Elite Trader's Gambler's Anonymous ETGA

Status
Not open for further replies.
Electric and others, we need a package that can backtest entries and exits, allowing us to change the range on pairs

Then we can get actual numbers

Example:

using a 30 pip "spread" how many pips can GBPUSD on average capture per day long / short?

change it to 40? what are the results now?

etc etc

anyone got any ideas?

I suppose I could do this in excel which is what I use for everything
 
Tripack,

These arethe "big guns" over at Oanda, discussing things way over my head. I have to read that damn thread several times and maybe I get 25% of it...(you would not believe the knowledge of those posters, I will not single anyone of them out, you guess)

Good to see you here Tripack :)

Michael B.


Quote from TriPack:

I think it is a fundamental assumption of this system that the sum of the back and forth movements is always greater than the sum of the straight line moves. Even if a market went 4500 pips in one direction, I would be willing to bet that it did more than 9000 pips of back and forth movement along the way, before or after (though if it were sudden it could cause a margin call before you had a chance to recover).. I haven't studied the price series you refer to so maybe I'm offbase. Here is an interesting thread on this subject:

http://www2.oanda.com/cgi-bin/msgboard/ultimatebb.cgi?ubb=get_topic;f=15;t=002782#000000
More...
 
I agree..a package would be awesome...Sympatico, states that some FXmath wiz's could probably trade this more effectively than he does. (funny, I use excel for everything too...)

I want to attempt to first "tune" this then Automate it next year after trading it manually for a year...with a little help from my trading friends :)

Meanwhile, I am going to try and get flat soon in the current system #1 (waiting for positive parity)

Guys, I definitely have not wasted this month!

I believe there is a program named, "Behold"..that can do this sort of testing. Anybody have it?

Michael B.


Quote from jasonjm:

Electric and others, we need a package that can backtest entries and exits, allowing us to change the range on pairs

Then we can get actual numbers

Example:

using a 30 pip "spread" how many pips can GBPUSD on average capture per day long / short?

change it to 40? what are the results now?

etc etc

anyone got any ideas?

I suppose I could do this in excel which is what I use for everything
More...
 
Quote from mogul:


IMO, much much much better to trade only in the direction of the underlying trend
More...

We seem to have a situation here where the following is occurring:

Take short term gains.

Allow losses to accumulate long term, hopefully to be covered by the short term gains in the short term, and in the long term hopefully the market will retrace and our long term losses will turn into short term gains.

An underlying assumption that at some point the market is going to cover old ground.

A requirement to size positions so small that a long term range can be covered.

And if it is executed correctly there is no need to figure out things like which way it is trending. We don't care because if executed correctly the system will do fine regardless of which way the market trends (if it trends at all).

If you compare that to the traditional model where you first have to figure out the direction of the trend I can see why ElectricSavant finds this method appealing. However as you and others state the execution has to be just right or you run out of funds before you have a chance to cover your losses by taking advantage of the market's natural tendency to have more small moves than large ones.

I'm working on my own concept that is actually an attempt to merge the two ideas but I haven't resolved how it should work out <yet>.
 
Tripack,

I wish you the best. If you decide to share, you are always welcome here in this Journal. I can always count on your honesty. Or maybe were lucky and you start another ET "hall of fame" Journal!

All the others here have truly contributed in a positive way...not one post here has been interpreted to be negative, quite to the contrary, there is much thought provoking stuff here, translated into "plain speak" for me... :)

Michael B.

P.S. I am curious, did you solve this riddle: If the EUR/USD and USD/CHF go in inverted directions while being long in both...have we defined infinity? if the EUR/USD goes up and up and up to infinity? How could the USD/CHF go down below zero? (would the anti-correlation percent change?)


Quote from TriPack:

We seem to have a situation here where the following is occurring:

Take short term gains.

Allow losses to accumulate long term, hopefully to be covered by the short term gains in the short term, and in the long term hopefully the market will retrace and our long term losses will turn into short term gains.

An underlying assumption that at some point the market is going to cover old ground.

A requirement to size positions so small that a long term range can be covered.

And if it is executed correctly there is no need to figure out things like which way it is trending. We don't care because if executed correctly the system will do fine regardless of which way the market trends (if it trends at all).

If you compare that to the traditional model where you first have to figure out the direction of the trend I can see why ElectricSavant finds this method appealing. However as you and others state the execution has to be just right or you run out of funds before you have a chance to cover your losses by taking advantage of the market's natural tendency to have more small moves than large ones.

I'm working on my own concept that is actually an attempt to merge the two ideas but I haven't resolved how it should work out <yet>.
More...
 
I just coded excel to run a test on GBPUSD

using a "every 30 pip" long/short system since 14/02/2005 - using 5 mins DATA, I came up with the following result

(including spread in calculations of 4 pips)

the average daily yield on GBPUSD was 202.4 pips

the key here is to figure out which is the highest possible system number, because when this trends against you in a BIG way, you will get a smaller unrealized loss (less "trapped" trades)

also a larger pip number of the method will be less work, less trades per day


UPDATE: 40 pip method yielded 152.72 pips per day

UPDATE: 20 pip method yielded 257.45 pips per day


so IMO, the 40 pip method would be the best on GBPUSD, because we have 25% less pips than say the 30 pip method, with 33% less drawdown damage - PLUS it will be the least work



and thats how this thing would have to be logically worked out IMO



Another update: 100 pip method yielded 58.18 pips per day

so that has a yield 38% as good as the 40 pip method

but it has a drawdown damage of 40% of the 40 pip method

so thats a close call - the end result should be the same - but just way less work

what you all think?
 
Jason,

This will be exhaustive work for you...Start at 10 pips and go to 40....(maybe more). How would you incorporate Pivot Points into this?


Michael B.

Quote from jasonjm:

I just coded excel to run a test on GBPUSD

using a "every 30 pip" long/short system since 14/02/2005 - using 5 mins DATA, I came up with the following result

(including spread in calculations of 4 pips)

the average daily yield on GBPUSD was 202.4 pips

the key here is to figure out which is the highest possible system number, because when this trends against you in a BIG way, you will get a smaller unrealized loss (less "trapped" trades)
More...
 
Quote from ElectricSavant:

Tripack,

I wish you the best. If you decide to share, you are always welcome here in this Journal. I can always count on your honesty. Or maybe were lucky and you start another ET "hall of fame" Journal!

All the others here have truly contributed in a positive way...not one post here has been interpreted to be negative, quite to the contrary, there is much thought provoking stuff here, translated into "plain speak" for me... :)

Michael B.

P.S. I am curious, did you solve this riddle: If the EUR/USD and USD/CHF go in inverted directions while being long in both...have we defined infinity? if the EUR/USD goes up and up and up to infinity? How could the USD/CHF go down below zero? (would the anti-correlation percent change?)
More...

First the PS. In reality since being long EUR/USD is long EUR and Short USD, in order for this to go to infinity, either EUR would need to go to infinity or USD would need to go to zero. I'll assume that USD is not going to go to zero so that only leaves EUR going to infinity.

USD/CHF can only go to infinity if either USD goes to infinity or CHF goes to zero. I'll assume CHF is not going to zero, so USD would have to go to infinity for this to work out.

Of course the two positions are mostly offsetting on the USD portion, so it is really EUR/CHF with a minor USD. Is CHF going to zero? Probably not. Is EUR going to uncorrelate with CHF? That could happen to some extent but as long as Switzerland is in Europe CHF will move in a similar manner to EUR. Is EUR going to infinity? Has anything ever gone to infinity?
 
Status
Not open for further replies.
Back
Top