Clemmo's Trading Odyssey

Something wicked-ass this way comes...

https://www.forexfactory.com/showthread.php?t=635625

Reading this one thread > Reading all of BabyPips

Easy.. look how many points it moves on average daily.

It's not just about volatility actually. the instrument/pair the has to respect the technicals first.

Also I avoid too illiquid pairs.

So something like this provides enough info for you?
https://www.actionforex.com/markets/top-movers/

it's not rocket science so I calculate it myself.

But the damn language barrier makes it difficult to explain.
well let's try.....

here's the formula:

- on the daily TF, subtract the low from the high of the last 300 candles, eg. candle 1 : 30; candle 2: 40 etc.
- divide the sum by 300

and voila.

So I've tried contracting 3 different people to build an EA for me. 3 failures so far.

The last one was particular surprising as they are well known on a different forum, and they code probably 3 (free) indicators a day for it and seem to be the only person who offers any help there; it seems like they can do anything.

I asked for a quote, got it, agreed to it, and a week later their response was basically. "I give up."

I had the same problem finding competent coders when I worked in the video game industry. Is it me? Am I hard to work with? I didn't think my idea was all that elaborate, but not being a coder myself it's hard for me to gauge these things.

Is there a sub-forum here on FF where you can hire coding help?

It seems like the scams in Forex have made traders so distrustful of any group effort, so much so that it's nearly impossible to get a small group of people together to work on a common goal.

Oh well, I started this on my own and I'm going to finish it on my own.

Also, my fourth developer has now gone dark on me. The most surprising one of all as his name is synonymous with professionalism. Actually I'm kind of worried. He hasn't been here or his usual forum haunts for nearly a week which is unusual.

Well it appears I finally got my red wedge which means it's officially time to start ignoring me.
Lots of things coming down the pipe though.

My trouble finding reliable MQL coding help has redoubled my desire to get competent enough to at least modify bits of code that I can find elsewhere. Hopefully in time I can become creative enough to get very basic versions of my ideas made. I'm taking Emerenko's course on Udemy, following Coders_Guru 'course' and going through code snippets I find online. I'm nearly finished the UofT's 'learning to program' course although that's in Python so it fails some of the directness requirements of good studying. Got to start somewhere though. Most comp-sci courses (maybe all of them) fail to en-train basic design patterns (using rote learning) that are necessary for things to really 'stick' in the mind of a newbie programmer with little background. They present a list of concepts (easy) then show you how they (the instructor) build a program (easy for them, hard for the student) and then they expect you to figure it out (impossible).


I don't think I'll know enough to make a basic MA crossover EA until I've gone through a few more courses.

I also bought Andrew R. Young's 'Expert Advisor Programming' and will go through it, although it's the older edition.

Does anyone have any recommendations for C tutorials/websites where I can learn the ropes of MQL?

I might start a new thread just for automated systems that I build or just stick them here, I dunno what's better. I don't know what I expect from ForexFactory or posting on forums. It's just a trading journal after all.

https://www.economist.com/leaders/20...softheuniverse

Markets, warfare, manufacturing, chess. Machines have already taken over it's only a matter of time before the masses realize it.

Been a long time since I rock and rolled posted anything and that's not because I don't have anything to report. If anything it's because I don't know where to begin.

Let's start with some bragging.

Who would think that the pound would rise after the relief of finally Brexiting?

This account started out as a $3k deposit.

Here's results from another system that has unexpectedly yielded some fruit. I am intrigued because that happens about 5% of the time. Meaning if I make 20 systems, one works out in the long run. The results on this are kinda weird because I worked on it, found it too volatile, put it aside over nearly a year, then came back to it and made some refinements. I am slow, but dogged.



What I like most about this curve is it's extremely unlikely to crater, as it did at the start. At least I suspect that because of the principles on which it's based - namely diversification and correlation avoidance. I expect slow steady progress. Want faster progress? Risk more. That's the kind of system I like.

I tried to get some love for this system in another forum but ZERO FUCKS WERE GIVEN and everyone focused on some other trader's system, which seemed like a whole lot of bunkem to me (moving average crossovers ffs) but to each his own and green pips to all, right? I only cared because I was hoping to get some free coding out of it. Also it validates - Clemmo’s Theorem - traders gravitate towards the worst systems, people and trades. Wherever traders flock, wherever the threads are busiest, you will find the most shit. The busier a thread gets, the worse it gets. The best threads are often a single post.

