also, with CFD's I would not long on stocks that might undertake a capital raising. You will take the full hit of the dilution with none of the upside from rights issue or SPP etc.

 

[samgribbles]

bochman, I note your analysis periods of 1 yr and 60 days. FYI I use 2-3 weeks for the SMA and SD calcs.

Are you looking to trade frequently or looking for longer term mean reversion investing with options or similar?

 

[samgribbles]

SKC. Noted. But in the absence of any better way to measure the profitability of an individual trade I use ROC, and the annualised ROC while crude and meaningless just helps introduce the time value. Do you have an alternative suggestion? I'm happy to change my ways.

I use IRR for the account growth over time which I believe is appropriate as it used balances net of fees etc and funds deposited/withdrawn.

 

[bochman]

bochman, I note your analysis periods of 1 yr and 60 days. FYI I use 2-3 weeks for the SMA and SD calcs.

Are you looking to trade frequently or looking for longer term mean reversion investing with options or similar?

I think PT mentioned that he uses 45 days or 60 days, hence I am using 60 days as a guide. 1 year is looking at the long term trend, in general, I found that if both 1 year and 60 days both indicates stddev of ~2, they have a higher chance of success.
Obviously with this strategy, a cut loss of within 10-15 days is appropriate, hence I am basing it more toward 60 days trend, with 1 year as a guide.

I generally think 2-3 weeks trend is a bit too short, it doesn't really give you the overall picture, I can be wrong here.

 

[samgribbles]

Bochman - no, no right or wrong, just different style.

CLOSED WBC / CBA
long WBC at 25.01 / 25.25 = +0.96%
short CBA at 50.38 / 50.16 = -0.44%
=> net move = +1.4% profit
=> with CFDs, after fees ROC = 9.9% in 2 days (opened yesterday)

 

[samgribbles]

long NAB at 29.37
short CBA at 50.16

 

[samgribbles]

long QAN at 2.82
short VBA at 0.455

not too sure about this one, first time this pair

SKC - any thoughts?

 

[bochman]

Samg:

is your ROC including commission? I found commission took a bit part out of profit normally.

 

[skc]

SKC. Noted. But in the absence of any better way to measure the profitability of an individual trade I use ROC, and the annualised ROC while crude and meaningless just helps introduce the time value. Do you have an alternative suggestion? I'm happy to change my ways.

I use IRR for the account growth over time which I believe is appropriate as it used balances net of fees etc and funds deposited/withdrawn.

As your position sizing is based on some sort of "notional capital", your ROC should be based on that "notional capital" amount. E.g If each leg is 30% of your captial ($50K x 30% = $15K) and you made $1000, you have returned 2% on your capital. How much CFD margin you have to put up really has no bearing what so ever.

Whilst what you are doing is not wrong, it could potentially lead you down the wrong path (e.g. chasing quick closes, going with lower margin requirements when another better trade is on offer).

Not sure how you use IRR. That's not my understanding of how IRR is used. Just look back at the end of year and keep it simple. If I started with $50K and now has $60K, I've made 20% so far...

I think PT mentioned that he uses 45 days or 60 days, hence I am using 60 days as a guide. 1 year is looking at the long term trend, in general, I found that if both 1 year and 60 days both indicates stddev of ~2, they have a higher chance of success.

This is interesting. Do share more as you do more research / trade.

Bochman - no, no right or wrong, just different style.

CLOSED WBC / CBA
long WBC at 25.01 / 25.25 = +0.96%
short CBA at 50.38 / 50.16 = -0.44%
=> net move = +1.4% profit
=> with CFDs, after fees ROC = 9.9% in 2 days (opened yesterday)

long NAB at 29.37
short CBA at 50.16

Why didn't you just keep the CBA short? Saves you some brokerage.

long QAN at 2.82
short VBA at 0.455

not too sure about this one, first time this pair

SKC - any thoughts?

I wouldn't do this pair because VBA is a lot more beaten down and can therefore potentially bounce violently. At their peak, QAN/VBA was ~2-2.5, now they are like 6 or 7. Also I have a decent long position in VBA. Try QAN/MAP pair which has a solid performance history - I just missed an entry this morning on the pair.

 

[bochman]

closed FXJ & APN

1.58 & 2.03

probably a tag too early.

 

[bochman]

oh, btw, looking at the 1 year, 60 and 45 days chart for VBA & QAN, I think it may work, all confirm stddev of >= 2 now.

 

[samgribbles]

bochman - yes, my ROC figures include brokerage of 0.011%. PM me your email address and I will send you the spreadsheet I use if you want.

skc
- re QAN/VBA. Good points, thanks.
- re CBA. Yes, should have. I closed my open trade, then checked old signals and found the NAB/CBA was on so re-entered CBA.

