My TRO EAs

Moderator: moderators

User avatar
newschool
rank: 150+ posts
rank: 150+ posts
Posts: 489
Joined: Fri Aug 21, 2009 2:26 am
Reputation: 0
Gender: Male

Postby newschool » Tue Sep 22, 2009 4:11 pm

Great contribution at120.

Obviously there are way too much false entries to me .

1. This is the only thing I dont get about those EAs, the backtesting while using a custom indicator can be very slow. Considering this one use an H1 timeframe, the everytick method is not helping either.
For an approximation you can try the ControlPoints method.

2. I added a MaxSpread setting, to not enter a trade while there is too much volatily (news). The point is if we are IN the trade before the news, then its allright :) but we dont want to enter during the release.
But I think this will only work during live testing.

3. Added the range for the last 5 candles , default is 40 pips, you can change it in the code.

4. Played with the default gap between trades or hours.

5. Uploaded as version 5, for me it reduce drastically the amount of false entries, the trick now is not to reduce good entries....

Image

Just a note, this is a reversal EA, this means the market --should not-- know where it is going, it should do cycles. So from the 29Jul to 3Aug, it is obvious there is a bullish trend hence the low performance of the system...

Btw , I see you have a special broker ;)

Patch,
if you look at my last posts you can read exactly what are my entries, 3sema H1 delayed, 2&3sema on M1... etc
I dont have yet the zlines indicator in this EA, I prefer to wait till the coding of the semas isn't beta.

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

User avatar
JESGPY
rank: 150+ posts
rank: 150+ posts
Posts: 289
Joined: Fri Dec 19, 2008 7:21 pm
Reputation: 0
Location: Asuncion
Gender: Male
Contact:

Postby JESGPY » Tue Sep 22, 2009 7:21 pm

Hey newschool nice work.

I know that the indicators can be added to the EA code, so you dont need the indicator to run the EA.

keep the good work!

JUAN

User avatar
forexbob
rank: 150+ posts
rank: 150+ posts
Posts: 263
Joined: Mon Apr 20, 2009 9:52 am
Reputation: 15
Location: Planet Earth
Gender: None specified

Postby forexbob » Wed Sep 23, 2009 12:08 pm

hi john,

do not know all the details, but newschool coding the semafor ea is trading reversals. which in my opinion is a hard task.

i follow the thread out of interest for EA's and being a programmer myself.

but i do not trade with an EA, just manual. following the 3clipz method, and on this moment just trading breakouts shortly before london open
(breakout of the 6 semafor)

I started with dtb, for a while traded reversals. but changed because to my idea its easier to trade breakouts then guessing tops/bottoms and go
for the reversal.

and i also just would call myself a beginner.
just by looking at price action learning a lot
The highest Forex rebates: http://www.cashbackforex.info

User avatar
newschool
rank: 150+ posts
rank: 150+ posts
Posts: 489
Joined: Fri Aug 21, 2009 2:26 am
Reputation: 0
Gender: Male

Postby newschool » Thu Sep 24, 2009 3:19 am

Tom,

when you look at the charts the sema indicator seems absolutely spot-on for every highs and lows, but in reality there is repainting. As you can see resulting in many (now some) false entries.

The breakout EA wont be THAT much complicated, its just that I dont want to code 2 things at the same time...

Forexbob,

yes I understand your point about guessing tops & bottom, but in the cyclic periods, this would be a perfect EA. Or as someone suggested, it could be used on futures as gold or oil.

=========
Updated the version5, corrected a non-noticeable bug in calculation of the range + tweaked some settings.

User avatar
forexbob
rank: 150+ posts
rank: 150+ posts
Posts: 263
Joined: Mon Apr 20, 2009 9:52 am
Reputation: 15
Location: Planet Earth
Gender: None specified

Postby forexbob » Thu Sep 24, 2009 7:35 am

pipingtom,

all attachments (also version 5) are attached to the second post
on page 1 of this thread.
The highest Forex rebates: http://www.cashbackforex.info

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

User avatar
newschool
rank: 150+ posts
rank: 150+ posts
Posts: 489
Joined: Fri Aug 21, 2009 2:26 am
Reputation: 0
Gender: Male

Postby newschool » Thu Sep 24, 2009 7:41 pm

at120 what do you mean about the ECN version? Just put 14 instead of 140 in the SL if you dont have 5 digits...

