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

Gravy Trains & Weather Vanes
https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=5238		 Page 15 of 17
Author:  Dragonian [ Sat Oct 21, 2017 1:45 pm ]
Post subject:  Gravy Trains & Weather Vanes
Thank you Charles, looking very promising.
:hi:
Author:  RisklessPips [ Mon Oct 23, 2017 5:44 am ]
Post subject:  Gravy Trains & Weather Vanes
GTWV Trade Manager Enhancements V 2.01 Released

This code is now available on page 6 of this thread Here

The manager now accommodates ranging price environments as follows

1) If prices are ranging at BE a new trade is not opened. This prevents the bounds of the range from being broken by illogical SL and TP levels.

2) At SL a new trade in the same direction as that closed can be opened if the range processing is set true AND prices are ranging.

3) At TP a new trade in the contra direction to that closed can be opened if the range processing is set true AND prices are ranging.

4) The ADX signal processing has been amended to allow users to determine their preferred averaging period for calculation of ADX. Default is 30. The standard period of 14 makes ADX too sensitive and it records retracements and rallies as trends when in reality they are short corrections within a ranging pattern. A 30 averaging period on H4 allows less sensitive, more lagging but more reliable recognition of ranges and trending movements.

5) All these modifications as usual are driven by user controlled switches and variables giving you ultimate control over the versatile functionality.

I have put this code on the release page as v 2.01 as I am 97% certain the code is doing what it should. There have been a few minor anomalies to do with the setting of SL and TP but I cannot recreate these after spotting them and that leads me to think either I misread the error or somehow my compiled code was corrupted ?!?@#

Charles
Author:  RisklessPips [ Fri Oct 27, 2017 5:27 pm ]
Post subject:  Gravy Trains & Weather Vanes
Kenya is a tad “hot” at the moment and I don’t mean the weather – so I have decided to post my results now after the final bar on H4 has commenced and I am (reasonably) sure there won’t be any more closures before market close my time !

Results GTWV on H4 - EOW 27th October 2017

Weekly results
Last week :clap: :clap:
Expectancy per dollar traded = +18.4%,
Expectancy per trade = +$0.30
Dataset 18 trades
Absolute gain / loss +$5.46 i.e +1.09% of initial account balance

This week ;)
Expectancy per dollar traded = -8.7%,
Expectancy per trade -0.15
Dataset 29 trades
Absolute gain / loss -$4.23 i.e -0.85% of initial account balance

Cumulative results (From start of open demo account 2/10/17 to date - initial balance 500 USD)
This week cumulative :clap:
Expectancy per dollar traded = -13.4%,
Expectancy per trade -$0.23
Dataset 92 trades
Absolute gain / loss -$21.55 dollars i.e -4.31% of initial account

Progress Report :
Weekly results - This week saw a decline in expectancy from +18.4% per dollar traded to -8.7% per dollar traded or alternatively from +0.30 dollar expectancy per trade to -0.15 dollar expectancy per trade closed.

This week there was a total of 29 trades closed which is 55% more than the average of 18 trades closed per week for this system at H4.

Why was there increased trading ? – I’m not sure. I suspect the ranging prices (as explained further below) may have contributed to increased frequency of trading with resultant poorer outcomes.

The modest expectancy decline is a cyclical event and to be expected.

Cumulative results - The method has now taken 92 trades since account open at start of October.

Although still statistically low, this is now becoming a number from which we can deduce possible future performance and trading actions required.

Since the beginning of October, GTWV has improved cumulative expectancy every week after the first. It now stands at -13.4% per dollar traded or -$0.23 per trade closed.

These headline expectancy numbers mask some very important details (and indicators of possible future action) as described below :

Using the GTWV Trade Analyser the following details are readily apparent when taking all closed trades across all pairs into account :

Average win is $2.16 per closed trade
Average loss is $1.77 per closed trade
Win rate is 39%
Loss rate is 61%

These figures are extremely pleasing and auger well for long term profitability if maintained !! A system that wins 40% of the time can be successful if it wins more on average than it losses. GTWV is meeting this requirement – Work to be done but the numbers don’t lie. :cheer:

Drilling down deeper we find :

Worst performing pair SGDJPY

Average win is +$1.41 per closed trade
Average loss is -$2.37 per closed trade
Win rate is 20%
Loss rate is 80%
Expectancy per dollar traded -68% , Expectancy per trade -$1.62
10 / 92 trades = 11% of closed trades

Best performing pair EURCHF

Average win is +$3.11 per closed trade
Average loss is -$2.84 per closed trade
Win rate is 64%
Loss rate is 36%
Expectancy per dollar traded +35% , Expectancy per trade +$0.99
14 / 92 trades = 15% of closed trades

What is the key difference between these pairs ?

