FPI - Fractional Product Inefficiency: The Impeccable Hedge

NeoTicker indicators

Moderator: moderators

bwilhite
rank: <50 posts
rank: <50 posts
Posts: 5
Joined: Wed Nov 01, 2006 8:07 pm
Reputation: 0
Gender: None specified

Postby bwilhite » Thu Nov 02, 2006 2:56 am

michal.kreslik wrote:
bwilhite wrote:One of these latter pairs will need to be traded at other than a "round lot" size. How do I determine the proper lot size for my synthetic position? I think I know the answer, but I want to be sure.

I don't want any other information than that


Hello, bwilhit,

the answer is here:

http://kreslik.com/forums/viewtopic.php?p=2354#2354

Michal


A picture can be worth a thousand words...but not always. So, please be pedantic here with me. I would go long EUR/USD 1 lot and long USD/CHF a number of lots equal to the EUR/USD quote? Is that correct? That's the ratio shown in the example. (Which I did see, btw, I just need confirmation that I'm understanding it correctly.) Thank you for your patience.

By the way, options market makers do this kind of thing all the time...it's called a conversion. But it's actually simpler for me to understand over there because of this ratio issue. Once that's cleared up...I have other uses for this idea also.

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

bitcy
rank: <50 posts
rank: <50 posts
Posts: 5
Joined: Thu Nov 02, 2006 6:11 am
Reputation: 0
Gender: Male

Postby bitcy » Thu Nov 02, 2006 7:35 am

Hello everybody.

I am trader since many years and I brought many of my friends to forex market. One of them, new to forex, redirected me to this thread.

People! You are trading water! This is the best joke I ever read and I can't stop laughing. Stop looking for holygrail and start studying and learning.

This idea is not new, it was born with the market and it was written down later by John Nash in his group's theory, many years ago. If there is a group of guys (currency pairs) and one of the guys will go out of the grup, sooner or later he will come back, or the others will folow him. The point is that you never know which one of the two will be, but you can guess by looking to the reaction of group members. In forex this is called correlation theory, and there are plenty of indicators and tools trying to predict the movement of a pair by looking to the other pairs around. The idea is that the market is not moving instantly. If one big australian guy wants to buy one japonese company, he will sell a big amount of AUD and buy a big ammount of JPY to pay for it. This will push the price for the pair, the AUDJPY "guy" is getting out of the group. He can come back or or he will drag AUDUSD and USDJPY or all the others, with it, depends how big is the initiator of the movement. This is what you, peple here, call rings, sum=1, all the stuff.

Guess what: this is not working in real life. You, as small trader, will be the last to be able to take advantage of such movements. There are arbitrators, there are other big sharks.

For the initiator of the trend:

Please REVIEW YOUR MATH and stop cheating people. In your example there is a qty of about 25K jpy not hedged, and you make the 6$ proffit from this quantity ONLY. You can get the same red dot over the 1's line and get out of your trade sooner or later also if the prices move in the other direction (against JPY) and you will get out with a big lose. If you make the complete hedging as you show later in the other post with the nice colored table (with 126k insted of 100k, buy you say nothing about profit there) then your example is producing some dollars... LOSS... Have a look again to the calculus.

2. There is a tool on the web, called market gauge. I won't place here any link, because first I don't know the link, second I don't want people say that I advertize something. But this could be interesting for you, if you want to develop a trading strategy based on this method. This tool is considering four currencies, USD, EUR, JPY, GBP. It paints all the 6 possible pairs on the screen in the corners of a big hexagon and draw all the posible lines between them as a complete K6 graph. There are 15 lines at all, that have labels on them. These labels are computed, function of the exchange rates, and in normal market conditions, all the labels are 1. So the sum of all graph is 15. If one of the pairs run away, the values on the lines attached to that pair change to 1.1, 0.9 and so on, pretty much as in your method, in such a way that the sum of the graph is all the time 15. This is considering the EMH stating that the market is all the time in an equilibrum. You can set a "minimum difference" and if the difference between the highest line value and the lowest line value is higher then the minimum difference, that a buy/sell signal will be generated for the respective (3, 4, 5, 6) pairs, that is what you call here the perfect hedge. The minimum difference is like a parameter for distance to 1 in your FPI diagram. Nice in theory, nice in testing: it produces small profit all the time when tested on historical data. IT NEVER PRODUCE PROFIT in real time, or at least not enough profit to cover the spread. First of all, the minimum difference must be set enoug high to cover the spreads. This occur very seldom, maybe once in the year, on some bad events as 9/11. There are only some guys in the world able to make profit of this gauge: 6 or 7 central banks, about 10 funds, and about other 20 arbitrators. You, as end-line traders will never be able to make any profit of it. But you can try...

