Gravy Trains & Weather Vanes

Locked
User avatar
RisklessPips
Trader
Posts: 246
Joined: Mon May 09, 2016 2:24 pm
Location: Nairobi, Kenya

Gravy Trains & Weather Vanes

Post by RisklessPips »

Wavegarrick » Sun Oct 01, 2017 3:09 pm wrote:Hi Charles,

"Being in the Trenches",
I've had a long and hard look at the words you describe in your thread. Powerful words indeed and so true. I know, I slept and fought in the trenches in my younger days, a warfare I did not want to partake in but something I was forced to do. The longer you stay in the trenches the more chance of getting into harms way.

I believe the same applies to trading. Why not be a sniper? Get into the trenches, make the kill and get out.
Something I am trying now in my trading, thanks to your thoughts.

I have gone back over many Ea's and Re assesed my original thinking and believe I am making headway.

I hope to share at some point.

Many thanks
Leon
Leon

I think I know what you are referring to when you say you slept and fought in the trenches - can't even begin to think what that must have been like - I hope my analogy is not offensive.

Let me address your question more directly.

If the expected outcome is profit then sniping may work as well as trench warfare. - The military have developed the terminology and traders call them scalpers and day traders at one end of the spectrum (Snipers) and swing or position traders at the other (trench warriors).

On that (profit making) level it probably does not matter which one you are.

However at another more basic level I do think it matters and in fact it ultimately impacts the ability to make profits as well.

You may have seen my latest book review on Dr Tharp's work. A key area he identifies as weak in most traders is decision making, stress control and emotional balance.

Now we don't need Dr Tharp to tell us that these three areas and particularly making decisions is fraught with difficulties for most people in many aspects of their lives let alone something as difficult as trading. Ever seen people in a supermarket trying to decide what flavour of something or other to buy ? You get the drift.

Sniping (scalping, day trading, low TF trading) requires continuous decision making that many people find difficult and uncomfortable. This discomfort induces a level of stress and emotional imbalance which makes the next decision even more difficult to make and so on.

Again with a military analogy - you know when you have had a number of straight losses how it becomes increasingly difficult to "pull the trigger" on the next trade.

I think this is why people often look for affirmation that their "sniping picks" are valid. They want their decisions validated and given meaning by others.

In an earlier post I wrote this decision making responsibility is shifted squarely back onto price in GTWV. In the trench you have no option but to do what price tells you. The critical element here is price action is making the decisions by virtue of confirming the direction and levels of participant interactions in the markets.

As a trader you then have confirmation and affirmation that what you are doing is reasonable and right. You just follow the ultimate leader and decision maker - Price.

Is there more risk doing it this way ?

I honestly don't think so. The only thing may be that as a sniper I erode my position far slower, by virtue of trading less, but if I am a poor sniper I will still be taken out eventually.

Charles
ps - Ever wondered why Index tracker funds are generally more profitable than actively managed ones ? - Decision making ! The former just follow the index no thought, no invention. The latter try to "beat" the market with (very) mixed results mostly not good.
Trading is a mind game - good job I have a brain
User avatar
RisklessPips
Trader
Posts: 246
Joined: Mon May 09, 2016 2:24 pm
Location: Nairobi, Kenya

Gravy Trains & Weather Vanes

Post by RisklessPips »

mjws00 » Mon Oct 02, 2017 11:37 pm wrote:Charles,

It's a pleasure watching you work through this, as your posts are interesting and well written. If nothing else you are building an interesting trade management tool.
Thank you Mike.
1) You are trying to enter in the direction of stacked moving averages. Which we know will work ok in a trending market, and get utterly destroyed in choppy or ranging conditions. Your win rate, and losses larger than wins is pretty much textbook.
Correct
2) You are manually using support and resistance to confirm or reverse your trade.
Price and nothing else. confirms or reverses my trades.
3) You've added indicators now as filters and you've picked strong ones. But hard to say if they are +EV over the longer term when market conditions change.
The indicators are not used as trade entry filters - they are used as trade management filters. Price alone tells us when to get in or out, the indicators alert us as to different actions necessary while we are in.
The rest is trade management, which is almost never an edge in and of itself. It adds no EV, it can only help you maximize what is already there. You can dial in Stacking, Drawdown, SL, TP, BE+ jumps etc over time.
I respectfully disagree. Trade management which includes risk and money management, personal psychological control including knowing when to get out seem more important to me than entering a trade at one or other (ultimately) subjective level.

