A truly riskless system?
- Thread starter MathAndLogic
- Start date
Quote from failed_trad3r:
If you go long EUR/USD and long USD/EUR at same time, you can make money
for example if EUR/USD goes from 1 to 2, a 10.000 pip move up then USD/EUR naturally goes from 1 to 0.5, a 5000 pip move down.
One trade will make double the pips the other one loses, and then you make 5000 pips profit on this trade.
A complicated version of this strategy is going long USD/GBP, long GBP/EUR, and long EUR/USD, threeway trade. This is equal to three seperate trades in one!:eek:
If you had ever traded forex you'd know there's a EUR/USD pair but no USD/EUR pair. The way you short the euro in dollars is to sell EUR/USD short. Obviously going long EUR/USD and short EUR/USD at the same time cancels out so you may as well have not traded (unless you enjoy paying gratuitous commissions and bid/ask spreads).
Now suppose there *was* a USD/EUR pair to trade? Then the pips for EUR/USD would be in 10,000ths of a dollar while the pips in USD/EUR would be in 10,000ths of a euro. If EUR/USD goes from 1 to 2 you gain 10,0000 dollar pips and lose 5,000 euro pips. Each euro pip is worth twice as much as a dollar pip so you lost as much as you gained.
Now you know why you're a failed_trad3r.
There's been a lot of uneducated argument on this - let me give you a simple example for you to argue about.
Lets assume that the Eur/USd rate is 2.00, That means the Usd/Eur rate is 0.50
We trade equal amounts (in money terms) of each currency pair.
We buy 1000 Eur/Usd @2.0 (value $2000)
We buy 4000 Usd/Eur @0.5 (value $2000)
Lets say Euro doubles so rates are now 4.0 and 0.25
We sell 1000 Eur/Usd @4 = $4000
We sell 4000 Usd/Eur @0.25 = $1000
That gives us a profit of $1000
If the Euro halves against the dollar rates will be even...
Well sell 1000 Eur/usd @ 1.0 = $1000
We sell 4000 Usd/Eur @ 1.0 = $4000
That also gives us a profit of $1000
Lets assume that the Eur/USd rate is 2.00, That means the Usd/Eur rate is 0.50
We trade equal amounts (in money terms) of each currency pair.
We buy 1000 Eur/Usd @2.0 (value $2000)
We buy 4000 Usd/Eur @0.5 (value $2000)
Lets say Euro doubles so rates are now 4.0 and 0.25
We sell 1000 Eur/Usd @4 = $4000
We sell 4000 Usd/Eur @0.25 = $1000
That gives us a profit of $1000
If the Euro halves against the dollar rates will be even...
Well sell 1000 Eur/usd @ 1.0 = $1000
We sell 4000 Usd/Eur @ 1.0 = $4000
That also gives us a profit of $1000
The link to the calculations from my first posting can not be downloaded anymore due to restrictions from rapidshare. This is why I uploaded the file here:
http://docs.google.com/fileview?id=...jYtZWViZi00ZDMyLWFhYjQtNzBmNjc3NDhhNDFk&hl=en
Please, before doing more calculations with only 1 or 2 price movements, take a look at this file.
Column A is a random number, 0 or 1.
Column B is eur/usd, it doubles when the random number is 1 and it halves when the random number is 0.
Column C is usd/eur, it is is always 1/(eur/usd). This column is not really necessary, but I hope it makes things more clear.
Column D is the amount of dollars which the first portfolio holds on every step. It is always 1/2 of the whole portfolio.
Column E is the amount of euros which the first portfolio holds on every step. It is always 1/2 of the whole portfolio multiplied by usd/eur.
Column F is the balance of Portfolio 1, which maximizes dollar returns. It measures how the strategy performs in dollars. It shows how much dollars Portfolio 1 has after each movement. This is what many people should look at before posting here calculations of 1 or 2 movements.
Column G, H and I are identical to D, E and F. They are also not really necessary, but I hope they make things more clear. They show how the system would perform if we measure the balance in euros.
http://docs.google.com/fileview?id=...jYtZWViZi00ZDMyLWFhYjQtNzBmNjc3NDhhNDFk&hl=en
Please, before doing more calculations with only 1 or 2 price movements, take a look at this file.
Column A is a random number, 0 or 1.
Column B is eur/usd, it doubles when the random number is 1 and it halves when the random number is 0.
