The Pouria EA Development Thread | Page 5 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

The Pouria Method 564 replies

Lessons Learned during System Development 29 replies

(pouria method II on the test) 218 replies

System Development Secrets 5 replies

Metatrader 4 Development Course 3 replies

Attachments: The Pouria EA Development Thread
Exit Attachments

The Pouria EA Development Thread

  • #81
  • Feb 11, 2007 8:41am Feb 11, 2007 8:41am
Quoting mikhaildgreat
mwilkinw, ilanr thank you thank you thank you GREAT GREAT GREAT Stuff! Now this could be applied to all the programs I am doing and using! the EA must be able to collect all the statistical data during forward testing, It must save it in the CSV file then use that to compute the optimal Lot size per trade. WOW! this is like the best money management system i ever came across! The more trades the EA makes the better it will be able to assign the optimal position size!

Thank you to both of you I will start working on this immediately!


Thank you once again,
Mikhail
MikhailDGreat,
I know this might sound ridiculous and too much to ask but could you do a reverse of the buy/sell order for the EA for me. I have an idea that I think might be beneficial in certain types of market conditions for certain pairs. I need it to buy when we should typically be selling and VV. Thanks
BigDog.
  • #82
  • Feb 11, 2007 9:12am Feb 11, 2007 9:12am
LOL!

No Problem, Here you go! tell us if you find it be profitable.

Regards,
Mikhail

Quoting JamDown
MikhailDGreat,
I know this might sound ridiculous and too much to ask but could you do a reverse of the buy/sell order for the EA for me. I have an idea that I think might be beneficial in certain types of market conditions for certain pairs. I need it to buy when we should typically be selling and VV. Thanks
BigDog.
Attached File(s)
File Type: mq4 Pouria EA Beta - Reversed.mq4   13 KB | 542 downloads
  • #83
  • Feb 11, 2007 9:22am Feb 11, 2007 9:22am
Hello Attached is the new version of the EA with a few minor but important bug fix. If you are forward testing the original version pls stop using the older version and use this instead.

Thanks,
-Mikhail
Attached File(s)
File Type: mq4 Pouria EA Beta 2.mq4   13 KB | 791 downloads
  • #84
  • Feb 11, 2007 9:23am Feb 11, 2007 9:23am
Quoting mikhaildgreat
LOL!

No Problem, Here you go! tell us if you find it be profitable.

Regards,
Mikhail
Mikhail,you are the man!!!!!!!!!

can't wait the new week start!!
  • #85
  • Feb 11, 2007 9:27am Feb 11, 2007 9:27am
Quoting mikhaildgreat
Hello Attached is the new version of the EA with a few minor but important bug fix. If you are forward testing the original version pls stop using the older version and use this instead.

Thanks,
-Mikhail
Thanks Mikhail,

Is it possible for you to document what changes have been made?

Thanks again.

BP
  • #86
  • Feb 11, 2007 9:36am Feb 11, 2007 9:36am
Quoting BearPaw
Thanks Mikhail,

Is it possible for you to document what changes have been made?

Thanks again.

BP
Not much really, just that the first version might generate false signals with MACD signals due to my mistake in the encoding.

-Mikhail
  • #87
  • Feb 11, 2007 9:42am Feb 11, 2007 9:42am
Quoting mikhaildgreat
Not much really, just that the first version might generate false signals with MACD signals due to my mistake in the encoding.

-Mikhail
Thanks....wonderful....we shall see how it goes!!

So, do you think the backtesting is not reliable? And we shall just see with the forward testing?

Thanks again.

BP
  • #88
  • Feb 11, 2007 9:50am Feb 11, 2007 9:50am
Quoting BearPaw
Thanks....wonderful....we shall see how it goes!!

So, do you think the backtesting is not reliable? And we shall just see with the forward testing?

Thanks again.

BP
I prefer manual backtesting, I dont trust metatraders Strategy tester. especially with small targets. If this EA had at least 50pip or 30pip SL and TP then strategy tester could generate more or less accurate results. But with such a tight SL and TP the results will be truely flawed. Just like cyberia trader which does insanely well during backtest with strategy tester but during actual trading it trades very very differently with poor results.