Happy and prosperous trading to all, and sorry for my japonese english. I dont think I will come back here soon, but reading your joke worth the time spent for this post...

User avatar
Nicholishen
rank: 50+ posts
rank: 50+ posts
Posts: 52
Joined: Thu Oct 26, 2006 8:44 am
Reputation: 0
Real name: Nick
Gender: Male

Postby Nicholishen » Thu Nov 02, 2006 7:47 am

bitcy wrote:Hello everybody.

I am trader since many years and I brought many of my friends to forex market. One of them, new to forex, redirected me to this thread.

People! You are trading water! This is the best joke I ever read and I can't stop laughing. Stop looking for holygrail and start studying and learning.

This idea is not new, it was born with the market and it was written down later by John Nash in his group's theory, many years ago. If there is a group of guys (currency pairs) and one of the guys will go out of the grup, sooner or later he will come back, or the others will folow him. The point is that you never know which one of the two will be, but you can guess by looking to the reaction of group members. In forex this is called correlation theory, and there are plenty of indicators and tools trying to predict the movement of a pair by looking to the other pairs around. The idea is that the market is not moving instantly. If one big australian guy wants to buy one japonese company, he will sell a big amount of AUD and buy a big ammount of JPY to pay for it. This will push the price for the pair, the AUDJPY "guy" is getting out of the group. He can come back or or he will drag AUDUSD and USDJPY or all the others, with it, depends how big is the initiator of the movement. This is what you, peple here, call rings, sum=1, all the stuff.

Guess what: this is not working in real life. You, as small trader, will be the last to be able to take advantage of such movements. There are arbitrators, there are other big sharks.

For the initiator of the trend:

Please REVIEW YOUR MATH and stop cheating people. In your example there is a qty of about 25K jpy not hedged, and you make the 6$ proffit from this quantity ONLY. You can get the same red dot over the 1's line and get out of your trade sooner or later also if the prices move in the other direction (against JPY) and you will get out with a big lose. If you make the complete hedging as you show later in the other post with the nice colored table (with 126k insted of 100k, buy you say nothing about profit there) then your example is producing some dollars... LOSS... Have a look again to the calculus.

2. There is a tool on the web, called market gauge. I won't place here any link, because first I don't know the link, second I don't want people say that I advertize something. But this could be interesting for you, if you want to develop a trading strategy based on this method. This tool is considering four currencies, USD, EUR, JPY, GBP. It paints all the 6 possible pairs on the screen in the corners of a big hexagon and draw all the posible lines between them as a complete K6 graph. There are 15 lines at all, that have labels on them. These labels are computed, function of the exchange rates, and in normal market conditions, all the labels are 1. So the sum of all graph is 15. If one of the pairs run away, the values on the lines attached to that pair change to 1.1, 0.9 and so on, pretty much as in your method, in such a way that the sum of the graph is all the time 15. This is considering the EMH stating that the market is all the time in an equilibrum. You can set a "minimum difference" and if the difference between the highest line value and the lowest line value is higher then the minimum difference, that a buy/sell signal will be generated for the respective (3, 4, 5, 6) pairs, that is what you call here the perfect hedge. The minimum difference is like a parameter for distance to 1 in your FPI diagram. Nice in theory, nice in testing: it produces small profit all the time when tested on historical data. IT NEVER PRODUCE PROFIT in real time, or at least not enough profit to cover the spread. First of all, the minimum difference must be set enoug high to cover the spreads. This occur very seldom, maybe once in the year, on some bad events as 9/11. There are only some guys in the world able to make profit of this gauge: 6 or 7 central banks, about 10 funds, and about other 20 arbitrators. You, as end-line traders will never be able to make any profit of it. But you can try...