Column C is usd/eur, it is is always 1/(eur/usd). This column is not really necessary, but I hope it makes things more clear.
Column D is the amount of dollars which the first portfolio holds on every step. It is always 1/2 of the whole portfolio.
Column E is the amount of euros which the first portfolio holds on every step. It is always 1/2 of the whole portfolio multiplied by usd/eur.
Column F is the balance of Portfolio 1, which maximizes dollar returns. It measures how the strategy performs in dollars. It shows how much dollars Portfolio 1 has after each movement. This is what many people should look at before posting here calculations of 1 or 2 movements.
Column G, H and I are identical to D, E and F. They are also not really necessary, but I hope they make things more clear. They show how the system would perform if we measure the balance in euros.
Ok, if your mother tongue is English and you say that I can't use "bet" in this sense, I agree with you
This game exactly represents the payoffs of a person who buys something in a random walk world and waits until the price either rises 100% or decreases 50%.
If you play that game all day long and each time your "tying up capital" is 100% of your capital, you will not achieve anything. This is why I gave this example to MathAndLogic, because he said that the Kelly bet should be 100%.
But if I understand you correctly, you say that you will "tie up" only a fraction of your capital. This would be right thing to do, but the optimal thing to do is to "tie up" every time what the Kelly formula says - half of your capital. If you play like this, your profits in the long run will be optimal and I wouldn't play this game against you
This is incoherent because when you "buy" EUR/USD you are actually simultaneously buying euros and selling US dollars. When you "buy" USD/EUR you are simultaneously buying US dollars and selling euros.
So in your first transaction you buy 1000 euros and sell 2000 dollars. In your second transaction you buy 4000 US dollars and sell 2000 euros. You now have a net position of:
short 1000 euros
long 2000 dollars
If the euro goes up to $4 you lose $2000.
If the euro goes down to $1 you gain $1000.
No surprises here.
sambian, I love the spreadsheet. I added a column to convert the euro account to usd, then another column to sum both account totals in usd.
Then I recalc'ed the table to generate new random numbers, and watched my ending balance. I like it when it said I had $5 billion! But I saw that roughly 1 in 20 recalcs showed a net loss. Not account wipeout, but less than starting capital. My formal math is not strong enough to prove it, but my belly tells me that a 5% chance of losing money seems about right with your method.
(and about a 30% chance of making > $100 million, and around 1% chance of making $5 billion. This is based on me counting on my fingers with a ridiculously small sample set of a few hundred trials. I would happily throw $2k at this strategy if there was a 1% of coming out with $5 billion however
In a practical sense, there aren't any currency pairs that bounce between 50% gain/loss fast enough for this to be a money maker in a lifetime. I wonder if we can still flesh out a viable trading method?
Then I recalc'ed the table to generate new random numbers, and watched my ending balance. I like it when it said I had $5 billion! But I saw that roughly 1 in 20 recalcs showed a net loss. Not account wipeout, but less than starting capital. My formal math is not strong enough to prove it, but my belly tells me that a 5% chance of losing money seems about right with your method.
(and about a 30% chance of making > $100 million, and around 1% chance of making $5 billion. This is based on me counting on my fingers with a ridiculously small sample set of a few hundred trials. I would happily throw $2k at this strategy if there was a 1% of coming out with $5 billion however
In a practical sense, there aren't any currency pairs that bounce between 50% gain/loss fast enough for this to be a money maker in a lifetime. I wonder if we can still flesh out a viable trading method?
Hey sambian,
do you make money with your system?
Another consideration is that the market is not random. I agree that your system works well in ranging markets, but in trending markets it must lose.
I upload a file with the end of week EUR/USD data for about a year. Since you are very good at math, would you finish the file with your system in order to see the real results? You can consider 2 pips ask/bid sprerad.
Looking forward to your reply.
do you make money with your system?
Another consideration is that the market is not random. I agree that your system works well in ranging markets, but in trending markets it must lose.
I upload a file with the end of week EUR/USD data for about a year. Since you are very good at math, would you finish the file with your system in order to see the real results? You can consider 2 pips ask/bid sprerad.
Looking forward to your reply.
At the risk of stepping on sambian's toes, I will take a stab at it.
The Sambian Strategy (as I will now call it) doesn't work like that. It doesn't take a point in time and decide what to do, it watches price (without regard to the time it takes to move) and makes a decision to rebalance the accounts when price hits one of two thresholds.