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

Gravy Trains & Weather Vanes
https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=5238
Page 8 of 17
Author:  lwtrade [ Sat Sep 02, 2017 8:36 pm ]
Post subject:  Gravy Trains & Weather Vanes

This week I had as much gain as open dd so no gains really.
Of the gains I have, 70%+ were red BE's.
(RisklessPips, I know you have addressed this situation and will look back for your comments).
Interestingly about 1/3 of the gains came from one chart I had on 15mins. The rest were 1 hr.
Netted about 10x the pips as P/L. Ha, wish this was the other way around!
May try higher TF's this week as you mention RisklessPips.
Author:  RisklessPips [ Sun Sep 03, 2017 7:34 am ]
Post subject:  Gravy Trains & Weather Vanes

lwtrade » Sat Sep 02, 2017 10:36 pm wrote:This week I had as much gain as open dd so no gains really.
Of the gains I have, 70%+ were red BE's.
(RisklessPips, I know you have addressed this situation and will look back for your comments).
Interestingly about 1/3 of the gains came from one chart I had on 15mins. The rest were 1 hr.
Netted about 10x the pips as P/L. Ha, wish this was the other way around!
May try higher TF's this week as you mention RisklessPips.
Hey lw

Great to see your report.

This very much confirms my own results at and below h1. The sweet spot is harder to find because price switchbacks are sharp AND frequent. Thus getting a SL and profit stop setting that increases average wins and lowers average loss is challenging.

I do think manual trades added to spiky prices may help but how much is not a test I'm ready to conduct :smile:

At H4 and above price movement is smoothed out and the pedestrian pace we have discussed previously gives GTWV space to work out it's logic in a more sedate environment.

One other thing August was notable for low volatility - summer holidays I guess. While I think GTWV will work in ranging markets because price decided actions are just as valid in those, as a rule it will do better in trending volatile markets. Lets hope September kicks it.

Charles
Author:  RisklessPips [ Tue Sep 05, 2017 9:26 am ]
Post subject:  Gravy Trains & Weather Vanes

GTWV and PoM

A while back I designed an indicator, coded for me by Rene which went on to become a fully fledged EA called PoM.

If you want to know more about PoM read about PoM here

The problem was I could never get PoM to trade profitably or consistently. On reflection maybe I never gave it time or I was using it in the wrong market environment.

Which brings me neatly to the crux of my post today.

We all know markets and individual pairs have identifiable characteristics. Thus we have trending and ranging markets, volatile pairs, overbought / oversold pairs etc etc.

Whilst GTWV relies entirely on price to dictate new market entries (at BE) and exits (at SL/TP) there are a number of important characteristics price does not "readily" reveal about itself that may help us to improve returns using GTWV.

It is ideal (but difficult) to know

1) Is the price currently trending or ranging ?
2) What is the current "speed" of movement of price based on its recent range ?
3) What is the current "strength" of that speed i.e is the speed sustainable for some time or not ?
4) What is the current broad market status in terms of 2 and 3 above for all the pairs you are trading ?

Enter PoM

With the exception of point 1 above, PoM was designed to tell you the information on price characteristics identified above.

I recognize GTWV is a brilliant EA, but running one instance as I initially started testing cannot adequately manage pair and market differences if you are trading several pairs.

You must have at the very least one manager per pair working a different setfile per market condition.

My current test environment depicted below indicates what I mean.
GTWV PoM 1.png
Here you see a profile with 8 charts

7 of the charts are for the pairs I am testing. On each of these I have applied an instance of GTWV EA and depending on my reading of price action, I apply either a Ranging or a Trending setfile.

Ranging setfile - This is a cautious setting
1) SL is tighter to entry
2) BE is 1:1 with SL
3) Take profits at BE and trails are much closer to price so that winners do not become losers

Trending setfile - This is a more adventurous setting
1) SL is further away from entry (although still within < 1% of account balance)
2) BE is 1:1 with SL (Because SL is further away BE is also extended to give price more room to play)
3) Take profits at BE and trails are also further away from price so that winners can run.

On the 8th chart (bottom left) I position PoM set up to monitor (not trade) the pairs I am trading at the relevant TF.

PoM tells me the individual pair speeds, strength, direction and general market condition.

Using this information I can decide :

1) To add more manual trades for any "performing / volatile" pair. This should help to increase average wins.
2) To apply the ranging or trending setfile to relevant pairs. (I also use the slope of the EMA's to try and guage the price's movement before applying the relevant setfile)
3) Customise my trading for each pair to reflect what I see not what I think.

Charles
Author:  TraderJoeForex [ Wed Sep 06, 2017 11:09 am ]
Post subject:  Gravy Trains & Weather Vanes

Smart thinking :good:
Author:  RisklessPips [ Wed Sep 06, 2017 5:43 pm ]
Post subject:  Gravy Trains & Weather Vanes

A strategy is only as good as the numbers it returns be it profit, loss, number of wins etc.

It's not only important to know your strategy's metrics you need to keep them evaluated to determine changes in the strategy’s performance due to changes in the market and/or perhaps dues to changes in characteristics of traded instruments.

With this in mind - I went cap in hand as usual to Steve and asked him to code an analyser for me to tell me what GTWV is doing in numbers and the result in the Analyser EA attached below.

Thank you Steve ! :clap:

The setfile is very easy to understand and perhaps needs no changes for most users except for display purposes.

The output is also very easy to understand but allow me to elaborate a few points

The EA will report on what is in your account history tab based on the filter applied - Thus Today will show today's results and all history will show ..... yip ! You got it in one !

In the setfile you have two switches :

1) Showsorteddata - This allows you to see how each pair you trade is contributing (or not) towards your bottom line
2) Showunsorteddata - This allows you to see a consolidated total rather than broken down by pairs

These switches are not mutually exclusive and you can set both to true so that you see individual pair results AND a consolidated total at the bottom of the chart.