SGDJPY has been is a long term ranging pattern at the same time that EURCHF has been trending up !

October 2017 Conclusions :

Clearly trading ranges is where I need to focus most on in terms of my micro trade management as well as in terms of the management style settings used in the GTWV manager.

Charles
SGDJPY Oct.png
EURCHF Oct.png
Author:  RisklessPips [ Fri Nov 03, 2017 6:31 am ]
Post subject:  Gravy Trains & Weather Vanes
RisklessPips » Sat Sep 23, 2017 6:12 pm wrote:Musings In Between Trading

But as things stand, in real terms of cold hard cash it is losing money.

Why?

I have come up with 5 possible reasons.

1) I am trading an account size that is insufficiently capitalised.
2) I am trading the wrong instruments
3) I am trading the wrong TF’s
4) Sub Optimal combinations of BE / TP / SL and trails to suit different market and price conditions
5) Something else

........

By deduction, if none of the reasons above are responsible for my dollar loss of equity it must be due to something else.

For instance

Challenge : GTWV does not work as a strategy or as incarnated
Solution: Go back to the drawing board.

Challenge: I am applying the strategy incorrectly
Solution: Review the strategy, ask more questions and apply the lessons

These are difficult questions to face but to ignore them is to blind myself – only by dealing with each in turn will I find answers.

Charles

Ps That something else could also just be a simple "give it time" :smile:
Musings Part 2

At the close of October we had initiated and developed the GTWV method for 3 months.

It was time to relax, kick back and reflect on lessons learned. Ploughing ahead without periodically stopping to think about any new venture is not a good idea :D

The quote above is from a post I made mid way through September i.e about half way through this experiment and it leads me to the crux of what I want to write today.

Even back then I had identified that the method GTWV had (and continues to have) merit. However, even then I was not content with performance and I noted the 5 areas listed above as possible weak links in the system.

I have spent this week thinking and analysing results across all pairs, TF's and combinations of manager setup. Here are my conclusions and proposed next steps.

1) I am trading an account size that is insufficiently capitalised.

I started of trading a small demo account of $500 USD. This was on purpose because an ambition of mine is to create a trading system that people in "less financially affluent" parts of the World can use to gain second incomes. - Lofty, I know, but what is life without uplifting dreams ?

The small account presents unique challenges because of need for risk management and the consequential difficulty to trade on a scale that allows significant cash accumulation. However, $1 USD in parts of the World goes a long way and therefore if I can devise a system that generates a few dollars consistently most weeks it might fly :cheer:

Conclusion : I am going to persist with a $500 USD account. (A side benefit of this is if it works on this size account it will probably work on larger accounts as well.)

2) I am trading the wrong instruments

I tackled this problem earlier when I recognised my trading of majors unquestioningly was a mistake. I then sought out pairs with ATR's that matched the risk tolerance afforded to me by the small account size.

This means there is no point trading a pair that moves 60 pips as normal when I can only cover 30 pips in my risk management.

Conclusion : The pairs I am currently trading with an average of 27 / 28 pips ATR at the height of session crossovers will continue to be in the portfolio.

There is however, one new factor in regards to traded pairs.

Originally as a leverage management tactic I only traded 5 / 6 pairs at 0.01 lots. That works out at 5000 units of currency traded on a a 500 USD account. That is a manageable 1:10 leverage position.

Whilst i still intend not to have more than 5 /6 trades open at any one time, I will now fish for possible trades in a wider pool of pairs that meets the ATR requirement. As discussed below my trade entry decision making is changing and I need a bigger pool of pairs to fish from because I will be taking fewer trades.

3) I am trading the wrong TF’s

I have tested TF's from M15 thru to W1. The lower TF's are too whippy and even though GTWV is direction blind the constant shifts whipsaw the system too much.