I believe its best to know the strategy then backtest using your eyes instead...

-Mikhail
  • #89
  • Feb 11, 2007 9:57am Feb 11, 2007 9:57am
Mikhail - awesome stuff.

the next thing is to get the managetakeprofitEA code into it so we can take partil profits and makethe mostof some of the big trends this method gets.
TRADE FOR SWAP
  • #90
  • Feb 11, 2007 10:17am Feb 11, 2007 10:17am
Quoting cgldsmth
Mikhail - awesome stuff.

the next thing is to get the managetakeprofitEA code into it so we can take partil profits and makethe mostof some of the big trends this method gets.
Yup I know but the question here is when to take profit? We need to find a better exit strategy to find the optimal time to exit.

If you or anyone has a good idea how to exit that would maximize the big trend please post it. because we cannot simply exit using when we get reverse entry signal because we will be exiting too late sacrificing profits as well...

-Mikhail
  • #91
  • Feb 11, 2007 10:55am Feb 11, 2007 10:55am
One way would be to backtest and find the average pips each signal gives on each pair. Although that's very laborious. How about a good old trail-lock?
TRADE FOR SWAP
  • #92
  • Feb 11, 2007 11:19am Feb 11, 2007 11:19am
Quoting cgldsmth
One way would be to backtest and find the average pips each signal gives on each pair. Although that's very laborious. How about a good old trail-lock?
we can give it a try... will post EA with trail stop tomorrow.

-mikhail
  • #93
  • Feb 11, 2007 11:41am Feb 11, 2007 11:41am
Well I can think of two exit strats that would work well with this approach.

1. use fibs off of the tunnel just like Vegas style

2. use ATR to determine exits in the same style as Turtle trading.
  • #94
  • Feb 11, 2007 11:48am Feb 11, 2007 11:48am
so divide the trade into 3, then exit at 55, 89, and 144 pips?
TRADE FOR SWAP
  • #95
  • Feb 11, 2007 11:50am Feb 11, 2007 11:50am
Correction, fibs off the fast moving ma not the tunnel, because price is usually no where near the tunnel

I was more thinking about looking at each currency and seeing what fib level price tends to congregate and using that level, but I see no reason why you could not have multiple exits... The only problem with this, is it does not maximize the potential profit, but it does make for a larger number of profitable trades.
  • #96
  • Edited 2:01pm Feb 11, 2007 1:39pm | Edited 2:01pm
Quoting mikhaildgreat
Hello Again,I just finished reading the paper about the Kelly system sounds very interesting and very relevant to our goals as traders. But to be honest i didnt understand a single thing with all those formulas. I have to admit I'm not very good in calculus with all those F'(x) and stuff made my head numb. were you able to understand the paper and are you able to apply it? I really would like to learn but not with calculus language. I'm very good with mathematical logic, thats why im a programmer, But when it comes to those kind of formulas i suddenly get dyslexia.

mwilkinw can you kindly explain as simple as possible the practical approach of using the formulas when it comes to forex trading? I think this is very important, it deals with position sizing to maximize investment growth but I just cant undertand it.

-Mikhail
Hi, after reading Fortunes Formula and Ed Thorp's mathematics of gambling, this is how I have interpreted Kelly Criteria. I put it in a small spread sheet.

Don't know if it will help you with this, but I think you can see what the formula is getting at...
The problem is that this style of MM is as aggressive as is possible, so there is a lot of risk of loss. Just not a great risk of ruin, because you can always get closer to zero. Mathematically it is true, but when your account is $10 the point is esoteric at best.