 

[samgribbles]

CLOSED NAB / CBA
long NAB at 29.37 / 31.27 = +6.5%
short CBA at 50.16 / 52.68 = +5.0%
=> net move = +1.5% profit
=> with CFDs, after fees ROC = 10.2% in 3 days

NAB up 4.3% today after a 0.3% rise in Wall St. Tempting to short it for a little pullback before the close

 

[bochman]

told u I closed out FXJ & APN pairs too early....
ah........

I seriously think VBA & QAN will work now, but it will be volatile.

 

[skc]

told u I closed out FXJ & APN pairs too early....
ah........

I seriously think VBA & QAN will work now, but it will be volatile.

May be, may be not. I've never topped up when a trade's gone against me.

If I was doing 30% a leg, it becomes hard to manage "single name risk" if I was to increase it to 50%. Just gets a bit scary for me.

QAN / MAP pair has been going off in the past few days.

5/10/09 Long MAP 2.59, Short QAN 2.81
6/10/09 Sell MAP 2.81, Cover QAN 2.84

7/10/09 Short MAP 2.95, Buy QAN 2.83
8/10/09 Cover MAP 2.87, Sell QAN 2.90

8/10/09 Buy MAP 2.80, Short QAN 2.99

And I managed to miss ALL of them

Completely off topic... did anyone see beauty and the geek australia? That air hostress was hot - I bet you she works for VBA.

 

[samgribbles]

very very good QAN/MAP trades! love the head bangin icon, if only I could insert that into some work emails!

 

[skc]

A very good trade setup at the moment is long PPT against AMP/AXA/PTM.

All of these wealth management companies have spiked up and made new highs, except for PPT still in consolidation mode.

I have attached PPT and PTM charts for information. Don't have a screen shot from the software at the moment, but divergence is well above 2.5 std dev.

Just an example how reading individual share charts can help me judge the potential merits of a trade.The pair trade will rely on the breakout of PTM be retested (or close the gap) while PPT moves into new high (market condition pending).

Some catuion required as the sector is under takeover rumours at the moment. Banks awash with capital are said to be eyeing the likes of AMP, while AXA is always prone to prey by it's French parents. But if someone was to move on one of them, the whole sector will follow suit no doubt.

 

[tigger36]

mmm, i'm not doing very well at this pairs trading lark... i've only just started so hopefully just a bad luck start, but out of 5 trades, i've 2 wins and 3 losses. Oddly the 2 worst losses are the ones I thought were good setups. I suspect I need to look at the charts in more detail, as it seems when a pair goes out of sync but caused by one of them breakingout, it can mean that stock keeps going out of sync as other traders are possibly buying into it as a breakout, it doesn't mean the paired stock is going to follow suit...
For instance i'm currently holding onto a trade in MGGT.L/RR.L, 96% correlation, and rising, but it really got out of sync with a breakout of MGGT.L on 6 Oct, and has carried on marching upwards, with RR.L going downwards! Maybe it will snap back together (I hope!), or can you see any fundamental reason why the rating of MGGT.L to RR.L has changed?
Thanks

 

[skc]

mmm, i'm not doing very well at this pairs trading lark... i've only just started so hopefully just a bad luck start, but out of 5 trades, i've 2 wins and 3 losses. Oddly the 2 worst losses are the ones I thought were good setups. I suspect I need to look at the charts in more detail, as it seems when a pair goes out of sync but caused by one of them breakingout, it can mean that stock keeps going out of sync as other traders are possibly buying into it as a breakout, it doesn't mean the paired stock is going to follow suit...
For instance i'm currently holding onto a trade in MGGT.L/RR.L, 96% correlation, and rising, but it really got out of sync with a breakout of MGGT.L on 6 Oct, and has carried on marching upwards, with RR.L going downwards! Maybe it will snap back together (I hope!), or can you see any fundamental reason why the rating of MGGT.L to RR.L has changed?
Thanks

I have no idea about these stocks as I don't trade outside the Australian market. I assume you've checked that there were no news when you opened the trade?

How long have you had the trade open?

Frankly 2 wins 3 losses in the first 5 means very very little.

 

[tigger36]

I have no idea about these stocks as I don't trade outside the Australian market. I assume you've checked that there were no news when you opened the trade?

How long have you had the trade open?

Frankly 2 wins 3 losses in the first 5 means very very little.

Yes, I know you're right.

4 days in trade now, but I think putting that post on here had the right affect, the trade is starting to move in my favour now...

I think my key question really was people's view on "breakout" deviations?, which can go a long way from the mean before reverting?

Thanks