FPI - Fractional Product Inefficiency: The Impeccable Hedge
Moderator: moderators
- Nicholishen
- rank: 50+ posts
- Posts: 52
- Joined: Thu Oct 26, 2006 8:44 am
- Reputation: 0
- Real name: Nick
- Gender:
mistake...double post
Last edited by Nicholishen on Sat Nov 04, 2006 5:41 am, edited 1 time in total.
Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.
Thank you for your support.
- Nicholishen
- rank: 50+ posts
- Posts: 52
- Joined: Thu Oct 26, 2006 8:44 am
- Reputation: 0
- Real name: Nick
- Gender:
i am a slave to my employer. he owns me, and i do not want to jeoparize my livelyhood. why did you join the forum (just today) for the purpose of breaking my balls? should i not share tools unless it contains source?
Last edited by Nicholishen on Sat Nov 04, 2006 9:31 am, edited 1 time in total.
Nicholishen wrote:i am a slave to my employer. he owns me, and i do not want to jeoparize my livelyhood. why did you join the forum (just today) for the purpose of breaking my balls? should i not share tools unless it contains source?
I join this forum just today because i just found this forum.
No you can share everything you want, i ask you for code just because it very simple and i can not start yours file.
I did start work on EA for this strategy, let see what i can do....
Good luck!!
- Gert Frobe
- rank: 150+ posts
- Posts: 177
- Joined: Fri Jun 30, 2006 9:51 pm
- Reputation: 0
- Location: Arlington, Texas
- Real name: ben arata
- Gender:
hi Nicholishen & EvgeniX on page 4 of this form there are two good mt4s one is by davidf and one is by eagel. both Jack(sorry, i can't resist Cap'n Jack) and David were very kind to us by taking the time to write these for us here at kresilk to use, and hopefuly make improvements on. if you come up with any improvements please post it here at kreslik.com so we can work togather and make the FPI better and better. remeber about Moore's Law and how all those small improments add up real fast.
Michal and all who have worked on this thank yall very much.
ben
Michal and all who have worked on this thank yall very much.
ben
I'm sure most have already noticed this but for those who haven't:
Each FPI number can only be calculated by the corresponding Bid/Ask price or the FPI number is meaningless.
So a BBS FPI is calculated by Ask * Ask * Bid & a SSB FPI by Bid * Bid * Ask.
This makes it impossible to test on historical data unless one has data with Bid/Ask.
Each FPI number can only be calculated by the corresponding Bid/Ask price or the FPI number is meaningless.
So a BBS FPI is calculated by Ask * Ask * Bid & a SSB FPI by Bid * Bid * Ask.
This makes it impossible to test on historical data unless one has data with Bid/Ask.
Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.
Thank you for your support.
Luke wrote:I'm sure most have already noticed this but for those who haven't:
Each FPI number can only be calculated by the corresponding Bid/Ask price or the FPI number is meaningless.
So a BBS FPI is calculated by Ask * Ask * Bid & a SSB FPI by Bid * Bid * Ask.
This makes it impossible to test on historical data unless one has data with Bid/Ask.
I never trust any BackTests!!!
Just make forward TEST!!
lot size
Hi Michal,
I'm very interested in your concept.
But I've a probem with sizing lot.
What would be the size for the following rings ( at the rates indicated )
AUDJPY ( 90.84) AUDUSD ( 0.7698 ) USDJPY ( 118.01 )
EURGBP ( 0.6689 ) EURJPY ( 150.08 ) GBPJPY ( 224.36 )
GBPCHF ( 2.3835 ) GBPUSD ( 1.9012 ) USDCHF ( 1.2538 )
GBPJPY ( 224.36 ) GBPUSD ( 1.9012 ) USDJPY ( 118.01 )
Thanks
Annie
I'm very interested in your concept.
But I've a probem with sizing lot.
What would be the size for the following rings ( at the rates indicated )
AUDJPY ( 90.84) AUDUSD ( 0.7698 ) USDJPY ( 118.01 )
EURGBP ( 0.6689 ) EURJPY ( 150.08 ) GBPJPY ( 224.36 )
GBPCHF ( 2.3835 ) GBPUSD ( 1.9012 ) USDCHF ( 1.2538 )
GBPJPY ( 224.36 ) GBPUSD ( 1.9012 ) USDJPY ( 118.01 )
Thanks
Annie
- michal.kreslik
- rank: 1000+ posts
- Posts: 1048
- Joined: Sat May 13, 2006 2:40 am
- Reputation: 36
- Location: Monte Carlo, Monaco
- Real name: Michal Kreslik
- Gender:
Luke wrote:Each FPI number can only be calculated by the corresponding Bid/Ask price or the FPI number is meaningless.
So a BBS FPI is calculated by Ask * Ask * Bid & a SSB FPI by Bid * Bid * Ask.
This makes it impossible to test on historical data unless one has data with Bid/Ask.
Precisely, Luke. However, the historical Bid/Ask prices can be recreated from either the Bid or Ask historical data if we know the average spread in points during the sample period. Needless to say, such a recreation is not perfect nor accurate, but at least it makes it feasible to perform the backtest with the spread in mind.
Michal
michal.kreslik wrote:Precisely, Luke. However, the historical Bid/Ask prices can be recreated from either the Bid or Ask historical data if we know the average spread in points during the sample period. Needless to say, such a recreation is not perfect nor accurate, but at least it makes it feasible to perform the backtest with the spread in mind.
Michal
Hello, Michal
This is good idea,
I did start work on EA for MT4, for this strategy.
And I have question for you, maybe you can share some results(statement) for this strategy?
Thank you
- michal.kreslik
- rank: 1000+ posts
- Posts: 1048
- Joined: Sat May 13, 2006 2:40 am
- Reputation: 36
- Location: Monte Carlo, Monaco
- Real name: Michal Kreslik
- Gender:
EvgeniX wrote:I never trust any BackTests!!!
Just make forward TEST!!
Hello, EvgeniX
Very simply put, the only difference between the backtest and forward-test is in the optimizing algorithm seeing the particular dataset sample when searching for the optimal solution or not.
Backtest (backward test) iterates thru the data sample until it reaches the iteration count constraint or until it finds the solution with the desired (or higher) fitness. So the backward data sample may be reused a zillion times. This fact alone makes it a perfect way to overfit your trading model to the training data. This overfitting begins at the point when the validation error stops dropping or starts rising while the training error continues to drop.
On the other hand, fortest (forward test) can only be done once on the data sample. Once you've done the fortest, this "forward" data sample can no longer be considered as an unseen (forward) sample. A brilliant way on how to combine the backtest and fortest systematically in trading models optimization is desribed in Robert Pardo's book Design, Testing, and Optimization of Trading Systems
Michal
Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.
Thank you for your support.