Here is a thread that talks about it at length and this post in particular
http://www.forexfactory.com/showpost...2&postcount=16
Attached File(s)
File Type: zip Kelly Risk Percentage.zip   2 KB | 548 downloads
  • #97
  • Feb 11, 2007 2:09pm Feb 11, 2007 2:09pm
Quoting mikhaildgreat
mwilkinw, ilanr thank you thank you thank you GREAT GREAT GREAT Stuff! Now this could be applied to all the programs I am doing and using! the EA must be able to collect all the statistical data during forward testing, It must save it in the CSV file then use that to compute the optimal Lot size per trade. WOW! this is like the best money management system i ever came across! The more trades the EA makes the better it will be able to assign the optimal position size!

Thank you to both of you I will start working on this immediately!

Thank you once again,
Mikhail

Hey, No Problem, Your Welcome! THANK YOU for your EA's you're the man! Like you I usually do all of my backtesting manually.. there are always trades that your eye would take that the automatic back test misses.

I calcualte my percentages manually and then implement them in "half kelly".

If you really give it some study, you'll see that if your profit% is 20-25% you'll still be very profitible, if you keep your betting size at your kelly% and add to your positions based on your increased equity using the kelly formula. That's how gamblers win.. because their win percentage is around 5%.. with a good trading strategy you can desecrate those numbers!

Steven A. Cohen from SAC Capital Partners uses the half-kelly and he's at 25 profit%. Needless to say he's aggressive and very strict with his system!

Happy Trading!
  • #98
  • Feb 11, 2007 2:17pm Feb 11, 2007 2:17pm
Quoting smjones
Hi, after reading Fortunes Formula and Ed Thorp's mathematics of gambling, this is how I have interpreted Kelly Criteria. I put it in a small spread sheet.

Don't know if it will help you with this, but I think you can see what the formula is getting at...
The problem is that this style of MM is as aggressive as is possible, so there is a lot of risk of loss. Just not a great risk of ruin, because you can always get closer to zero. Mathematically it is true, but when your account is $10 the point is esoteric at best.

Here is a thread that talks about it at length and this post in particular
http://www.forexfactory.com/showpost...2&postcount=16

smjones! I like the spreadsheet... an easy way to calculate. Although this is the normal kelly percentage... you can change your leverage to receive a half-kelly result!

I use 80%-Kelly Value... can that be input into excel?

Thanks for the tool!

Happy Trading!
  • #99
  • Feb 11, 2007 3:27pm Feb 11, 2007 3:27pm
Quoting mwilkinw
smjones! I like the spreadsheet... an easy way to calculate. Although this is the normal kelly percentage... you can change your leverage to receive a half-kelly result!

I use 80%-Kelly Value... can that be input into excel?

Thanks for the tool!

Happy Trading!
Hello mwilkinw,

I'm trying to integrate kelly in the code of one of my programs but I have a cercern though. when you compute the kelly% and you plug in the values for the (win / loss) do you plug in the values for the trade you are about to make? TP and SL? or does the Win / loss in computing kelly% refer to the average win and loss per trade produced by your system during previous trades? I ask because most of the systems don't have fixed TP and SL value.

Thanks Again,
Mikhail

PS. Sorry if my question is really noobish I'm still trying to absorb the concept, thanks for your patience.
  • #100
  • Feb 11, 2007 3:40pm Feb 11, 2007 3:40pm
Quoting mikhaildgreat
Hello mwilkinw,

I'm trying to integrate kelly in the code of one of my programs but I have a cercern though. when you compute the kelly% and you plug in the values for the (win / loss) do you plug in the values for the trade you are about to make? TP and SL? or does the Win / loss in computing kelly% refer to the average win and loss per trade produced by your system during previous trades? I ask because most of the systems don't have fixed TP and SL value.

Thanks Again,
Mikhail

PS. Sorry if my question is really noobish I'm still trying to absorb the concept, thanks for your patience.
That is actually and interesting question. For myself, I interpret it to mean the average win ratio the average pip win and the average pip loss, but when the formula was conceived, it was originally used to get the maximum bet for inside information about horse races. So, in it's original form it was applied to a specific horse race and a specific bet..

The reason I chose to apply it to an average, is because there are so many trades in FOREX and the idea is to get an average win ratio and average wins and losses. Thereby no individual trade have a devastating effect on the account balance... Just a personal preferance.
0 traders viewing now