Geoff's FX Journal

December 6, 2005

Account Balance: $97,894
Unrealized P&L: $8,508
Net Asset Value: $106,402

AUD/JPY: Long 4 units.
CAD/JPY: Long 1 unit.
EUR/AUD: Short 2 units.
GBP/CHF: Long 1 unit.
GBP/JPY: Long 2 units.
NZD/USD: Long 2 units.
USD/JPY: Long 1 unit.

Everything's still riding, added 1 unit of EUR/AUD short. I'm starting to get nervous, seems like things are a bit overextended. Wouldn't mind if the system did some profit taking in the next few days/week.
 
December 7, 2005

Account Balance: $98,124
Unrealized P&L: $5,734
Net Asset Value: $103,858

AUD/JPY: Long 4 units.
CAD/JPY: Long 1 unit.
GBP/CHF: Long 1 unit.
GBP/JPY: Long 2 units.

Overextended, eh? Whipsawed out on EUR/AUD and USD/JPY. Got stopped out at ~.71 when NZD/USD went to hell.

AUD/JPY looks like it's hitting resistance at 90.40 - it's my current darling at 4 units, here's to hoping it keeps on running.
 
Quote from esuperbo:

Whipsawed out on EUR/AUD and USD/JPY.
More...

hi - good journal and you are doing good trading work - the usd/jpy has been in the 50%-70% of the average daily range for the last two days... - i don't quite understand how your system gets "whipsawed" when there's pretty much no movement... - i guess your stops are pretty tight, looks like you got stopped out of the same pair trade a few days ago when there was not much of an adverse move either.. - how much backtesting have you done on your system? - good luck with the trading, all the best.
 
Perhaps I'm using the wrong terminology - a whipsaw to me is something that just barely took out my stop, or hit my stop in an extremely sudden manner. Should I be using something different?

I'll describe my basic strategy using one pair. A unit (U1) is added with a volatility based stop (S1) if the pair shows a surge above weekend levels (somewhat of an ORB). The stop tends to be extremely tight (typically 0.5-0.75%). The stop is not moved during the week. Assume in the next week the pair surges again, and another unit (U2) is added. S1 is now moved up to S2. This got extremely ugly when I tried to type it out - it's quite simple, a table will illustrate well.

Week 1
Unit 1 - Bought at 100 - Stop at 99.5

Week 2
Unit 1 - Bought at 100 - Stop at 102.5
Unit 2 - Bought at 103 - Stop at 102.5

Week 3
Unit 1 - Bought at 100 - Stop at 102.5
Unit 2 - Bought at 103 - Stop at 105.5
Unit 3 - Bought at 106 - Stop at 105.5

Week 4
Unit 1 - Bought at 100 - Stop at 105.5
Unit 2 - Bought at 103 - Stop at 105.5
Unit 3 - Bought at 106 - Stop at 108.5
Unit 4 - Bought at 109 - Stop at 108.5

Now we're fully loaded - assume the trend keeps going. Instead of adding units, we keep adding stops (think of it as an extra unit of 1 dollar with appropriate stop), hopping our way up the trend week by week, one tight one (the profit taking stop) and last weeks (the stay in the game stop). Assume we hit 112.

Week 5
Unit 1 - Bought at 100 - Stop at 108.5
Unit 2 - Bought at 103 - Stop at 108.5
Unit 3 - Bought at 106 - Stop at 111.5
Unit 4 - Bought at 109 - Stop at 111.5

Now it goes to hell! We go back to 110.

Week 6
Unit 1 - Bought at 100 - Stop at 108.5
Unit 2 - Bought at 103 - Stop at 108.5

Next week another surge to 112.

Week 7
Unit 1 - Bought at 100 - Stop at 108.5
Unit 2 - Bought at 103 - Stop at 110.5
Unit 3 - Bought at 111 - Stop at 110.5

Does this make sense? The aim is to scale in with gradually larger money/stops/risk, ride big trends, and sell units off during churning or falling periods.