Some thoughts for conversation:
What purpose does the first trade serve, if it is only a pointer to SR, and something you literally trade against later almost 50% of the time? A dot on the chart would tell you the same thing without trade costs or drawdown. Feeler trades CAN be awesome if and when they connect a trader to a market.
The first trade gets you into the trenches where you can start to follow price continuously - see my post above in response to Leon. Like you say seeds can be awesome - and in fact with proper trade management letting the 50% that price approves win should make the (smaller) losses on the other 50% a "business cost" worth bearing.
I often enter light and pyramid in, because it connects me to price and helps with clarity and focus. But they are NOT pseudo random entries.
Exactly what GTWV does. For a system taking 40 trades per week - that works out at 8 trades per day of which 6 are not random. They are entirely driven by price action. The other two are based on a simple evaluation of MA's, S/R. Simplicity, of course, should not be confused with randomness.
If you are reacting correctly to SR those trading decisions are your edge.
Crux of the matter right there. However, I am trying to react to price alone using strong management rather than entry decisions. Whilst the SR zones are critical to a "cherry picking" strategy I believe they are less so to a continuous price following methodology - we move past and beyond or fall back from these points in tandem with price.
Do Gravy Trains and Weather Vanes add value? Or have you learned to trade SR and built a tool that will help you with money and trade management?
6 million dollar questions :D
A few simple rules and your skill, brain, and experience may yield better results than automation and trying to curve fit a system to a market that always changes.
Always true Mike. Always true.

Thanks

Charles
Trading is a mind game - good job I have a brain
User avatar
RisklessPips
Trader
Posts: 246
Joined: Mon May 09, 2016 2:24 pm
Location: Nairobi, Kenya

Gravy Trains & Weather Vanes

Post by RisklessPips »

GTWV - Trade Management Under The Microscope

Below are two pairs superimposed with GTWV's trade management filters ADX and HGI.

My initial thoughts

1) HGI makes the market condition call before ADX

2) ADX makes more sensitive (i.e sharper) calls than HGI

Lessons ?

Working on it.
HGI vs ADX USDHKD.png
HGI vs ADX AUDUSD.png
Charles

ps - And BTW - the MA's are giving a good account of themselves too. Look at their slopes and where they cross each other in relation to the HGI and ADX :D
You do not have the required permissions to view the files attached to this post.
Trading is a mind game - good job I have a brain
User avatar
RisklessPips
Trader
Posts: 246
Joined: Mon May 09, 2016 2:24 pm
Location: Nairobi, Kenya

Gravy Trains & Weather Vanes

Post by RisklessPips »

Results GTWV on H4 - EOW 6th October 2017

This is the first week's report for the new currency pairs and the more versatile and stable EA.

Weekly results :( (i.e from start of open demo account 2/10/17 to date - init balance 500 USD)
End of week - Expectancy per dollar traded = -35.9%,
Expectancy per trade -0.59 (Dataset 18 trades)
Absolute gain / loss -10.53 = -2.1% of account

Progress Report :
Weekly results - First week of October saw a difficult trading environment with a lot of ranging price environments. My thinking on ranging price trade management is evolving and I will be posting some thoughts during the course of this week. I did not use the new functionality in GTWV that closes trades in a ranging market if they are in profit - perhaps my thoughts should include that option - we shall see.

The method took 18 trades during the week which is statistically insignificant so the figures above are not hugely reliable. What they do though, is tell me there is much work to do.

Charles

PS - Although I am not reporting on them the test between HGI and ADX on H4 500 USD accounts is interesting. HGI is ripping it having held steady (Virtually no DD) and is overall flat after 2 weeks and in the trading environments on such a small account that feels like a win.
Trading is a mind game - good job I have a brain
User avatar
RisklessPips
Trader
Posts: 246
Joined: Mon May 09, 2016 2:24 pm
Location: Nairobi, Kenya

Gravy Trains & Weather Vanes

Post by RisklessPips »

GTWV TM with ADX HGI & RR - Update

I have been looking at the results for the various tests I am doing and it appears to me there is (occasionally) a cash profit particularly just before the Asia close.

This is then dissipated during the busier (more erratic ?) European and American sessions leaving the position flat or in loss by start of next Asia session.

I have therefore added a basic basket close to the Trade Manager to allow you to set a cash level at which to close ALL trades. Note - This is a cash level NOT Pips at which trades will be closed.

You will then have to enter new seeds manually ! Stay in touch with your charts :smile:

I have also "hijacked" the trade comment field because I want to analyse the results and types of environments in which GTWV TM is opening and managing trades.

