# Martingale EA - page 49

[Deleted]

measuring the ...well the "stochastic waves" with a lagging indicator may be what is needed here...Ralph...are you with me?

ES

Late Edit: Guppy will end up the best pair to trade!...ooops sorry going too fast here...getting ahead of myself.

1037

ralph.ronnquist:
That is, you should consider adding short on short, or long on long for a negative P (i.e., when the first trade is negative), because that leads to higher takeprofit over stoploss ratio. That's regardless of lot size, which "merely" affects which the resulting takeprofit and stoploss are.

I agree with this assessment. One thing to keep in mind, most people here are using a VERY large stop loss (basically none at all), in hopes the market will eventually move enough in the right direction to be successful.

ralph.ronnquist:
It may be worth for someone to verify these equations, because it is a fairly strong result.

I like this statement! Makes me think we're at least on the right track.

1037

ElectricSavant:
What interested me with the kayvan method, was its look at the previous candlestick...perhaps looking at the last two candlesticks could have a smoothing effect if the EA were not hardcoded for TF. ES

Looking back at ANY candles ties you to a TF. But so does an indicator. You have to use something.

1037

ElectricSavant:
looking back too far...may detract from the mathematical progression that this EA method wants to take.

Exactly. This is very much a "right in the NOW" EA.

[Deleted]

I mean that the "architecture" should remain open allowing the user to load the EA on any TF. If a user wants a larger frequencey of trades he could load it on the M15....if he wants less trades over a longer span...he could use the H1. If the TF is hardcoded into the EA then the user would not have any choices. If he wants more trades with lower TP's...he would have no choice, in the case of an EA that has the TF hardcoded.

We should attempt to present an EA that works in all TF's and gives the user a choice at which one.

There could be a input to lookback at a rolling 1, 2, or three candles to determine the high or low breaches...at these breaches there could be an input to use the averaging downmethod...or the pyramiding inmethod which are exact opposites.

This way wolfe, you present an EA that allows us to test various set files without you having to code and recode....

ES

wolfe:
Looking back at ANY candles ties you to a TF. But so does an indicator. You have to use something.
1037

ElectricSavant:
I mean that the "architecture" should remain open allowing the user to load the EA on any TF. If a user wants a larger frequencey of trades he could load it on the M15....if he wants less trades over a longer span...he could use the H1. If the TF is hardcoded into the EA then the user would not have any choices. If he wants more trades with lower TP's...he would have no choice, in the case of an EA that has the TF hardcoded.

We should attempt an EA that works in all TF's and gives the user a choice at which one.

There could be a input to lookback at a rolling 1, 2, or three candles to determine the high or low breaches...

ES

Yes ES, you already have this.

Let's look a some code-

This basically says this- The next buy is the high of the previous bar + the spread of the currency + 2 additional pips.(kayvan's idea)

High[1] will always return the value of the high price of the previous bar (no matter what TF you are attached to).

Now we know use this to enter a trade-

{

All the order sending stuff here...

Attach to whichever TF you want the EA to read the high or low of the bar previous the current one.

1037

ElectricSavant:
wolfe and others...

I do not think Ralph would participate if there was not something that kayvan hit upon.

I am using the present EA with low DD and many TP's thus far...I will stick with my tests there...and do not see it blowing up yet...next week will tell...

Meanwhile... a more stable and perhaps an EA with more choices for the various personalities out there might be polished and mounted...

Remember kayvan claims he has been trading his method profitably...for howw long I do not know...I suspect cable H! that is giving me such good results is the current preferred pair and TF..i simply do not know.

ES

I know, these are the best results you've had so far right? I'll still work on it.

[Deleted]

Yes Yes...I know...but when Ralph was considering the concepts this EA presents, I wanted to think out loud a little and I think we should preserve the TF openess...in the redesign..I was hoping that you and Dr. Ronnquist could work together.

I think my friends that I have in wolfe and Ralph are going to teach me how to code whether I know it or not...hehe..I am a stubborn old man...and there is more than one way to skin a cat! and...wolfe if you get the chance to work with Dr. Ronnquist...Take it!

I am sorry to say all of this below looks like Chinese to me...unless I was a copy and paste jockey, I would need to know what every space and every symbol and phrase meant...and why it was placed where it was...thus to explain this to me would be for another thread.

ES

wolfe:
Yes ES, you may not know, but you already have this.

Let's look a some code-

This basically says this- The next buy is the high of the previous bar + the spread of the currency + 2 additional pips.(kayvan's idea)

High[1] will always return the value of the high price of the previous bar (no matter what TF you are attached to).

Now we know use this to enter a trade-

{

All the order sending stuff here...
Hope you understand.
[Deleted]

On this EA...kayvan and you made Lcfx's TFX profitable for me..and I again watch it with interest.

kayvan states there are fill problems...and he wants a modified chain which extends the progression beyond a primitive doubling from the beginning...but regardless... the EA works in its present state with my settings thus far on Cable H1 with primitive doulbling and the stated errors.

If you work on this EA further, please scrole up and reread some of my suggestions for open architecture. This will minimize much future work for you wolfe. This thread may become one of the all time hall of fame threads in tsd...and we might just make some profit to boot...

ES

wolfe:
I know, these are the best results you've had so far right? I'll still work on it.
[Deleted]

to reiterate...now this is my opinion only...

Us Testers need from you coders the following....
• A candlestick lookback for breaches that we can set in the user friendly input menu.
• An Average Down or Pyramid Up method to choose in the user friendly input menu
• Number of PIPS breach zone that we can set in the user friendly input menu

This is what I have so far...and these proposed options present a myriad of possibilites to test in an idiot-savant way Wolfe you can do it!

ES