Happy and prosperous trading to all, and sorry for my japonese english. I dont think I will come back here soon, but reading your joke worth the time spent for this post...


Dems fightin' words! :lol:

User avatar
michal.kreslik
rank: 1000+ posts
rank: 1000+ posts
Posts: 1047
Joined: Sat May 13, 2006 2:40 am
Reputation: 36
Location: Monte Carlo, Monaco
Real name: Michal Kreslik
Gender: Male

Postby michal.kreslik » Thu Nov 02, 2006 9:36 am

bitcy wrote:People! You are trading water! This is the best joke I ever read and I can't stop laughing. Stop looking for holygrail and start studying and learning.


bitcy, thanks for your "valuable" input. Could you please direct us to some of your works so that we may finally stop trading water and looking for holy grail? Please lead us to the light.

bitcy wrote:This idea is not new, it was born with the market and it was written down later by John Nash in his group's theory, many years ago.


As Confucius said some 2500 years ago, "No idea is really new". I've come up with FPI idea by myself about year ago, but it would be naive to think that nobody before me was ever thinking/working/writing about a similar idea.

By the way, Nash didn't write any "group theory". I guess you are referring to his famous Equilibrium points in n-person games work. I'm sure Confucius would be glad to know that even Nash's idea "was not really new". If you studied some more, you would know that the idea Nash got Nobel prize for in the nineties is actually sometimes called Nash-Cournot equilibrium. And I'm sure that if we went back to Mr. Cournot's time we would find out that even his idea "was not really new".

Also, it seems like you didn't get the basic principle of Nash's non-cooperative games: Nash equilibrium is a set of rules that produces the best total sum of outcomes for all the involved players in the game. This equilibrium is reached if no player would gain anything by changing his personal behavior without the others changing their behavior, too.

But I can't really see any direct link between Nash's non-cooperative games and the Forex market. In Forex market, the game players are not equally capable of playing the game, so, quite obviously, the big player has much to gain by changing his personal behavior at the moment without consulting his game strategy first with all the rest of the game players.

bitcy wrote:For the initiator of the trend:

Please REVIEW YOUR MATH and stop cheating people. In your example there is a qty of about 25K jpy not hedged, and you make the 6$ proffit from this quantity ONLY.

Now come on, bitcy :) In a later article, I corrected that obvious error - I was writing the article you are referring to deep in the night :smt015

bitcy wrote:If you make the complete hedging as you show later in the other post with the nice colored table (with 126k insted of 100k, buy you say nothing about profit there)

Correct! I say nothing about profit there, because if you look closely at that post, it does only show how to calculate the position sizes for FPI correctly.

If I were to say something about profit/loss, I would have to include not only the open transation, but the close transaction, too. I reckon 99.9% of the readers noticed there's only one table, not two.

bitcy wrote:Nice in theory, nice in testing: it produces small profit all the time when tested on historical data. IT NEVER PRODUCE PROFIT in real time, or at least not enough profit to cover the spread.


If you would kindly read my introductory article, you would notice my sentence "the success of using FPI in a real trading environment is much dependent on a fast and literally razor-sharp execution and low spread". Obviously, this condition is not met with the majority of retail FX brokers.

bitcy wrote:I dont think I will come back here soon


Too bad. I will be missing you badly.

Michal

User avatar
michal.kreslik
rank: 1000+ posts
rank: 1000+ posts
Posts: 1047
Joined: Sat May 13, 2006 2:40 am
Reputation: 36
Location: Monte Carlo, Monaco
Real name: Michal Kreslik
Gender: Male

Postby michal.kreslik » Thu Nov 02, 2006 1:19 pm

Hello,

I calculated the distribution of FPI values accross the timespan of 6 months from Dec/29/2005 to Jun/30/2006.

I employed this FPI ring in the statistics:
  • EUR/USD
  • GBP/USD
  • EUR/GBP
I used a set of 5 min historical close Bid data exported from Tradestation and filtered out all the obvious bad ticks. If the data for a particular date and time was incomplete (missing price for one or more pairs), I discarded all data for this date and time and didn't make up for the missing data in any way. The nature of the distribution statistics lends itself to discarding such a missing data since the order of data in a sample is irrelevant.