So it's good news for my system too. If it was interesting to the crowd that would make it suspect.

Still, I find it a little frustrating that traders don't work together more, I mean, in a way that's not illegal. It's one reason they're so easy to exploit by scammers.

Speaking of coding - how's that going? Slow, really slow. And not so steady. I'm still terrible but I've been able to make a couple of things and will post more on that later. I might do some livestreams where I work through Andrew Young's books (shelled out for the newer edition - it's cheaper on his website than Amazon) and maybe a few others. I think I just need to follow numerous patterns before I can start getting creative.

On a personal note I'm starting day 4 of a 7 day fast. It's been brutal as I am a food lover. It might turn out to be a five day fast since I still feel pretty bad.

Everything I've read about it says things are supposed to get easier starting from today, so here's hoping. However everything I've read also says I should be losing 2 lbs a day and the scale is not cooperating. Why does my body love being fat so much? I guess I'll stick with it, after all...we're supposed to be disciplined, right traders? In theory I guess there are no fat successful traders. And I guess since I can't bring one to mind right now it must be true (availability bias alert).

That's another thing. Where's all this mental clarity I'm supposed to be getting?!

TGIF.

Clemmo’s Theorem - Ah Clemmo you make me chuckle.

Anyway what's wrong with manual trading ? I figure Forex has a good dash of chaos all through it , ask someone to build an EA based on chaos ..

Personally I think success has to really come down to money management. IC markets helps me out by closing my trades when I'm at max leverage and things turn south. They clearly want me to succeed. ( NB I'm actually trading ok atm, I havent watched my trades drop off for a while.. ok i mean since yesterday). I've discovered the trick is to have a $100 account , double it , pay your self $100, then flog the profit until it's gone and start again. I'm surprised I'm not a CEO somewhere. Imagine if I had staff under me.

Don't starve yourself, start walking. Walking is great. Get a dog and go for long walks..
Cheers h

Good to hear from you HB.

I used to walk and run frequently but when the weather dropped to -30 to -40 with the wind chill I slacked off. I will start again now that it's warming up. My wife wants a dog so it's probably inevitable.

You are right about money management, and there's nothing wrong with manual trading for some people. I think my success will depend at least partly on automated strategies. Know thyself.

Interesting idea about the $100 account. I guess the small amounts keep things on an even keel psychologically.

ps. keep me in mind when you start hiring.

Of course automated trading has its own set of pitfalls. Here's a system that was performing extremely well, but for the last couple of weeks has been doing this.

Is it time to make adjustments or just let this market 'phase' run its course? Someone with more experience would have a certain answer. For me it's trial and error.

Sore losers.

This trade went past my TP by 7pips and stayed that way for nearly 12 minutes. What is the deal, TradersWay?
Little tricks like this add up to losing accounts no matter how skilled the trader.

When I glanced at this account on my phone the title had changed from the broker name to this cryptic message. I think DJ = Dow Jones, and Sterling Spot Rates must refer to GBPUSD rates. Anyone seen this before? It was only up for a minute or two before the broker name was put back.

Truth, Beauty, and Justice
A price action modus
(note: these ideas are from my personal journal but some notions might be useful to others so I'm posting them here - not ready to talk specifics about most of this yet)

Background:
At some point I’ll release the videos on price action theories from a math/random walk perspective but here are some shower-ideas that occurred to me.

First of all, these ideas are all related to Extrematron, which I should not be so bullish about (because in systems dev you need to kill your idea babies, often and continually), but I can’t help it. Secondly, as I’ve written elsewhere, the term ‘price action’ makes me cringe as it is so loaded with poorly defined preconceptions by so many other trading gurus (and disciples), most of whom can’t define their ideas with any accuracy.

However, a lot of the concepts in this system are basically derived from ‘price action’ and even the idea of the ‘swing high/low’ which I first heard about and saw in none other than a Huddleston video, is the core of the whole thing. So I owe a debt even to the price action gurus.

Secondly this is a move away from my usual band-heavy focus and an attempt to narrow down the bare essentials needed to trade the ‘Law of the Jungle’ concept (maximum/optimum capital efficiency on all TF).

Custom Concepts
RE (aka high/low swings/aka fractals/aka peaks/troughs)
PZ Protrusion Zone/Phoenix Zone
Phoenix line
Brush-off
CCC
I might add links later but you can find all these ideas in other docs

Precepts
I had forerunning ideas - can I remember them?

