stevehopwoodforex.com
https://www.stevehopwoodforex.com/phpBB3/
Print view

Holy Graily Bob 'n Dotty
https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=4892
Page 17 of 26
Author:  Maximilian [ Tue Jan 10, 2017 9:34 pm ]
Post subject:  Holy Graily Bob 'n Dotty

Hi,
yep, SL will be the way to go for me also,
We will see,
Max
Author:  SteveHopwood [ Wed Jan 11, 2017 2:23 pm ]
Post subject:  Holy Graily Bob 'n Dotty

Here are a couple of pics that continue the GJ hedged basket. A steep fall:
11_30.png
Consolidation:
11th Jan.png
You can see from the Position Monitoring indi that I have 17 buys and 15 sells open, but look at the upl difference between the two groups. I bet that the position would have closed at my hedged basket tp some time ago, had HGBnD not offset those early sells following each brief retrace.

This suggests to me that those of us using hedging do not want offsetting once we are hedged. Early days, so we shall see.

In the meantime, I await the market making up its mind again. A decent move in either direction will eventually close everything down.

:xm:
Author:  pakeha [ Wed Jan 11, 2017 6:16 pm ]
Post subject:  Holy Graily Bob 'n Dotty

Steve

I have been thinking about the hedged basket issue and while I don't have a "brilliant" idea, one question did cross my mind. What is the optimal grid size per pair?

Given the volatility of pairs is different (i.e. peak to trough differences) then is it worthwhile increasing the grid size on the more volatile pairs by 2x or 3x? This would of course mean that the hedging function would kick in only after a really significant move and give the algorithm a bit of breathing room on any given move up or down.

Just a thought.
Author:  Dovilo [ Wed Jan 11, 2017 7:46 pm ]
Post subject:  Holy Graily Bob 'n Dotty

The problem with trades opposite to the trend closing on a minimal profit is still here. Although I don't think it has to do with CloseOnOppositeSignalOrTrendChange parametr because I experienced it with it set on "true" and on "false".
Author:  DerPat [ Thu Jan 19, 2017 5:25 am ]
Post subject:  Holy Graily Bob 'n Dotty

pakeha » Wed Jan 11, 2017 7:16 pm wrote:Steve

I have been thinking about the hedged basket issue and while I don't have a "brilliant" idea, one question did cross my mind. What is the optimal grid size per pair?

Given the volatility of pairs is different (i.e. peak to trough differences) then is it worthwhile increasing the grid size on the more volatile pairs by 2x or 3x? This would of course mean that the hedging function would kick in only after a really significant move and give the algorithm a bit of breathing room on any given move up or down.

Just a thought.

I like your thought and I would like to add, that you even need session based average true range per pair, as every pair behaves different throughout the sessions. And especially at London Open the Asian session ATR gets killed. Or at New York open, the europe lunch break atr gets killed.
While it sounds nice to have an atr adjusted, based on session choosen and lookback period of days to consider, I do not know what a headache this would cause to a programmer and what the outcome would be.

EDIT:
Just had an idea for an overkill on this, example: Session based ADR for the last 5 thursdays for the respective pair.
Author:  Expbloom [ Sun Jan 22, 2017 11:28 pm ]
Post subject:  Holy Graily Bob 'n Dotty

Hi good day to everybody. I'm a newbie in this forum but do not have coding background. Don't know whether any one ever suggests in SHF to avoid / suspend for the HGBnD to place new order for any new pair despite the HGI signals for long / short as per below user variables :-

1) PauseOnHighImportanceNews = True / False
2) PauseMinPriorToHighImportanceNews = To stop the HGBnD trigger new order for new pair prior to high important news in minutes
3) PauseMinAfterHighImportanceNews = To stop the HGBnD trigger new order for new pair after high important news in minutes
4) PauseOnMediumImportanceNews = True / False (I prefer to ignore this medium important news)
5) PauseMinPriorToMediumImportanceNews = To stop the HGBnD trigger new order for new pair prior to medium important news in minutes
6) PauseMinAfterMediumImportanceNews = To stop the HGBnD trigger new order for new pair after medium important news in minutes
7) PauseOnLowImportanceNews = True / False (I prefer to ignore this low important news)
8) PauseMinPriorToLowImportanceNews = To stop the HGBnD trigger new order for new pair prior to low important news in minutes
9) PauseMinAfterLowImportanceNews = To stop the HGBnD trigger new order for new pair after low important news in minutes

We need normal market conditions but news are always in sentiment, sometimes act abnormal from technical analysis and price movements behave like roller coasters. Please bear in mind that the HGBnD will only pause on the relevant currency. For example, if the high important news is pertaining to GBP, only those 7 pairs of GBPXXX will be paused but not other pairs.