After this, there was a total of about 105 thousand valid Bid prices available for this research. It translated to about 35 thousand prices for each of the three FX pairs.

Unfortunately, I don't have any true Bid/Ask price history data at my disposal, so I constructed the Ask prices from the Tradestation historical Bid prices by adding a constant (average spread) to every Bid price as follows:
  • 0.0003 for EUR/USD
  • 0.0004 for GBP/USD
  • 0.0005 for EUR/GBP
Certainly, this is not the true spread, but it's the closest I could get to the real thing.

From the resulting set of Bid and Ask data I then calculated four FPI distribution samples:
  • BSS 011 sample: EUR/USD_Ask * (1/GBP/USD_Bid) * (1/EUR/GBP_Bid)
  • BSS 100 sample: (1/EUR/USD_Ask) * GBP/USD_Bid * EUR/GBP_Bid
  • SBB 011 sample: EUR/USD_Bid * (1/GBP/USD_Ask) * (1/EUR/GBP_Ask)
  • SBB 100 sample: (1/EUR/USD_Bid) * GBP/USD_Ask * EUR/GBP_Ask
The above 4 calculation variants represent 4 valid ways on how to calculate the FPI value for a given ring. I guess the samples' naming convention is self-explanatory; B stands for Buy, S for SellShort, 0 for natural multiplication, 1 for fractional multiplication.

From these samples I removed the statistical outliers (rare FPI values that are too far from the mean) and marshaled the data into the distribution diagram with a resolution of 100. For a better readability, I transformed the original FPI values as follows:
  • displayValue = 1 000 000 * (FPI - 1)
I calculated means, 1st standard deviations, kurtosis and skewness of the sample distributions of FPI values:










As you can see, the FPI values are all distributed normally, with a slightly leptokurtic kurtosis. The skew depends on the choosen variant of calculation (either 011 or 100).

It's also interesting that none of the distribution graphs shows FPI oscillating around FPI = 1 (displayed as 0).

Michal

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

ryan
rank: <50 posts
rank: <50 posts
Posts: 24
Joined: Wed Oct 18, 2006 9:57 am
Reputation: 0
Gender: Male
Contact:

Postby ryan » Thu Nov 02, 2006 3:03 pm

Michal, you graphs are very encouraging even if they do use data which has been 'articifically' created in part.

I have been wondering for a while if we can dervive a $ value from the FPI differential? Is this possible?

davidf
rank: <50 posts
rank: <50 posts
Posts: 24
Joined: Fri May 19, 2006 6:10 pm
Reputation: 0
Gender: Male
Contact:

Postby davidf » Thu Nov 02, 2006 3:03 pm

Michal, GREAT WORK again !!!
DavidF

makosgu
rank: <50 posts
rank: <50 posts
Posts: 8
Joined: Thu Oct 26, 2006 8:29 pm
Reputation: 0
Gender: Male

Postby makosgu » Thu Nov 02, 2006 4:01 pm

@bitcy...

I am so thankful for traders like yourself. Perhaps in japan, things are a bit different. Unlike yourself, I deal with transactions that are at a minimum $.5B a clip. That means, every time we hang up the phone, we have put a $.5B of risk on the line. Do you know what it is to trade $500M at a single click of the button? Do you understand the mentality that is required to look at such structures? What it all leads to is the whole Buffet issue. Every market has a saturation point for how much it can handle. It turns out that the institutions play in a very different arena then individual traders do. Where you play is dictated by how much capital you are looking to enter the market with at a sigle clip. Buffet's billion dollar company has great difficulty in finding arenas that can handle the amount of capital he has to put to work. In my personal trading account, I bound transactions at roughly six figures. My buddies bound their trading at 7 to 9 figures at a single clip. We never worry about the effects of our trades on each other simply because the capital amounts dictate where it is that you can safely play. Unfortunately, I don't see how it is that you do not get what we are talking about, considering your tenure. Any time there is a distribution, the information is tradeable. Why? VERY SIMPLE! In order to make profits, the market has to move. Without movement, no one profits. Michal's charts clearly show that there is a distribution of movement. The data is irrefutable. It is not possible to have rings perfectly move in step hence the opportunity. Put another way, imagine a crowd of thousands. If you tell everyone to take one step forward, you will HEAR a distribution of footsteps if you plot it with respect to time, you will find a mean time offset at where the majority of persons will have taken a step. This is the true heart of the matter in trading. We are talking about a mass of trades acting. Depending on which aspect of this mass you measure, it will show that there is a proportion of the mass that consistently lags and more importantly, that there is proportion that consistently lead. CORRELATION and proper analytics make this explictly clear when you set up the analysis properly. Folks like yourself give institutions too much credit. You may think that they do the math right but in actuality, they depend on folks like myself to build the system. How do I know this, because I've been doing this for institutions for the better part of a decade. Furthurmore, banks view forex very differently than your average forex daytrader.