1. All charts are abstractions of reality - but they are a useful abstraction. Bars/candles don’t tell us the whole story, but where I used to think of them as completely useless they provide a frame/outline/construct for tracking time and fractality.
2. Time frames are fractal in the sense that a chart on M5 is indistinguishable from a chart on W1 at the same zoom level. Again, we thank Mandelbrot
3. Mandelbrot assures us that past prices affect future prices even though I’m also sure that almost all traders/sell-side ‘advisors’ believe too much in the forecast-ability of prices. There is more randomness than most traders are willing to admit. This is part of the ‘bounded randomness’ I mention in other versions of GUFT. Trading this bounded randomness properly is where dumb/blind/simple systems rule.
4. Wyckoff identified the only real pattern in trading which is trend-distribution-trend

Principles

1. When price changes direction it necessarily creates a peak or a trough. It doesn’t matter over what timeframe, it cannot NOT produce evidence of a turn in the form of at least a 3 or more bar swing/RE, where one bar is higher than its brothers or one bar is lower than its sisters. This is a fact that few would dispute. Therefore all bar patterns of this type signal a change in direction even if it’s only very brief.
2. Once price has reversed it can either continue or reverse once more. This is also a fact.
3. If price reverses again and then EXCEEDS the peak or trough that was created during the original bend/turn it signals a higher degree of likelihood that price will continue in that direction. This is because whatever reason(s) price had to reverse the first time has now been overruled. Or maybe it’s because some other general condition(s) conducive to the trend has re-asserted itself. Whatever the reason it’s now more likely for the current ‘trend’ to dominate as both external (speculator) and innate (microstructure) factors would see this break away from a past sign of reversal as supportive of the new trend. This idea won’t be as undeniable as the first two principles but still I suspect it would have wide support. I’ve seen many forum posts where someone tries to improve on the basic pattern of a reversal and detect it earlier, but I have so far found all these methods to be unreliable compared to the simplicity and ‘truth’ of a peak/trough break. Every valid RE break should be entered with a few exceptions (see below). The risk of failure increases with trend maturity.
4. Despite these facts there are a couple of problems with RE breaks.
   1. One problem is that these peaks/troughs can look identical and there are so many that it becomes a case of too many signals. This is because price is always moving up and down on all TF. The idea of trend essentially becomes meaningless in such an environment since any trend in progress can turn at any time. Price is really only moving up or down when compared to the most recent trade.
   2. While price can continue past a break it can also quickly reverse as the conditions that favored the retrace re-assert or external factors cause price to stop trending (essentially to resume retracing).

   
   

5. The solution to the first problem is to assign significance to peaks/troughs based on
   1. The amount of time that has passed since their peak/trough levels were last broken. This is an ultimate distillation of Gann/Hurst wave superpositions.
   2. The depth of their ‘elbows’
   3. Potentially other factors, like the brushoff, explained below

6. The solution to the second problem is to demand a minimum safe amount of continuation for price beyond the peak or trough level that was broken. The more price moves beyond this level the more likely the ‘breakout’ is part of an enduring trend. This is the PZ.
7. When price breaks free and clear of a PZ it won’t always do so cleanly. The solution is the phoenix line and setting stops to break-even once price has CCCe’d that line.
8. Despite all the probabilities and our best efforts, price can still ‘range’ after breaking an RE. The solution is add a loss-recovery rule to take profits more opportunistically when there is a trend and offset losses sustained during the range.
9. A trend is over when the floor drops (for longs) or the ceiling rises (shorts). Some credit to Laurent Bernut for this.
10. When price fails to make a higher high or lower low but does not exceed the PZ, that’s a ‘brush off’ a sign of weakness and so caution is advised when adding trades to an existing trend bias. Some REs could be ignored in such a state.
11. This is all we know or care to know about the future price action. All the complexity of price arises out of the 4 tendencies (see GUFT4) and to interpret more in their random squiggling is to follow the whip-poor-wills into the mists. This has been the doom (and still is) of many otherwise sensible traders.

This is what I’ve been trying to define - the simplicity of the whole thing - is as elegant as I can make it? I have trouble finding the words for this feeling, but it's almost a relief, like finding a proof, though of course I have a long way to go.

