FPI - Fractional Product Inefficiency: The Impeccable Hedge

NeoTicker indicators

Moderator: moderators

User avatar
michal.kreslik
rank: 1000+ posts
rank: 1000+ posts
Posts: 1032
Joined: Sat May 13, 2006 2:40 am
Reputation: 0
Location: Monte Carlo, Monaco
Real name: Michal Kreslik
Gender: Male
Contact:

Postby michal.kreslik » Tue Aug 28, 2007 6:19 am

TheEconomist wrote:Can you trade something that doesn't exist?

Like these rings: CL/$, WTI/$, CL/WTI ; or CL/$, KE/$, CL/KE

Those crosses (oil among themselves, oil to wheat...do they exist)? If no, how do you close the ring?


TheEconomist,

you can trade value1/value2 pairs if there exist interconnected direct quotes between these two values.

For instance, if you want to trade 1 barrel of oil / 1 ounce of gold and no one is quoting this value pair directly, you can look around and find out that there exists 1 barrel of oil / 1 USD quote as well as 1 ounce of gold / 1 USD quote.

Then, if you want to be long oil / gold, you simply go long oil/USD and short gold/USD having the position size constant based on the number of USD units.

Having said the above, you can close the above incomplete FPI ring that includes oil, gold and USD, by tapping into another commodity market. For instance, you can do this by opening positions in oil/EUR and gold/EUR having the position sizes constant based on the number of commodity units.

You can see that if there were no currencies available, you would not be able to trade oil against the gold as there would be no medium of exchange between these two commodities. And that's also the beauty of currencies: most people think money is a commodity by itself. But money is only worthwile if you can use it as a medium of exchange. And also, it's only worthwile if enough people believe they can use a particular medium of exchange to exchange their commodities and services.

Remember that Charles II of England was using wooden sticks as a medium of exchange. As enough people believed this is the currency, it worked as a currency.

Now do you think a paper rectangle with something printed on it (a dollar bill, for instance) possesses a higher value by itself than a wooden stick? :lol:

Michal
Last edited by michal.kreslik on Tue Aug 28, 2007 6:28 am, edited 1 time in total.

Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.

User avatar
michal.kreslik
rank: 1000+ posts
rank: 1000+ posts
Posts: 1032
Joined: Sat May 13, 2006 2:40 am
Reputation: 0
Location: Monte Carlo, Monaco
Real name: Michal Kreslik
Gender: Male
Contact:

Postby michal.kreslik » Tue Aug 28, 2007 6:27 am

dbw451 wrote:Hi Michal,

I'm a long time IB user. I'm not overly familiar with your FPI strategy and whether or not it is dependant on accurate real-time quotes. IB does not broadcast accurate tick data, but rather their quotes are timed snapshots. The snapshots leave out a lot of ticks. This leads to some interesting situations like my orders often are filled a second or so before my quote/chart displays a touch on my order price. The advantage of their quote update strategy is that they keep up well with fast moving markets (unlike Tradestation which sometimes falls several seconds behind). The obvious disadvantage is that you don't get every tick which could make arbitrage difficult.

Also, if you need historical data for backtesting or backfilling charts, be aware that the lowest pratical timeframe IB provides is in 1 minute bars.

IB is a good broker but if you need accurate real-time or historical tick data, you'll need to supplement with another source like eSignal.

Regards,

David


David,

thanks for the information. I'm now collecting data from IB as well as from EFX for use in one project and I'm definitely going to test my automated FPI framework on this data as well. I was using the Gain capital data ticks that are freely available, but Gain having their spread marked up, I found the Gain data utterly worthless for any FPI calculations.

Concerning the IB ticks vs. snapshots: does that mean that you are occasionally being filled with prices that were never fed by IB itself?

Michal

User avatar
dbw451
rank: 50+ posts
rank: 50+ posts
Posts: 50
Joined: Sun Jul 22, 2007 9:45 pm
Reputation: 0
Gender: Male

Postby dbw451 » Tue Aug 28, 2007 11:39 am

Michal,

Yes, I have had fills that did not show up on my charts. In two cases that I remember, my stops were hit and my charts did not show the bars touching my price. In both cases the market was fast and volatile. I called IB and in both cases and they verified that my stops were hit. Later I forced backfills on my charts and the bars in question showed their proper highs and lows. So quotes you collect in real-time could differ from IB's historical data (which is a separate data stream from a separate database).

My experience is with futures quotes. The FX quotes could be different, but I would assume IB takes the same approach. They are more concerned with the speed of their software than the tick accuracy of their quotes. I think this tradeoff is acceptable to most styles of trading (but I'm not sure about arbitrage).

David

juangecko
rank: <50 posts
rank: <50 posts
Posts: 4
Joined: Fri Aug 31, 2007 9:36 pm
Reputation: 0
Location: Buenos Aires
Gender: Male
Contact:

Postby juangecko » Thu Sep 13, 2007 4:25 pm

thanks for the information. I'm now collecting data from IB as well as from EFX for use in one project and I'm definitely going to test my automated FPI framework on this data as well. I was using the Gain capital data ticks that are freely available, but Gain having their spread marked up, I found the Gain data utterly worthless for any FPI calculations.

Concerning the IB ticks vs. snapshots: does that mean that you are occasionally being filled with prices that were never fed by IB itself?