So what you have brought up but missed is how to trade correlation. It is true that CORRELATION is the real culprit here but you have clearly missed what correlation tells you. If I remember correctly EUR/USD and USD/CHF have high negative correlation -.9. THis means the pairs move opposite one of another but in synch. Correlation is the degree to which movements are in synch. I mentioned earlier that I could easily double the lenght of this thread to make things as complicated and/or comprehensive as possible. Howevr, it was difficult to gauge how much thought folks had put in to the matter. In any event, so what of correlation for the moment? Well if you think about everyone taking one step at the same time, you will notice that there is a tendency for certain individuals to cosistently take the first step before the rest do. What you may have not considered is how you define a step for each asset. There is a neat realtime link that I monitor for real time correlation for dozens of pairs. It allows me to trade concepts that are orthogonal to FPI. However, since we aren't discussing these, I won't divulge and make tangential comments...

@Michal

I don't understand the rational for the displayvalue equation you have posted. I would ask to post the underlying data but I'm sure that violates some tradestation policy. Secondly BSS 100 and BSS 011 should show the similar distributions. In otherwords, it is counterintuitive as to how a 100 would give a different distribution then 011.

Example...
(a/c)=(a/b)(b/c)

011
(a/c)*(1/[a/b])*(1/[b/c])=1

100
(1/[a/c])*(a/b)*(b/c)=1

In other words...
whether calculating FPI as 011 or 100 should give the same exact FPI value. Nonetheless, we do find a distribution. What is difficult to gauge from your charts is the overlap between charts 1 and 3 and charts 2 and 4. Judging how quiet things are in the thread from earlier participants, either folks are trading what they know and keeping quiet for fear of ruin, or moved on to greener pastures. Just to put things into perspective, to really optimize you want to put all rings in perspective. As a result, there dozens and dozens of the above distributions at play in realtime. The idea is to continually be engaging in the extremes of ANY pair that is now at extreme. In other words, you want to be indiffferent to any given pair, and biased to ALL opportunities across all pairs that arise at any given moment. It is super advanced I admit. However, this is the material they don't talk about in literature and how I deal with the markets in the equities arena...

Nonetheless Michal, SUPER post.
Last edited by makosgu on Thu Nov 02, 2006 4:08 pm, edited 1 time in total.

makosgu
rank: <50 posts
rank: <50 posts
Posts: 8
Joined: Thu Oct 26, 2006 8:29 pm
Reputation: 0
Gender: Male

Postby makosgu » Thu Nov 02, 2006 4:02 pm

@ryan...

Yes, we can explicitly translate the FPI to a money value. I just haven't thought about the equation yet...


Umm sorry ryan. I take the statement back. FPI is a coefficient with no units since they all cancel in the equation. I THINK!!! :?:

ryan
rank: <50 posts
rank: <50 posts
Posts: 24
Joined: Wed Oct 18, 2006 9:57 am
Reputation: 0
Gender: Male
Contact:

Postby ryan » Thu Nov 02, 2006 4:25 pm

Makosgu,

Yes it is an interesting question. In theory where FPI=1 profit = $0, if by implication the FPI hedges 3 or more pairs relative to each other I am not convinced it isnt possible to attribute this to pips, unless this is impacted by the 'weighting' of each pair in the FPI hedge. I guess evaluating some examples will help us determine this.

BTW ... I have sent you a PM.

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


Return to “NeoTicker indicators”