For this reason the trade comment field will now tell you
1) if a trade was opened in an HGI Trend, RAD or Range environment. (If using HGI signals)
2) if a trade was opened in an ADX Trend or Range environment. (If using ADX)
3) If a trade was opened in a default environment

By reviewing the history tab results for profit / loss over time and doing some cross referencing I'm hoping this will allow us to see if opening trades in a particular environment is better or worse than others. Informed decisions can then be made.

As usual this code is now beta until proven bug free

Charles

EDIT - Updated code available on page 6.

ps - There is also a fix to the TP addition processing
Trading is a mind game - good job I have a brain
User avatar
RisklessPips
Trader
Posts: 246
Joined: Mon May 09, 2016 2:24 pm
Location: Nairobi, Kenya

Gravy Trains & Weather Vanes

Post by RisklessPips »

Results GTWV on H4 - EOW 13th October 2017

Apologies for the early post (which skews things somewhat) - I would have liked to see Friday out but I have noticed a small problem where the EA added a new trade when I would not have expected it to do so. I have to clear that up and because it is happening at BE it will take some time to trace and fix the code (or setfile) before close of market.

Weekly results (i.e from start of open demo account 2/10/17 to date - init balance 500 USD)
Last week
Expectancy per dollar traded = -35.9%,
Expectancy per trade -0.59
Dataset 18 trades
Absolute gain / loss -10.53 i.e -2.1% of initial account balance

This week :clap:
Expectancy per dollar traded = +1.2%,
Expectancy per trade +0.02
Dataset 22 trades
Absolute gain / loss +0.39 i.e +0.08% of account balance

Cumulative results
Last week was first week. Results from that week combined with this week reveal the following :
This week cumulative :clap:
Expectancy per dollar traded = -16%,
Expectancy per trade -0.25
Dataset 40 trades
Absolute gain / loss -10.14 dollars i.e -2% of initial account

Progress Report :
Weekly results - This week saw a positive expectancy of 1.2% per dollar traded and a 0.02 dollar expectancy per trade taken, of which there were 22 up on 18 from last week.

The small positive expectancy this week was obtained in a slightly less ranging market environment but was not sufficient to overcome the large negative results from the previous week. However the cumulative results below show the trend (pun intended) is in the right direction.

The method took 22 trades during the week and cumulatively that makes for 40 trades in 2 weeks. This is still statistically insignificant so the figures above are evolving and not to be taken as best indicators of the method's ability. What they do though, is tell me there is much work to do.

FYI - In the separate unreported test between HGI and ADX - the former has a positive expectancy of +12.2% per dollar traded and +0.36 dollars per trade. It has taken 34 trades and is equity up 7 dollars. Yesterday that HGI signal account had it's first basket closure.

Charles
Trading is a mind game - good job I have a brain
User avatar
RisklessPips
Trader
Posts: 246
Joined: Mon May 09, 2016 2:24 pm
Location: Nairobi, Kenya

Gravy Trains & Weather Vanes

Post by RisklessPips »

GTWV - Trading Edge

If you have been around here and the Forex world in general for any length of time, you know leaving an EA "to do its thing" without periodic intervention marks the user out as an ...... (fill the blank)

I am trying to develop a working relationship with GTWV with ADX HGI and RR that compliments the EA and perhaps establishes what may be called my "trading edge."

I welcome your views.

As noted previously, in my book, trade, risk and money management are where the deal comes together or falls apart.

I will use an analogy and then give an example of what I am talking about.

Analogy

As any business person will attest, the relationship between business owner and managers / employees is critical to the well being and success of that business. A co-ordinated business strategy ensures end customer satisfaction, resulting in wins all round.

My trading accounts are my business. GTWV EA represents my manager and the setfiles are my employees.

Clearly, leaving my manager and employees to "get on with it" is serious abdication of my personal responsibility in terms of successful outcomes.

Whilst looking at this problem it became apparent the EA will never do everything I need or want without structured complimentary involvement from myself.

This structured complimentary involvement from myself defines "the edge."

I use the setfile to give "my manager / employee" instructions on trade management tasks. Each setfile (work manual) describes many different combinations of market environment (Customer) and how to handle them.

The manager is also given instructions of how to notify me of activity that warrants my personal involvement for a decision. This is in the form of the various stops (SL / TP / BE Trails)

On being notified of an activity that requires my intervention I duly look at the situation and either decide to act or let the manager carry on as per work rules.

Example

On one of my tests I am trading AUDUSD and USDSGD. These two pairs have strong negative correlation.

In one scenario this week I had shorted both.

Wrong move right there ! But which should close ?

GTWV is unable to make this type of call because it cannot see both markets. This is where unnecessary and unacceptable losses arise.

Looking at the charts, I saw the USDSGD was in a trend short. The AUDUSD had MA's that were moving up sharply.