Michal[/quote]

Michal, I wanted to ask you if FPI arb works in fx brokers. I heard that one of the problems is that you might requotes or other issues (in bucketshop brokers) or with ECN brokers (as mentioned above) you might not very good fills.
What is your personal experience and what broker is the best to do this.
Thanks

Juan

User avatar
dbw451
rank: 50+ posts
rank: 50+ posts
Posts: 50
Joined: Sun Jul 22, 2007 9:45 pm
Reputation: 0
Gender: Male

Postby dbw451 » Thu Sep 13, 2007 7:26 pm

Yes, I've been filled at prices that were not sent by IB and thus did not show on my charts. It's more of a rare event than an occassional event.

David

Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.

juangecko
rank: <50 posts
rank: <50 posts
Posts: 4
Joined: Fri Aug 31, 2007 9:36 pm
Reputation: 0
Location: Buenos Aires
Gender: Male
Contact:

Postby juangecko » Thu Sep 13, 2007 8:42 pm

Yep, thxs, but I wanted to ask you if FPI arb works in fx brokers. I heard that one of the problems is that you might requotes or other issues (in bucketshop brokers) or with ECN brokers (Interactive brokers, MBtrading,etc) you might get not very good fills.
Is FPI possible ?? and in what brokers?
Thanks

Juan

User avatar
stevegee58
rank: <50 posts
rank: <50 posts
Posts: 19
Joined: Mon Oct 01, 2007 7:34 pm
Reputation: 0
Gender: Male

Postby stevegee58 » Mon Oct 01, 2007 8:04 pm

Hi there Michal,

I just thought I'd let you and the other denizens of the board know that I've been experimenting with this on Oanda's FXGame and what I've experienced so far.

My experiments have centered around the EURUSD/USDJPY/EURJPY.

First, I naively entered my initial trades with the same number of units (1,000) and made money quickly. I was very excited so I closed out the trades for a good profit and re-opened immediately with the same units. I immediately lost money and waited a few days to see if I could get it back.

Finally I closed it out and re-read this whole thread. Turns out I needed to do something like:

Buy 1,000 EURUSD @1.4132
Buy 1,413 USDJPY @115.42
Sell 1,000 EURJPY @163.11

Those quotes were at the time I simultaneously entered the trades by hand.

This was more like it. I checked the "exposure" tab and saw I had 0 units of exposure in EUR and USD and about 20 JPY (almost 0). Upon entry I immediately had an unrealized loss of $0.34 which was about what I'd expect from the sum of the spreads on my positions. Over the course of the last few days I've watched the unrealized profit go slightly positive (about $0.10) and back negative again where it sits mostly. Mostly the unrealized is around -$0.20.

I've concluded that I did not enter at an optimal moment and should have waited. However I didn't have an easy way of observing the FPI at the time. I'm an IB customer and I'm aware of some API goodies available for their TWS. I've modified the Excel spreadsheet from IB (TwsDde.xls) that connects to the TWS using DDE (sorry for all the letters) to display the FPI in real-time. I'm just beginning to watch this value during the day to see what kind of range it has. There does seem to be some correlation between this value and the unrealized profit of my currently open positions at Oanda.

It's very interesting so far. My next step is to close out the current positions and re-open them when my Excel-based FPI shows a more favorable entry.
Last edited by stevegee58 on Mon Oct 01, 2007 8:41 pm, edited 1 time in total.

obx
rank: <50 posts
rank: <50 posts
Posts: 41
Joined: Fri May 26, 2006 12:35 pm
Reputation: 0
Gender: Male

Postby obx » Mon Oct 01, 2007 8:35 pm

Michal,

Have you done any work on identifying which member(s) of a ring is "source" of (ie main player(s) in) an inefficiency?

Thanks,

obx

User avatar
michal.kreslik
rank: 1000+ posts
rank: 1000+ posts
Posts: 1032
Joined: Sat May 13, 2006 2:40 am
Reputation: 0
Location: Monte Carlo, Monaco
Real name: Michal Kreslik
Gender: Male
Contact:

Postby michal.kreslik » Sat Oct 06, 2007 10:18 am

obx wrote:Michal,

Have you done any work on identifying which member(s) of a ring is "source" of (ie main player(s) in) an inefficiency?

Thanks,

obx


No pair can explicitly be declared as the source of inefficiency in the FPI ring. As Forex traders, we know that the majors are dragging the cross pairs behind. But from the mathematical point of view, it can't be said that some particular pair (for instance, a major pair) is the source of an inefficiency.

It's because the inefficiency in the FPI ring is formed as a product of all involved pairs' values. The pairs' values are measured against each other. Saying that some pair in an FPI ring contributes more to the inefficiency mathematically would be like saying that some particular vertex of a triangle contributes more to the total area of the triangle than any of the other two vertices.

Michal

User avatar
michal.kreslik
rank: 1000+ posts
rank: 1000+ posts
Posts: 1032
Joined: Sat May 13, 2006 2:40 am
Reputation: 0
Location: Monte Carlo, Monaco
Real name: Michal Kreslik
Gender: Male
Contact:

Postby michal.kreslik » Sat Oct 06, 2007 10:25 am

stevegee58 wrote:Over the course of the last few days I've watched the unrealized profit go slightly positive (about $0.10) and back negative again where it sits mostly. Mostly the unrealized is around -$0.20.


Steve,

your FPI ring behaves exactly as it is supposed to. Remember that the oscillation between overall profit and loss in an open FPI ring is inevitably skewed towards the negative values (loss) as there is the spread as well as the commissions involved.

Obviously, you should close the FPI ring once the overall P/L is positive. It's as simple as that.

Michal

Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.


Return to “NeoTicker indicators”