Tell me if I'm not making sense and I'll do my best to explain this better.
 
Almost forgot -

In terms of backtesting, long term I've run this through Excel and obtained good results but it's a crude approximation. I don't have the data or tools to get certain things, like the magnitude of the stops or carry boost/drag right. Starving student - I'm not about to go out and buy the latest and greatest software package but I'll do my best.

Shorter term, I ran this on a FXGame account for 6 months and I found the results quite acceptable. It was driving me crazy watching USD/JPY run up like mad when I told myself I had to trade it for 6 months with fake money, but that's life.
 
Quote from esuperbo:

I don't have the data or tools to get certain things, like the magnitude of the stops or carry boost/drag right.
More...

hi Geoff - thanks for the explanation - the historical forex data is likely to be in abundance for free nowadays thanks to the advent of the forex bucket shops etc - and since your system sounds fairly straight-forward, you don't need any particular "tools", i am guessing you can do it in excel (i have backtested quite complicated strategies with excel... excel can handle basic backtesting) - your system is 100% mechanical.. why wait six months to see the results if you can backtest in a week's time to see what makes sense.. - i am speaking from my own experience, if you are mechanical system oriented, i believe that the sooner you get into the proper strategy evaluation / testing framework, the faster your progress is going to be - sounds like you have an interesting system going, i almost want to say a "modified turtle approach" - you say:

Intent of this journal:
1. Provide a record of mental states/intentions that can be later reconciled with real world actions.

i believe the point of 100% mechanical is to take the emotion out of trading - just imagine how much further ahead you are going to be if you run even the basic but decent backtesting and gain an understanding of the dynamics of your system rather than spend your time on "recording your mental states" when trading something you are not so sure about... - this is just a thought, in either case, good luck and all the best.
 
p.s. here is a whipsaw on the euro - i think in your case you are mostly getting jiggled out / stops tagged on your tight stops.
 

Attachments

December 8, 2005

Account Balance: $98,460
Unrealized P&L: $4,900
Net Asset Value: $103,360

AUD/JPY: Long 2 units.
GBP/JPY: Long 2 units.
GBP/CHF: Long 1 unit.

AUD/JPY looks to be running out of steam... 2 units hit the upper stop. The long awaited JPY bull run in the cards?

fader - thanks for the advice, I appreciate it. I acknowledge that I must seem a bit hesistant, but this is my first time trading outside of rebalancing my mutual funds. I've been looking to get into the game for a while, but couldn't find a combination of markets and brokers that wouldn't cause my account to bleed to death through commissions.

OANDA scales well for my account size (ie small) and with my time constraints a longer term strategy like trend following seemed best. If it seems like a modified turtle strategy that's because it almost is - those rules were the first system I read. Blew my mind frankly, the entry was only a new damn high! Really caused me to rethink trading systems and how an entry is simply that, an entry... plenty more to consider.

I'm still working on backtesting and developing new systems - I assume the confidence comes in time. Even with a purely mechanical system I still find that emotion comes into play - I haven't acted on it, but I must condition myself to it before I get into more intensive strategies.

Again, thanks for the help.
 
ok sounds good - your 3% return on equity looks good, but i can't tell what leverage you are using / assuming? i assume you allocate among the pairs based on the respective volatility...
 
Currently with 5 units I have 88% of my margin available and my margin percent is 40% (at 20:1, default OANDA). I'm not entirely sure what these mean, I'm off to look them up. I'm currently guessing that margin available is (NAV - (position value/20)) and margin percent is (position value / (20*NAV))

I should keep track of this, what is the best metric for me to use - margin used/available, margin percent, or position value?

I currently size each trade by setting a volatility based stop, limiting each trade to a certain loss of closed equity (currently 2.5% / 8 pairs), and then figuring out the appropriate trade size.
 
You must log in or register to reply here.
Back
Top