It was therefore clear, on probabilities, that waiting for AUDUSD to hit SL was going to incur additional and unnecessary loss. I shut it down for a small loss and opened up a trade in the long direction. That trade and the short USDSGD are now both in profit.

This is an example of how my involvement with the EA has added value.

Does this type of intervention and decision making in trade management qualify as an edge ?

In general terms I am more inclined to interfere and micro manage on the profit side and let the manager deal with the loss side.

What do I mean by this ?

If a trade breaks even, the manager is telling me good things are happening. I have given the manager instructions on how far behind current price to trail the stop. It is sometimes galling to see price almost reach TP and then retrace to hit a stop that is still some way away for a small profit.

So on being alerted that a price has gone past BE, I manually intervene if it gets to a point near TP. A point that I will determine as the business owner (discretionary trading).

At that point I will raise the stop so that if TP is not achieved I take more out of the trade than if I had left the decision to my manager. If TP is achieved - great. My manager will open a new trade and off we go again.

Comments please

Charles
Trading is a mind game - good job I have a brain
User avatar
RisklessPips
Trader
Posts: 246
Joined: Mon May 09, 2016 2:24 pm
Location: Nairobi, Kenya

Gravy Trains & Weather Vanes

Post by RisklessPips »

GTWV - Trend Vs Range Trade Management

In keeping with our attempts to create an ever sharper and more versatile trade manger I have been looking at how GTWV handles trade entry for trending and ranging price movements.

Currently the processing is
1) On hitting a SL, the manager adds a new trade in the contra direction
2) On hitting a TP, it adds a new trade in the same direction.

Problem : This logic can be justified in a trending market but does not stand scrutiny in a ranging market.

Ranging Market Logic
Conceptually, the ranging channel is defined by an upper and lower boundary. The higher bound being a resistance zone and the lower being a support zone. Price bounces up and down in this channel until a breakout at one of the boundaries possibly leading to a trend in that direction.

Better processing logic in a ranging price environment is as follows

Existing Long Trade
If SL is hit open a new long trade (in the same direction as the closed trade) at the support zone
If TP is hit open a new short trade (in the contra direction to the closed trade) at the resistance zone

Existing Short Trade
If SL is hit open a new short trade (in the same direction as the closed trade) at the resistance zone
If TP is hit open a new long trade (in the contra direction to the closed trade) at the support zone

Note: Depending on the width of the channel the new trades may go into loss before the bounds of the channel are attained providing the resistance / support that defines a range.

This range environment logic is now implemented in GTWV and as usual is switched on (default = off) by a setfile switch.

I will release this upgrade when I complete my testing.

The images below will make this post clearer.
EURCHF on first look.png
In this image we see a retreating EURCHF - under current processing on hitting the SL it would open a short trade.
EURCHF HGI Range.png
Here the trade management service HGI has identified a range and consequently (if the switch is set) the manager will now add a long trade on hitting SL IF the ranging price environment still persists at that time.

Charles
You do not have the required permissions to view the files attached to this post.
Trading is a mind game - good job I have a brain
User avatar
RisklessPips
Trader
Posts: 246
Joined: Mon May 09, 2016 2:24 pm
Location: Nairobi, Kenya

Gravy Trains & Weather Vanes

Post by RisklessPips »

GTWV - Even Followers Of Price Have To Read The Charts

I was reviewing my overnight trades this morning and came across this example of CADCHF which has 3 important lessons for me (and perhaps others).

The lessons are easier to follow by reference to the image below.

Lesson 1 -Taking Personal Responsibility

Just by a visual look at the chart I can clearly see this price is ranging and has been doing so since the end of September with a days worth of false breakout at the turn of the month on the short side.

When the last open trade closed I duly input my new seeds as per GTWV requirements.

This was where I made my mistake - I added a short pending seed at the bottom of the range (which has now triggered and is in loss) and a long pending seed at the top (which I deleted once the short side kicked in.)

Had I been reading my chart correctly :arrrg: , I would have seen this was a ranging environment and I would instead have added a long pending seed at the bottom of the channel (support zone) and a short pending seed at the top of the channel (resistance zone)

Remedy - I will shut this trade down and open a new long trade with a stop at the top of the channel. If the prices are still ranging when (and if) price gets there the new GTWV ranging feature will open a short trade and back down the channel we might go !

Lesson 2 - A Range Is Not A Trend

I do not trade with HGI and ADX superimposed on my charts - GTWV takes care of that in the background. However, I have superimposed both the signals to demonstrate lesson no 2.

HGI is showing very clearly ranging arrows (up and down) a little delayed from the start of the range.