Of course, the HGBnD will yet fully "take care" on the existing pairs' open trades. Thanks
Author:  SteveHopwood [ Mon Jan 23, 2017 12:15 am ]
Post subject:  Holy Graily Bob 'n Dotty

Expbloom » Sun Jan 22, 2017 11:28 pm wrote:Hi good day to everybody. I'm a newbie in this forum but do not have coding background. Don't know whether any one ever suggests in SHF to avoid / suspend for the HGBnD to place new order for any new pair despite the HGI signals for long / short as per below user variables :-

1) PauseOnHighImportanceNews = True / False
2) PauseMinPriorToHighImportanceNews = To stop the HGBnD trigger new order for new pair prior to high important news in minutes
3) PauseMinAfterHighImportanceNews = To stop the HGBnD trigger new order for new pair after high important news in minutes
4) PauseOnMediumImportanceNews = True / False (I prefer to ignore this medium important news)
5) PauseMinPriorToMediumImportanceNews = To stop the HGBnD trigger new order for new pair prior to medium important news in minutes
6) PauseMinAfterMediumImportanceNews = To stop the HGBnD trigger new order for new pair after medium important news in minutes
7) PauseOnLowImportanceNews = True / False (I prefer to ignore this low important news)
8) PauseMinPriorToLowImportanceNews = To stop the HGBnD trigger new order for new pair prior to low important news in minutes
9) PauseMinAfterLowImportanceNews = To stop the HGBnD trigger new order for new pair after low important news in minutes

We need normal market conditions but news are always in sentiment, sometimes act abnormal from technical analysis and price movements behave like roller coasters. Please bear in mind that the HGBnD will only pause on the relevant currency. For example, if the high important news is pertaining to GBP, only those 7 pairs of GBPXXX will be paused but not other pairs.

Of course, the HGBnD will yet fully "take care" on the existing pairs' open trades. Thanks
So you joined on Friday and think you can improve one of the best EA's created by a combo of some of the best minds here?

Wow.

Clearly not read my Welcome message. Bad move.

You are consigned to the group of dimwits unable to do anything other than read posts and download attachments here.

I did warn you.

:xm:
Author:  youcrazykids [ Tue Jan 31, 2017 3:32 pm ]
Post subject:  Holy Graily Bob 'n Dotty

Dear Tomele,

I followed with interest your "rat race" thread, and its trial of the different HGBn# variants, particularly how their settings compared to this version. I am only replying in this thread because I feel it is relevant, and anyhow the original "rat race" thread is locked.

Might I enquire about the settings being used in this version of HGBnD, in relation to the settings used during your trial? Please see a small comparison attached showing some differences (This thread's version is on the left, the "rat race" version on the right).

I am interested to know your rationale behind some of the choices, for example: the timeframe, use of a stop loss, use of offsetting... were these deliberate choices, random choices, or just the original settings that your trial version possessed? No matter how they were made, they did fairly well!

I realise that your "rat race" version on trial was a slightly older version than that currently in use here, nevertheless, it could be interesting to understand.

Thanks!
Author:  tomele [ Wed Feb 01, 2017 9:20 pm ]
Post subject:  Holy Graily Bob 'n Dotty

Hi youcrazykid.

The settings for the contestants in my rat race were the OOTB ones. This means I didnt change anything. Except the "trendfollowing" ones for which I published my settings in the first post.

Time has gone by and the EAs were subject to changes. Just read the threads. Some inputs disappeared while some new came up.

The results of my rat race are void for the actual versions. At the time they were posted, they served as a hint on which EA to concentrate as its results were the most promising ones. I couldnt repeat the results with more actual versions yet.

The EAs underly a permanent process of change. Sometimes this improves things, sometimes it doesnt. As soon I will have a version/setting combination that shows appealing results, I will start a new topic.
Author:  SteveHopwood [ Mon Feb 06, 2017 6:27 pm ]
Post subject:  Holy Graily Bob 'n Dotty

A quick update folks. Thomas and I have not forgotten HGBnD. Thomas has made some improvements to the code.

Old stagers, do you remember sixths? Selling in the top sixth of the chart and buying in the bottom? I shall be introducing them toe HGBnD soon.

Attached is a version of the indi I just introduced in the DFC thread - I developed this to create the code for HGBnD.

The attached version calculates the distance between the high and low then divides this by 6 - there is a user input that allows you to play with this divisor. The dotted yellow lines delineate the top and bottom sixths. The idea is to take HGI buy signals only when the market is below the bottom dotted line, and sell only when the market is above the top dotted line. Hedged grid trades will be placed no further than the peak hilo of the next time frame up.

Drag the indi and then HGI onto a few charts. The display looks very promising.

I am just working out the code for some of the detail. We should have an updated version in time for next week.

:xm:
All times are UTC Page 17 of 26