Quote:
Originally Posted by WizardHP
There is a lot of talk about the drawdown that may occur with C4 (CheckGrid poses drawdown issues as well, but nobody seems very interested in that excellent EA - too bad).
I am not exactly sure how to calculate historical draw down for C4. But since I have nothing else to do but spend my nights trying to understand trading, I took a stab at what I thought might give me some insight.
I took weekly data for the past 52 weeks on GBPJPY and CHFJPY. I calculated the weekly range for each pair and then compared the weekly range point difference for both. Long for GBPJPY and Short for CHFJPY. Weekly Range = Close - Open; Weekly Range Point Difference = Weekly Range for GBPJPY - Weekly Range for CHFJPY. (I'm not sure if this is the right way to do it and would certainly like to hear if someone has a different view.)
Anyway, the results ranged from a run-up of 630 pips for the week ending 1/5/07 to a draw down of 903 pips for the week ending 8/10/07. The average run-up/draw-down over the 52 weeks was .17 pips.
The results do not take into consideration that you might not have entered the trade at the week's open and may not have closed the trade at the close of the week.
Assuming this is the correct way to do it, the results would probably be more useful calculated on a daily time period.
I've been trading the demo for several months and it seems to me the daily draw down has never exceeded approximately 300 pips.
My purpose here was to try to figure out how much money I would likely need in my account to use C4 to earn roll over interest and benefit from occasional profit taking with this robot without having my account shut down by a margin call.
Seems to me that with $1,000 or so you could have traded 2 10k lots over the last year, earned the daily rollover, and made some additional profit with the Take Profit function.
Am I wrong, Muddy Guy? Where would you like the beer sent. You deserve a couple of six packs, at the very least. Am I wrong fellow thread followers?
|
Good morning,
First - we are talking about swap trading here. This is not correlation trading. Is everybody clear on that?
Second - the carry trade (swap trading) is not dead, contrary to popular news items in the past two months. In fact, it can be a solid strategy when use correctly. Many people got wiped out in their carry trades this year as prices moved significantly against them. Why did they lose money? One word - greed. They were too heavily leveraged in a market at historic highs and overripe for a substantial correction. Did any broker stop paying the daily swap? No. They paid every day. In fact, those who entered carry trades back when the USDJPY was in the 90's never broke a sweat during this last year.
So, the first thing that can hurt a swap trader is higher leverage. What is the second and only other thing that can hurt a swap trader? A reduction in the interest rate differential. That happens slowly, but if the trader is mired in a drawdown and the swap goes away, it is a slow financial death.
Two approaches to swap trading are to take a position and just sit on it, or to average in over time. I prefer the second. For example, if you put 10% of your account into the trade, then as $100 is earned in swap, you open new positions for 10% of that ($10). This creates a smoothing effect on the value swings. Interestingly, it also seems to create the same effect on the correlation factor.
It may also be a good idea to trade a basket of sorts with several currencies. My C4 robot can handle this. It is not always a good idea to be heavily exposed to one currency, such as the JPY.
So, to finally answer your question about exposure, one way to analyze historic divergence is to simply take the price of both pairs at a starting point and and ending point and do the math. For example, assume the GBPJPY is up 100 pips this year and CHFJPY is up 50 pips. If you are GBPJPY long and CHFJPY short your differential is a positive 50 pips. This is what I call correlation drift analysis. Drift is a function of long term fundamental strengths of the currencies involved.
Another important factor is spike analysis. You will find in my August posts that I described a negative spike in pairs I was trading and my feelings for British cows.
Looking at the previous example - what if one fine day the GBPJPY suddenly spiked down/back up 2000 pips without an offsetting move in the CHFJPY? Your transient differential is a negative 2000 pips. If you are leveraged too high, this will wipe you out, even if it "goes back" to where it should be.
So when you analyze a historic time period for the purpose of testing maximum drawdown, be sure to look for transient spikes, and calculate the differential at those points.
Trade carefully,
Bill