Currency Basket and Carry Trading

Quoting firehorse

Disliked

Did you have a system for picking those two particular pairs as opposed to any other pairs.

Ignored

Quoting firehorse

Disliked

Also, do you have a system for changing the pairs if circumstance change?

Ignored

Quoting BobMM

Disliked

Ok, here's a comparison of basket prices over 5 years. Included is: mine, ElectronicSavants, and DGCFX's. Those last two have a very similiar basket with a slightly different weighting. Both of them went into negative NAV territory in the start of 2002 at their default settings because they use up 50% of the account as margin when doing 15:1 leverage. (exotics and hedges)

If you guys find flaws with what I'm doing, point em out! I'm as amazed and impressed by this basket's performance as you may be, and I want to make sure I can make the most profitable system from it possible.

-Aksel

Ignored

Quoting sjburdi

Disliked

Thanks Aksel, I changed my basket pricing backtesting strategy to use cumulative percentage difference for each day, instead of pip price. I am testing ES's basket so I multiplied the daily percentage difference based upon the percentage weighting factor for that currency in USD. I sum up the total percentage each day and have a cumulative running total of basket price % change. The DD has gone as low as 4.4%. Any suggestions? Is there something I am doing wrong or should be adding to the formula?

Ignored

Quoting BobMM

Disliked

DD should be much larger than that, 2/20/04-5/11/04 shows up around -6% DD for me. From the way you phrased your methodology, it's hard to figure out where our methods diverge. I'll explain mine in more detail and if you do the same we can find where we differ.

I have a huge table of prices... so if AUDJPY is in the basket, I first make a new column of the USD equivalent of AUDJPY. (AUDJPY*1/USDJPY). This is the column I use for basket price determination. On the day it is brought in I set another column with the total change of AUDJPY USD equiv based on today's price change. So for say, 3 days where the price went up 10% each day, we'd see:

0%
10%
10%
10%

Final basket price is Weighting*%daily change of each pair + next pair, etc and denominated in dollars the whole time. Maybe the difference is in how we did our weightings. Looking off his exposure tab, I did the weighting as 11364/85512*(%chg) + 6818/85512(%chg), so I think I used the units of each pair/total # of units. This is a real head wrapper now, because those are units of the currencies not translated into USD, so did I make a mistake here? But the units shouldnt matter, since the final currencies are expressed in USD, right?

I'm thinking I did it correctly, since to decide how many of each to buy you start with his amount of NZDUSD, and then use ratios to buy units of the others, and it's in a set amount. So as long as you have the right ratio of units you should duplicate his performance methinks. Thoughts?

I suspect we did our weightings differently and that's the reason for the discrepancy in our findings. Anyone else want to chime in and offer their opinion? How do we do his weighting?

-Aksel

Ignored

Quoting sjburdi

Disliked

Aksel: You are right. I am doing the weighting totally differently. I do not use Units to determine weighting. What I do is determine the dollar amount invested for each pair. Then I determine what percentage of the total amount invested that pair represents. So now I have my weighting in todays dollars. Then i have to go back in time to determine my percent weightings 5 years ago - in order to get an accurate starting point based upon prices 5 years ago. To do this, I determine the price percentage difference for each pair from 5 years ago to todays price. Now I apply that percentage change to each pairs current today's price and determine a new 5-year ago weighting for each pair and start from there. Then I get a percentage difference each day for each pair and total up the basket. This strategy may be flawed as well, as I continue to think about it. Please let me know your thoughts here.

Ignored

Quoting BobMM

Disliked

So clear me up on this one point: Are you dollar-weighting it based on the prices 5 years ago or prices today? Because, prices change. I'm not quite sure which method is right, quite frankly. And you're not changing the weightings, right? Or are you adjusting the weighting daily based on dollar fluctuations?

Better yet, how many of each unit do you hold today? I'm thinking dollar-basing isn't exactly following his model, because he mentioned he didn't expect changes to weighting except very rarely, if at all. The point is to backtest his system as it is now, with the weightings he and dfgx came to in hindsight, right? Rather than what it may have been 5 years ago?

-Aksel

Ignored

Quoting JackJones

Disliked

...and this is all thanks to Electric and his posts

My style is to have the negative weights to counter big moves which still earn some Interest.

Have the Earners where I balance them around zero (not sure how to term this)

And my Trenders where I get net short/long and play the breaks.

Slightly more time consuming but it gives me lots of time to think. It does cause some head aches and I sometimes can not be bothered thinking about it to be honest but it's nice to have time on my side.

I'm also looking to pin point big fundamental news and get flat or square and ride out the change.

Fun for the whole family really.

Ignored

Quoting ElectricSavant

Disliked

Your assuming your exposure is not the culprit and WHEN you you "enter", "scale in" or "add to"...

Ignored