Note - I am trading 7 pairs so the EA is set up for displaying results for 7 pairs + the consolidated data - you may have to play around with the settings to get it to fit your monitor.

The Output

1) Closed - Total trades closed
2) Winners - Total winners
3) Losers - Total losers
4) Won - Total won
5) Lost - Total lost
6) Avg won = Total won / total winners
7) Avg lost = Total lost / total losers
8) Win rate = total winners / total trades
9) loss rate = total losers / total trades
10) Exp per trade = Expected return per trade in account currency over the long term
11) Exp per account unit = Expected return in percentage terms per unit of account currency traded over the long term

The last two numbers become more reliable the more data in your account however these are extremely important because NO strategy will earn its user a living if these are negative !!

Clearly the analyser gives dynamic feedback as each trade closes - but because it uses the entire dataset you provide in your history tab, unless you are changing strategy parameters daily and reporting on daily results, it will take time for new closed trades to influence past results.

However, you will be able to see what changes to your setfiles are doing to your results in due course.

In Practice

The formulae for expectancy (No 10 above) = (Avg win * win rate) - ( Avg loss * loss rate)
if this number is negative play around with the components above in your trade management to get it positive.

The formulae for expectancy (No 11 above) = (1+ (Avg win / avg loss)) * win rate - 1
if this number is negative play around with the components above in your trade management to get it positive.

Example of practical application:

As a result of last weeks results I decided I needed to increase average wins and lower average losses on my test TF's.

Well I have by playing around with BE, TP, trails

However, in doing so I also increased the loss rate too much by manipulating the SL. Consequently even though each trade is now on average winning more than the average loser, due to the high loss rate GTWV is losing rather than making money.

Now I can keep my BE etc steady and play around with the SL

I would have struggled to pinpoint this without the analyser !

Finally as usual - no change requests to Steve -This is coded as I requested - any and all output or logic errors are my sole responsibility.

Charles
Gravy Train Analysis EA.mq4
ps - As a bonus for other strategy users, if you leave the magic number as is at -1 it will report on whatever is in your history tab, they do not have to be trades created by GTWV.

All strategies can therefore have these key metrics reported by use of this EA :yahoo:
Author:  RisklessPips [ Thu Sep 07, 2017 5:36 am ]
Post subject:  Gravy Trains & Weather Vanes

"If you only have a hammer in your toolbox, every problem looks like a nail."

I've been thinking about this with regards GTWV and as I wrote earlier one setfile just doesn't cut it for all the different market and pair price conditions.

Further back before that I also wrote about my protocol for listening to price. These are the rules I use to listen to and do price's bidding (or asking) :smile:

When trading manually there is an element of discretion which improves continuously the more you interact with price and it's action on different charts. For instance you can adjust stops on the fly for individual pairs and market conditions.

When using GTWV EA much of that discretion is binned for the convenience of automation in the setfiles.

Question - How best to balance the benefits of being closely in touch with price thru manual trading and the convenience of automated trading thru the EA ?

Price has different characteristics and it is vital we respect those characteristics as we listen to what it is saying. To disregard them is to be deaf to the context in which price is communicating and that I fear is a certain recipe for loss!

I have put together my initial thoughts on different price characteristics and attach them below.

I would be interested in your thoughts on how this protocol can be made more practical (indeed do you think it is valid ?) so that it serves as a better reference point for working with price action.

Thanks

Charles
Price characteristics.pdf
Author:  RisklessPips [ Thu Sep 07, 2017 9:25 am ]
Post subject:  Gravy Trains & Weather Vanes

For the visually inclined :

In this image the price is clearly in "choppy mode." - It's like the proverbial elephant. Difficult to describe but you know it when you see it !

Long wicks all over the place, long and short candles jostling for attention - a very uncertain (choppy) price action.

Trading this with the same setfile as I would a calmer environment is surely asking for losses.

The candle before the current one would definitely have taken BOTH my trades out had I used the "standard" setfile.

As it is, I applied my "choppy setfile" with extended SL settings and I'm still in the game !
Choppy environment.png
Charles
Author:  lwtrade [ Thu Sep 07, 2017 11:31 am ]
Post subject:  Gravy Trains & Weather Vanes

A couple of thoughts...

The point is good. And most of these points in the PDF, could be programmed into a function which could alter how GTWVea behaves. I think the question is, are we in these moves long enough for the ea to adapt but early enough to take advantage of these "environmental indicators" (like if last 10 candle wicks average is greater than x)? As you know the problem is these differing environments blend into each other, to our great frustration!

I also further agree (not that I have studied it) with your earlier point of session characteristics, and wonder if this is repetitive enough to take advantage of (and easier than trying to define the rules for the changing environment?)

Thoughts?
Author:  SteveHopwood [ Thu Sep 07, 2017 10:20 pm ]
Post subject:  Gravy Trains & Weather Vanes

Charles, there are thingies that mathematicians, statisticians et al go in for that drive me nuts because they look like pretentious bollocks to me. You quote this:

(No 11 above) = (1+ (Avg win / avg loss)) * win rate - 1

Given the left hand side of the equation, what I coded was:
(No 11 above) = (Avg win / avg loss) * win rate

So what is it with the preceding +1 and the succeeding -1?

Please do explain how my simple equation fell fowl of some abstruse mathematical requirement.
Author:  lwtrade [ Thu Sep 07, 2017 10:47 pm ]
Post subject:  Gravy Trains & Weather Vanes

If when price closes on the other side of our support / resistance level, is our acceptance that price is now going wv, then does it stand to reason that we should close the original direction rather than wait for the sl?

LW
All times are UTC Page 8 of 17