1. Truth - the RE breaks are the end result of every correct forecasting method you can concoct, all the indicators, and setup patterns, the Al-Brooks-type pontifications and cyclical minutiae are simply seeking out this resultant behaviour, so why not cut out the middleman (and wasted effort) and just enter when it happens?
2. Beauty - the simplicity of drawing this phoenix line in the sand and riding out what may come. If it returns, we rinse and repeat, if it doesn’t return we have released our young into the wild and our parenting duties are done.
3. Justice - the only demanding part of the whole process since I don’t have an automatic way (yet) to foresee a breaking ceiling or floor, but we can simply do a check every bar I guess. We get our just come-uppance each time.

This used to be a longer post but then I realized I'm focused too much on what other people are doing and not staying in my own lane. Much like my wife's driving. So this is all that's left and it's meant as advice for me to heed, YMMV.

1. Clemmo's Theorem is alive and well. Traders are more interested in sticking with the herd and using what everyone else is using even if it puts them at risk. You can use this to your advantage by looking at what gets ignored. I found a cool adaptation of a 3 bar fractal entry (credit Carlos Luis F) that I'm testing now. Hint/warning: I'm the only one that likes it.
2. No one is much interested in testing and checking original claims on their own. Most people sort of wait around in a circle and ask each other 'is it working for you? Should I invest time in this?' Assuming that if enough people give it a thumbs up it must be worthwhile. However it's a case of blind leading the blind.
3. What you WANT might not be what you NEED. I see this a lot and I'm guilty of it myself. If you have an idea for a system or an EA and you carefully define its expected behaviour you'll still end up making a lot of on-the-fly changes because some assumptions were lurking in the background that you didn't know were there and when these are tested and disproved you have to change your theory slightly. Often this process goes on in a sort of tug-of-war that either completely changes your system or you bin it as non-viable. Recall Aesop, "We would often be sorry if our wishes were gratified."
4. Unexpected behaviour in a system isn't always detrimental; until you do some careful checking (MT4 backtests don't count) you won't know for sure, and even then you probably won't know for sure.

I'll title this lecture "An examination of relative extrema (fractal) break methods." The first examination is a 3 bar fractal break that takes a trade after the last bar in a sequence is broken by price moving up or down. The test looks simply at win/loss as a function of entries generated by signals and a static 30 pip (1:1 R:R) SL and TP. There might be ways to improve performance by using more sophisticated trade management, but additional rules could also interfere with the performance of the fractal break test so for my purposes this is enough.

The first chart shows a visual backtest using [useless method omitted], done by hand.


It might be tricky to see what's going on but the important bit is the stats. This is a random system. I was able to improve it slightly by creating a new rule. (re-breakouts which are basically phoenix entries)

However compared to the more traditional method of measuring and trading fractal breaks, this system is no improvement.



To be fair this chart was done by hand and so I'm likely favoring 'significant' peaks and troughs using intuitive criteria that I haven't defined and would be useless to anyone who wasn't me, especially to an EA. I did a second test placing SwingMan's properly coded indicator on a chart and overlaying my 'flawed' version beside it.



I extended the test out to the current date and again compared winning versus losing trades using the simple fixed stop and take profit method. As you can see neither system is great, but for some reason one indicator wins 8% more, a statistically significant difference. What's funny is I'm not even sure why the 'bug' in my code is making this better? It doesn't have re-breakouts coded in yet or I'd take the credit. It could be statistical variance; more testing is needed but with this many trades I would be surprised if it was pure luck.

I will continue to look closer at how these two indicators differ, because there might be a valuable insight into how fractal breaks can be improved. For now what matters most is that it doesn't matter what break-out method you use - they all appear to hover around the 60% success rate. This can be demonstrated by the following test which I made using peak/trough breaks.



OR SO I THOUGHT
Credit to LuisF with this method - which performs significantly better than anything I've tried so far.



What's the secret sauce? In his own words,

Very interesting. Two levels need to be broken before an actionable signal is generated - similar to the PZ idea, this stops entries from being made near the top of a crest or bottom of a dip. Something else that needs to be investigated - is it even beneficial to wait for a level break? Maybe n bars forming HLs (long) or LHs (short) is enough to justify an entry? Check it.

This leaves me with the notions that I wrote about in 'Truth Beauty Justice' which is to expand a zone around the entry (although my early tests suggest it needs a minimum width >= the width of the price channel and even then its value needs to be proven) and to make use of the 'significance principle' which is easy to see when marking a chart manually.

What's more difficult is coming up with a formal definition of what constitutes 'significance', especially the need to distinguish deeper, pointier 'spikes' from shallower rounded ones.