User avatar
newschool
rank: 150+ posts
rank: 150+ posts
Posts: 489
Joined: Fri Aug 21, 2009 2:26 am
Reputation: 0
Gender: Male

Postby newschool » Fri Sep 25, 2009 12:54 am

Ok I see. But I dont think it was a problem for the EA... I mean I had no errors when testing...

=====================================

Here I have something to choose for the EA, I think there are 2 global strategies :

1. We want the EA to take all the good trades, thus we use wide detection values , but we put a smaller s/l and a tighter exit. So we make more false entries, but at the end the average is positive.

2. We want the EA to take only the fewer very good trades, thus we watch for perfect conditions for an entry and then we use a bigger s/l. But if we miss the trade we don't re-try many times in the next minutes/hours. We just pass.

This is my dilemma at the moment...

jhtumblin
rank: 50+ posts
rank: 50+ posts
Posts: 113
Joined: Tue Jun 13, 2006 3:42 am
Reputation: 0
Location: Atlanta, GA
Gender: None specified

Postby jhtumblin » Fri Sep 25, 2009 8:52 am

Here I have something to choose for the EA, I think there are 2 global strategies :

1. We want the EA to take all the good trades, thus we use wide detection values , but we put a smaller s/l and a tighter exit. So we make more false entries, but at the end the average is positive.

2. We want the EA to take only the fewer very good trades, thus we watch for perfect conditions for an entry and then we use a bigger s/l. But if we miss the trade we don't re-try many times in the next minutes/hours. We just pass.


Just my opinion but I think you are approaching this the wrong way by ignoring money management rules, which in the long run will create losses.

1. If an entry is valid, it is valid correct? If it isn't then you need to throw this idea out altogether and go straight to number 2. If in fact it is valid, you leave the tight s/l in and use a risk/reward ratio which will give you a greater winning expectancy despite the winning percentage.

2. If you do have a way to filter the trades to take the "fewer very good trades," then use that instead and depending what your s/l is, you still have to use a target which gives you a winning expectancy determined by the expected winning percentage.

An entry signal should be the best signal you can find based on your entry strategy. It will not always be right, I can guarantee you that but you can't make any money if your EA never pulls the trigger either. This talk of finding the perfect signal is starting to sound like a grail search. Get an entry that is good enough, and tweak the managed side of the trade to create profits. We already know the semaphore isn't magical but that it can be used for valid signals, so if you treat it this way you will find what you're looking for faster.

User avatar
bredin
rank: 1000+ posts
rank: 1000+ posts
Posts: 1032
Joined: Mon Feb 16, 2009 10:41 pm
Reputation: 5
Gender: None specified
Contact:

Postby bredin » Fri Sep 25, 2009 9:42 am

I get better ea results basing sl on support/resistance lines, and setting position size accordingly.

If I move stops I move them on sup/res lines as well (fractals, pivots, past semas).

working on exits based on market conditions rather than a straight tp works better in the long run (IME) as static values of sl/tp can be invalidated by market conditions.

IME static sl just doesnt work well enough; a tight sl just means you lose less money more often :) a bigger sl might give the trade time to develop, but market conditions might change and invalidate your position while still in drawdown.

Its tough to get right, especially over the long term. Sometimes what appears to be a sensible and safe decision causes an otherwise profitable ea to hemorrhage :)

G.
Welcome to the era of "Guns and Maskies" Trading!
The previous message contains discretion. Viewer nudity is advised.
Insanity Industries Indicators Here
Insanity Industries on YouTube!

User avatar
TheRumpledOne
rank: 10000+ posts
rank: 10000+ posts
Posts: 12778
Joined: Sun May 14, 2006 9:31 pm
Reputation: 1332
Location: Oregon
Real name: Avery T. Horton, Jr.
Gender: None specified
Contact:

Postby TheRumpledOne » Sat Sep 26, 2009 12:36 am

FOR THE RECORD:

I DO NOT WRITE EAs.

THERE ARE NO TRO EAs.
IT'S NOT WHAT YOU TRADE, IT'S HOW YOU TRADE IT!

Please do NOT PM me with trading or coding questions, post them in a thread.

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


Return to “MetaTrader”