The longer TF's D1 and W1 have ATR's that really stretch the risk management profile of a $500 USD account.
For instance the average ATR of the pairs I trade on a D1 profile is about 60. This is really at the outer limit of the number of pips I want to risk per trade and it magnifies the effects of poor decision making by making the loss rate significant so that even a high win average does not compensate. :arrrg:

Conclusion : The account size and my comfort level dictate I restrict activity to H4. I am inclined to persist with D1 because I like the TF dynamics and if GTWV can tread water at D1 with $500 USD account, then I know if the account capitalization is raised, D1 will be a great TF to trade.

4) Sub Optimal combinations of BE / TP / SL and trails to suit different market and price conditions

This is a constant work in progress. I do however think I have attained an excellent balance between the parameters for all the different market conditions that GTWV EA manages.

Conclusion : Work in constant progress.

5) Something else

My reflections this week have centered on this point. The something else was me and my decision making.

As at the end of October GTWV had a small negative expectancy, which had improved consistently from the initial poor start to the month. This demonstrated to me the method is viable.

In its first incarnation, GTWV has been an "always in the market" method. Rarely out of the market. This has drawbacks but clearly the absolute loss of only 4% over the difficult 4 trading weeks of October shows they are not insurmountable because this was very low DD. (The weeks were difficult because of ranging prices on the traded pairs)

When I cut and diced my results I noticed something I instinctively knew but hadn't fully appreciated.

My seed trade entry method was a serious weak link in the method. Out of 56 seeds I entered in October, only 14 turned out winners i.e the overall win rate for the system was 40% but that of my manual seeds was only 25% !! :evil:

I now go back to posts from Leon (Wavegarrick) and Mike (mjws00).

Both guys in effect asked me why not snipe rather than engage in a trench (always in) strategy. My response was being always in took a major decision making problem away from the trader. I still think this is true however I have now concluded that entering the market with manual seeds needs better market entry signals than I have been using.

Conclusion : The system will now not be "always in" but the trade entry selection criteria will be mechanical based on chart patterns to ensure I limit the possibility of emotional or otherwise poorly determined entries.

Hopefully this will also increase the probability that seeds become gravy trains and raise the seed win rate above the dismal 25%.

I will re-commence trading and select no more than 5 /6 open seed trades at any one point from a pool of 14 pairs that have the ATR characteristics than allow my risk management tolerance level to work on a small $500 USD account at H4.

The trades will not be direction blind (as has been the case thus far) but will rely on a simple and mechanical entry system of chart patterns recognition.

I still have further analysis to do, but these above are the main ones and if all goes to plan I will recommence testing with a new setup (account and self) next week.

Thats all for now

Charles
Author:  lwtrade [ Sat Nov 04, 2017 5:50 am ]
Post subject:  Gravy Trains & Weather Vanes
Great post as always Charles. I personally appreciate your clear and honest explanations of your thoughts and analysis. I can see this thread being something people should thoughtfully walk through as a great example of practical development.

Great job !
Author:  RisklessPips [ Tue Nov 21, 2017 9:37 am ]
Post subject:  Gravy Trains & Weather Vanes
GTWV Mk II - Evolving Thoughts

On another thread Full post here naningbob says
When I do trade it's mainly HGI big arrows in the direction of my CSS reads. What currencies are running strong against what currencies are running weak and the big arrow that goes in the same direction at the beginning of a trade session. That pretty well sums up my trade strategy in two sentences.
Back in May last year when I put out the HGI and Sixths method I wrote this HGI and chart patterns

On this thread I wrote Gravy Trains preferred means to an end

Having spent November thus far revisiting the underlying basis for GTWV I am more open to influences regarding

1) How to keep it simple - as per nanningbob
2) How to look for confluence in a mechanically repeating method - as per HGI Sixths & chart patterns
3) How to produce more Gravy Trains - as per GTWV

Last week for the first time (lucky week :?: ) GTWV achieved an expectancy of more than one dollar per dollar risked for the first time. This on a 500 USD account.

I will document how I am now selecting trades in short order but before then does anyone want to share how they enter markets using a method that is

1) Simple
2) Mechanical and repeatable
3) Reliable (i.e works a minimum of 40% of the time

The new trade entry method I am using has also indicated to me that GTWV can be made less complicated in terms of user variables and options.

I will therefore be amending the code to create GTWV lite - but this is not critical to ongoing method development.

Comments welcome.