ADX on the other hand has readings of below 20 (the "normal" cut of for a range) up to near 40 which is a very strong trend - all within the ranging channel !!

HGI is hands down picking the price environment in this example more accurately.

Remedy - I will keep the environment analysis of the two signal services monitored. I may change the code if ADX is selected as the service of choice to make the cutoff for recognising a trend much higher than the current level of 25.

Lesson 3 - Summary

GTWV follows price and is nominally blind to direction. This example is a good lesson of why there are no absolute truths in trading. By following price when it made a small false move outside the range at 4PM my time I shorted the pair when I should have gone long by failing to mange the fact that in the wider picture price range was the dominant force.
Lesson learnt.png
Charles

EDIT On further reflection I think I am making a knee jerk decision to lesson 2. ADX is a tried and tested performer. Instead of amending its accepted cut off limits for ranges and trends I will manage how GTWV works based in its readings in my setfile.

EDIT 2 GTWV EA is not about re-inventing the wheel. Neither is it about curve fitting. But grey areas of trade management arise which require compromise.

For the next upgrade release I will allow user choice in the setfile for the period to average for the ADX signals. Default is 30.

Here is the reasoning.

Smaller averaging periods (such as the standard 14 period) generate more sensitive, less lagging indications for a trend / range. This leads to situations such as that indicated in lesson 2 above.

Higher averaging periods (such as the default above) generate less sensitive, more lagging indications for a trend / range. This will desensitize ADX and allow it to call trends in what is (visually and by definition) a clear trend rather than an up or down movement within a broad range pattern.

GTWV uses ADX signals for trade management rather than trade entry purposes. Lagging but more reliable price environment indications are therefore acceptable.

30 has been chosen after looking at several charts and comparing the signal with HGI signals and visual reads. As indicated to avoid curve fitting on my part, users will be able to change this setting in the setfile.

Also note - On H4 30 represents a week's worth of data being averaged. On a day chart, 30 is the previous month's data being averaged.
You do not have the required permissions to view the files attached to this post.
Trading is a mind game - good job I have a brain
User avatar
RisklessPips
Trader
Posts: 246
Joined: Mon May 09, 2016 2:24 pm
Location: Nairobi, Kenya

Gravy Trains & Weather Vanes

Post by RisklessPips »

Results GTWV on H4 - EOW 20th October 2017

Weekly results
Last week :clap:
Expectancy per dollar traded = +1.2%,
Expectancy per trade+$0.02
Dataset 22 trades
Absolute gain / loss +$0.39 i.e +0.08% of account balance

This 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 account balance

Cumulative results (i.e from start of open demo account 2/10/17 to date - init balance 500 USD)
This week cumulative :clap:
Expectancy per dollar traded = -15.5%,
Expectancy per trade -$0.27
Dataset 63 trades
Absolute gain / loss -$17.32 dollars i.e -3.46% of initial account

Progress Report :
Weekly results - This week saw a magnificent positive expectancy of 18.4% per dollar traded and a 0.30 dollar expectancy per trade taken, of which there were 18.

The week had a lot of ranging prices for the 6 pairs traded and this performance in that environment is very encouraging. My micro management efforts played a part in these results too as I started to accommodate ranges more actively than hitherto.

The trade manager also now has range trading logic embedded in its management styles and although the test is still too new to deliver conclusions about this management enhancement I am hopeful it will cut down whipsaws and profit leaks as we progress over the next few weeks.

If you cross reference the results from last week and this with the the cumulative, you will see a discrepancy. Last week I published results before Friday was closed out. Well wouldn't you know it... the system closed 5 trades later that day for a whopping -78.9% loss per dollar traded, -$2.53 per trade taken and a total of -$12.64 loss on the day :oops: . Those 5 trades were not accounted for last week nor this week resulting in the discrepancy.

The fact that this week reflected a large positive expectancy and slight improvement on the cumulative results is further testament to the good week the system had in difficult trading conditions.

Cumulative results - The method has now taken 63 trades since account open at start of October. This is still statistically insignificant so the figures continue to evolve but very clearly in a positive direction. Work to do.

However, after the first week we have had two consecutive weeks of positive expectancy in largely ranging markets - this is encouraging and holds promise for better performance when price trends re-emerge.

3% drawdown (1% per week) after 3 weeks is something to shout about. :yahoo: This rate of capital erosion is manageable and provides opportunity to recoup the loss with better management practices over the coming weeks.

Charles

- I will post the latest upgraded code and itemise improvements in a later post before open on Monday.
Trading is a mind game - good job I have a brain
Locked

Return to “Manual trading systems”