Charles
Author:  Wavegarrick [ Wed Nov 22, 2017 4:52 am ]
Post subject:  Gravy Trains & Weather Vanes
Hi Charles,
but before then does anyone want to share how they enter markets using a method that is

1) Simple
2) Mechanical and repeatable
3) Reliable (i.e works a minimum of 40% of the time
The most reliable method I find, no matter what system I use and whatever the market throws at me. This is the one that has been most successful for me over the years and I always fall back on it.

Nanningbobs 60 and 240 Linear weighted moving average set to open. There has been much debate over using different settings for different time frames. For me it does not matter.

I use 5min, 15min, 1hour, 4hour and Daily all with the same settings. All must have crossed in the same direction.
So for eg if you have 5min down, 15 min up, 1hour down, 4hour up, 1day up. It is simple you do not trade. The market is mixed.
All must have crossed in the direction of the daily trend before you enter. You can have multiple entries in an uptrend or downtrend so I would leave my gravy train running in the medium to long term and scale in and scale out as the trend becomes mixed again.
I use Rene's maddy dashboard matrix to get a quick visual over all the currency pairs to narrow down the pairs I want to trade and you could probably use Thomas's Multi time frame Ma indicator to time your entries.

This may not be for you or everyone and at times you wait for your entry. But reliable. Oh yes it is :)

Cheers
Leon
Author:  peterke [ Wed Nov 22, 2017 8:23 am ]
Post subject:  Gravy Trains & Weather Vanes
Charles, Here''s a basic trend entry using the HA smoothed and another indicator that sets the trend direction. Neither repaint and seem to be pretty consistent at getting entries early enough. At lower time frames it needs to be filtered a bit at the end of moves.
I've used the combination in several eas for defining the long and medium term trend direction on higher time frames for entering using different criteria on short timeframe and seems to work ok.

rgds

Peter
SIMPLE ENTRY EXAMPLE.JPG
Author:  RisklessPips [ Wed Nov 22, 2017 2:28 pm ]
Post subject:  Gravy Trains & Weather Vanes
Wavegarrick » Wed Nov 22, 2017 6:52 am wrote:Hi Charles,

The most reliable method I find, no matter what system I use and whatever the market throws at me. This is the one that has been most successful for me over the years and I always fall back on it.

Nanningbobs 60 and 240 Linear weighted moving average set to open. There has been much debate over using different settings for different time frames. For me it does not matter.

I use 5min, 15min, 1hour, 4hour and Daily all with the same settings. All must have crossed in the same direction.
So for eg if you have 5min down, 15 min up, 1hour down, 4hour up, 1day up. It is simple you do not trade. The market is mixed.
All must have crossed in the direction of the daily trend before you enter. You can have multiple entries in an uptrend or downtrend so I would leave my gravy train running in the medium to long term and scale in and scale out as the trend becomes mixed again.
I use Rene's maddy dashboard matrix to get a quick visual over all the currency pairs to narrow down the pairs I want to trade and you could probably use Thomas's Multi time frame Ma indicator to time your entries.

This may not be for you or everyone and at times you wait for your entry. But reliable. Oh yes it is :)

Cheers
Leon
Hi Leon

From the description this certainly achieves 1, 2 & on your recommendation 3. Simple, mechanical and repeatable, reliable.

One thing can you post a link here so I and others can find the meat of the method and how to use it as simply as possible for trade entry - I'm guessing it's in HGI somewhere but that is a lot of material to read.

On my test case plan.

Thanks

Charles
Author:  RisklessPips [ Wed Nov 22, 2017 2:33 pm ]
Post subject:  Gravy Trains & Weather Vanes
peterke » Wed Nov 22, 2017 10:23 am wrote:Charles, Here''s a basic trend entry using the HA smoothed and another indicator that sets the trend direction. Neither repaint and seem to be pretty consistent at getting entries early enough. At lower time frames it needs to be filtered a bit at the end of moves.
I've used the combination in several eas for defining the long and medium term trend direction on higher time frames for entering using different criteria on short timeframe and seems to work ok.

rgds

Peter
Peter

I will organise my entry methods test plan and compare this with what I am using as well as that suggested by Wavegarrick thru the next few weeks.

I will detail my test method once I have my head around these entry systems.

Thanks

Charles
All times are UTC Page 15 of 17

Powered by phpBB® Forum Software © phpBB Limited